From pollok at gmail.com  Sun Dec  3 18:47:19 2006
From: pollok at gmail.com (Alessandro Poletto)
Date: Sun Dec  3 18:47:55 2006
Subject: [ic] tree problem
Message-ID: <45737017.19832.911090@pollok.gmail.com>

I have a problem with the tree routine.

If I use the toggle/memo option of the tree option so i'll open a 
branch/field of the tree and I don't close the others.

I'll try to explain better...
I have a tree:

A
--1
----x
B
--2
---Y

Starting position is:
A
B

all works fine if selecting A

A
--1
B

but if after I select B

A
--1
B
--2

My need is this:
A
B
--2

I haven't found nothing about this problem and into the documentation 
there's nothing about the automatic collapsing of the unused branch.

Many thanks in advance to all.




============================================
Toth di
Alessandro Poletto
http://www.toth.it
E-mail: alessandro@polet.to
============================================

From prejnydyl at ozonetec.com  Mon Dec  4 16:34:17 2006
From: prejnydyl at ozonetec.com (years)
Date: Mon Dec  4 16:34:57 2006
Subject: [ic] upon parent
Message-ID: <000401c717eb$f3dbf320$00000000@tommeliten>

31547

From kdeugau at vianet.ca  Mon Dec  4 17:34:20 2006
From: kdeugau at vianet.ca (Kris Deugau)
Date: Mon Dec  4 17:34:37 2006
Subject: [ic] tree problem
In-Reply-To: <45737017.19832.911090@pollok.gmail.com>
References: <45737017.19832.911090@pollok.gmail.com>
Message-ID: <4574A26C.8060509@vianet.ca>

Alessandro Poletto wrote:
> I haven't found nothing about this problem and into the documentation 
> there's nothing about the automatic collapsing of the unused branch.
                             ^^^^^^^^^^^^^^^^^^^^
Please don't.

I don't know about anyone else, but "helpful" tricks like this generally 
just frustrate me while trying to navigate an interface.

-kgd
From pollok at gmail.com  Wed Dec  6 04:42:08 2006
From: pollok at gmail.com (Alessandro Poletto)
Date: Wed Dec  6 04:42:03 2006
Subject: [ic] tree problem
In-Reply-To: <4574A26C.8060509@vianet.ca>
References: <45737017.19832.911090@pollok.gmail.com>,
	<4574A26C.8060509@vianet.ca>
Message-ID: <45769E80.18146.A5A3895@pollok.gmail.com>

Yes... I think the same thing but customers are customers and what 
they ask me I have to made...



> Alessandro Poletto wrote:
> > I haven't found nothing about this problem and into the documentation 
> > there's nothing about the automatic collapsing of the unused branch.
>                              ^^^^^^^^^^^^^^^^^^^^
> Please don't.
> 
> I don't know about anyone else, but "helpful" tricks like this generally 
> just frustrate me while trying to navigate an interface.
> 
> -kgd
> _______________________________________________
> interchange-users mailing list
> interchange-users@icdevgroup.org
> http://www.icdevgroup.org/mailman/listinfo/interchange-users


============================================
Toth di
Alessandro Poletto
http://www.toth.it
E-mail: alessandro@polet.to
============================================

From DB at M-and-D.com  Wed Dec  6 14:59:01 2006
From: DB at M-and-D.com (DB)
Date: Wed Dec  6 14:59:21 2006
Subject: [ic] Setting title of results page
Message-ID: <45772105.5040309@M-and-D.com>

I have a page with the following search:

[query
    type=list
    more=1
    ml=150
    sql=|
        SELECT  distinct name
        FROM    cat
        |]
[list]
<h3>[sql-param name] items -
Click for a list of <a href="[area search=|
    se=[sql-param name]
    sf=category
    st=db
    co=yes
    op=eq
    sp=brands_results
    va=category_name=[sql-param name]
|]"> [sql-param name] items </a></h3>
[/list]
[/query]


Now in the body of the results page, the tag
[value category_name]
gets displayed as expected. However I also want that value to be used
for the results page's title. In the results page's template I have:

[comment]
ui_template: leftonly_2
ui_type: template
ui_name: leftonly_2
ui_version: 5.4.1
ui_label: Page with top/left areas 2.
ui_source: templates/leftonly_2
ui_template_layout: LEFTONLY_TOP_2, UI_CONTENT, LEFTONLY_BOTTOM

page_title:
        label: Page Title
        width: 50

members_only:
        code: members_only
        label: Members only?
        options: 1=Yes,=No*
        type: yesno

meta_keywords:
        code: meta_keywords
        db: ^@
        label: Meta keywords
        type: text
        width: 50

meta_description:
        code: meta_description
        db: ^@
        label: Meta description
        type: text
        width: 50

[/comment]
[set page_title][/set]
[set members_only][/set]
[set meta_keywords][/set]
[set meta_description][/set]
<!-- BEGIN PAGE_PICTURE -->

but if I put
[value category_name]
into the title slot in the admin content editor, the title of the page
is literally "[value category_name]" and is not getting interpreted as I
want.

Does anyone have an idea why the tag gets displyed correctly in the page
body but not in the title?

DB
From ic at 3edge.com  Wed Dec  6 14:11:19 2006
From: ic at 3edge.com (ic@3edge.com)
Date: Wed Dec  6 15:11:42 2006
Subject: [ic] Setting title of results page
In-Reply-To: <45772105.5040309@M-and-D.com>
References: <45772105.5040309@M-and-D.com>
Message-ID: <20061206191119.80C9D178A8@mail.papanikos.com>

DB writes: 

> I have a page with the following search: 
> 
> [query
>     type=list
>     more=1
>     ml=150
>     sql=|
>         SELECT  distinct name
>         FROM    cat
>         |]
> [list]
> <h3>[sql-param name] items -
> Click for a list of <a href="[area search=|
>     se=[sql-param name]
>     sf=category
>     st=db
>     co=yes
>     op=eq
>     sp=brands_results
>     va=category_name=[sql-param name]
> |]"> [sql-param name] items </a></h3>
> [/list]
> [/query] 
> 
> 
> Now in the body of the results page, the tag
> [value category_name]
> gets displayed as expected. However I also want that value to be used
> for the results page's title. In the results page's template I have: 
> 
> [comment]
> ui_template: leftonly_2
> ui_type: template
> ui_name: leftonly_2
> ui_version: 5.4.1
> ui_label: Page with top/left areas 2.
> ui_source: templates/leftonly_2
> ui_template_layout: LEFTONLY_TOP_2, UI_CONTENT, LEFTONLY_BOTTOM 
> 
> page_title:
>         label: Page Title
>         width: 50 
> 
> members_only:
>         code: members_only
>         label: Members only?
>         options: 1=Yes,=No*
>         type: yesno 
> 
> meta_keywords:
>         code: meta_keywords
>         db: ^@
>         label: Meta keywords
>         type: text
>         width: 50 
> 
> meta_description:
>         code: meta_description
>         db: ^@
>         label: Meta description
>         type: text
>         width: 50 
> 
> [/comment]
> [set page_title][/set]
> [set members_only][/set]
> [set meta_keywords][/set]
> [set meta_description][/set]
> <!-- BEGIN PAGE_PICTURE --> 
> 
> but if I put
> [value category_name]
> into the title slot in the admin content editor, the title of the page
> is literally "[value category_name]" and is not getting interpreted as I
> want. 
> 
> Does anyone have an idea why the tag gets displyed correctly in the page
> body but not in the title?

[seti page_title][value category_name][/seti]  ? 

From peter at pajamian.dhs.org  Wed Dec  6 15:17:07 2006
From: peter at pajamian.dhs.org (Peter)
Date: Wed Dec  6 15:17:40 2006
Subject: [ic] Setting title of results page
In-Reply-To: <45772105.5040309@M-and-D.com>
References: <45772105.5040309@M-and-D.com>
Message-ID: <45772543.30305@pajamian.dhs.org>

On 12/06/2006 11:59 AM, DB wrote:
> Now in the body of the results page, the tag
> [value category_name]
> gets displayed as expected. However I also want that value to be used
> for the results page's title. In the results page's template I have:

<snip>

> [set page_title][/set]

<snip>

> but if I put
> [value category_name]
> into the title slot in the admin content editor, the title of the page
> is literally "[value category_name]" and is not getting interpreted as I
> want.
> 
> Does anyone have an idea why the tag gets displyed correctly in the page
> body but not in the title?


You need to change the [set]...[/set] tag above to [seti]...[/seti].
Then it will work.  See...

<http://www.interchange.rtfm.info/icdocs/tags/set.html>
<http://www.interchange.rtfm.info/icdocs/tags/seti.html>

Peter
From DB at M-and-D.com  Wed Dec  6 15:34:59 2006
From: DB at M-and-D.com (DB)
Date: Wed Dec  6 15:35:09 2006
Subject: [ic] Setting title of results page
References: 45772105.5040309@M-and-D.com
Message-ID: <45772973.1070503@M-and-D.com>

Thanks guys - that did the trick :)
From emailgrant at gmail.com  Sat Dec  9 13:34:26 2006
From: emailgrant at gmail.com (Grant)
Date: Sat Dec  9 13:34:45 2006
Subject: [ic] ALERT: bad pipe signal received for /page.html
Message-ID: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>

Hello, I've been plagued by apache2 segfaults ever since I started
using Interchange::Link years ago.  The latest Link.pm has ALERT
messages accompanying the segfaults in error_log:

ALERT: bad pipe signal received for /page.html
[Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
Segmentation fault (11)

Does anyone have any advice on solving this?  I'm using apache-2.0.58
and mod_perl-2.0.2 in Gentoo Linux.

- Grant
From emailgrant at gmail.com  Sat Dec  9 13:37:04 2006
From: emailgrant at gmail.com (Grant)
Date: Sat Dec  9 13:37:17 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
Message-ID: <49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>

> Hello, I've been plagued by apache2 segfaults ever since I started
> using Interchange::Link years ago.  The latest Link.pm has ALERT
> messages accompanying the segfaults in error_log:
>
> ALERT: bad pipe signal received for /page.html
> [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> Segmentation fault (11)
>
> Does anyone have any advice on solving this?  I'm using apache-2.0.58
> and mod_perl-2.0.2 in Gentoo Linux.

Also, here is the portion of Link.pm that the ALERT seems to come from:

# Return this message to the browser when the server is not running.
# Log an error log entry if set to notify

sub die_page {

    my $r = shift;
    my $msg;

    warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";

    $r->content_type ("text/html");
    $r->print (<<EOF);
<HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
<BODY BGCOLOR="#FFFFFF">
<H3>Someone pressed stop...</H3>
<P>
We have aborted this request because someone terminated it.
Please try again soon.
</BODY></HTML>
EOF

}

Please let me know if you have any ideas.

- Grant
From qck at galaxy102.co.uk  Sat Dec  9 14:25:44 2006
From: qck at galaxy102.co.uk (Roderick Quinn)
Date: Sat Dec  9 14:31:16 2006
Subject: [ic] It's almost impossible to follow a guru's advice in long
	periods of decline.
Message-ID: <002801c71bc8$8d5cb220$cfad8340@uyi>

The American IRA and Canadian RRSP are two of the last greatest tax
loopholes left.
Venture CapitalVenture Capital is a great way to light money on fire.
Thanks to Wenger and his boys.
I was researching "investment clubs" yesterday with Google; I wanted to
find active clubs to send a free copy of my book.
That's a good excuse to go to the Brazen Bean for a delicious martini.

dddd

Date : Dec 11 2006
Company : Amerossi International
S t o c k  : AMSN
Current Price : $0.0006
Projection 2 to 5 Days : $0.006-$0.008
Status : 300-500%

dddd

but hung out with the rich and famous, guzzling booze, enjoying all the
fine thing in life. It's the only financial paper I read. It's almost
impossible to follow a guru's advice in long periods of decline.
Again you must believe in your ideas and stick with them.
The stock must break-out and prove to me that it's going up.
Historically, bonds underperform the stock market.
People are comfortable with the stock market.
They often tempt investors into "adjusting" their system. With too much
risk comes almost certain loss as in the lottery and gambling where
mathimatically you cannot win in the long run.
BondsBonds and CDs offer a relatively safe investment, but the return is
low.
It doesn't happen all the time, but it's often the case that a small
number of stocks in your portfolio will make all the big gains. Click
below for more news about clubs, balls and games . I have been using it
for years and love it. Congrats to all arsenal-fans all over world.
They test and retest your mettle.
This is the same topic I talked about earlier in the week with regards
to owning X shares of ABC stock at Y price. You will pay stock based
commission on trades which might be higher than the costs of mutual
funds and so you need to do a little math and determine which is best
for you. They test and retest your mettle.
If a stock is going up over time, on a linear chart, this will begin to
look like a ski jump and offer us little visual value. Traders like
Warren Buffet an Martin Zweig would use PE since they are value
shoppers.
com      USA Today              Venturi has heart surgery, expected to
recover  Former U.
Golfernews is a member of the HarWester Network. Trading Range StopTake
any stock chart and pick a timeframe.
Just another day closer to when they tax us for the air we breath. Never
stopped going on about the workers, socialism, Lenin, Marx etc.
In the end they both had the exact same sized portfolio. Some years are
more volatile. PercentYou can pick a percent decline level for any stock
or portfolio. Will the home end of Emirates still be called the
Northbank?
Six nominees, five English, one foreign which seems a little strange
with all the overseas players in the Premiership.
Click below for more news about clubs, balls and games . I continually
reinforce this.
Golfernews is a member of the HarWester Network.
If you understand this can be very normal, you won't be alarmed when it
occurs.
I would even go as far as saying.
The big dangers in an up market is the euphoria of thinking you can do
no wrong.
With too much risk comes almost certain loss as in the lottery and
gambling where mathimatically you cannot win in the long run. Wignan did
play well in first half, but arsenal destined to grab the fourth spot. "
This is simply not true. I like to say that if stocks are arithmatic,
then futures are calculus. It was a nervous time in the second half but
once Henry had nailed his hatrick and West Ham got a second, the party
could begin. Wigan played quite well, but Thierry Henry is simply a
lethal weapon on a football pitch. It helps if an investor is mentally
ready for all types of market conditions. Hope you enjoyed it as much as
I did. Small BusinessHave you read the Millionaire Next Door? Joe Cole
and John Terry have been solid all year, and Chelsea will be champions
again but neither has pulled up any trees. Some banks have recovered
nicely. In the end they both had the exact same sized portfolio.
Index funds generally outperform the vast majority of index funds in the
long run. If Gilberto can stick with Riquelme like he did in the last
game we have a great chance but it might just come down to needing an
Henry special moment to get us through.
Traders like William O'Neil "canslim" would not.
Arsenal had to get a better result than Spurs, who had tried to get
their game with West Ham postponed at the last minute because of some
supposed food poisoning.
com      Misc sources      PGA. These are just the lowest forms of
species walking the earth; too bad they get rich for the crap they put
out there. SchwagerThe New Market WizardsJack D. SchwagerThe New Market
WizardsJack D.
Find the highest high and the lowest low for that stock in that
timeframe. I watched the game live on local TV here in Bali.

From kevin at cursor.biz  Sat Dec  9 15:33:12 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Sat Dec  9 15:33:51 2006
Subject: [ic] ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
Message-ID: <200612092033.13174.kevin@cursor.biz>

Grant <emailgrant@gmail.com> wrote:
> Hello, I've been plagued by apache2 segfaults ever since I started
> using Interchange::Link years ago.  The latest Link.pm has ALERT
> messages accompanying the segfaults in error_log:
> 
> ALERT: bad pipe signal received for /page.html
> [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> Segmentation fault (11)
> 
> Does anyone have any advice on solving this?  I'm using apache-2.0.58
> and mod_perl-2.0.2 in Gentoo Linux.
> 
I just use Apache 1.3 (currently 1.3.34-r14) and mod_interchange with
Gentoo GNU/Linux.  It works perfectly. :-)

I'll create mod_interchange-2 for Apache 2 one day - once there's a
demand for it, or I find that I need it for some reason.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From kevin at cursor.biz  Sat Dec  9 15:46:23 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Sat Dec  9 15:46:32 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
Message-ID: <200612092046.24353.kevin@cursor.biz>

Grant <emailgrant@gmail.com> wrote:
> > Hello, I've been plagued by apache2 segfaults ever since I started
> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> > messages accompanying the segfaults in error_log:
> >
> > ALERT: bad pipe signal received for /page.html
> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> > Segmentation fault (11)
> >
> > [snip]
> >
> We have aborted this request because someone terminated it.
> Please try again soon.
>
Perhaps the page request really was terminated.  Do you have any
extra-slow pages or searches etc. that users might bored waiting for?

I tend to do that myself;  If a page hasn't loaded within 10 seconds,
I'll press "stop" and either attempt a reload, or just go elsewhere.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From emailgrant at gmail.com  Sat Dec  9 16:35:22 2006
From: emailgrant at gmail.com (Grant)
Date: Sat Dec  9 16:35:50 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <200612092046.24353.kevin@cursor.biz>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<200612092046.24353.kevin@cursor.biz>
Message-ID: <49bf44f10612091335r2199a3b4wf9b50f18f22ef211@mail.gmail.com>

> > > Hello, I've been plagued by apache2 segfaults ever since I started
> > > using Interchange::Link years ago.  The latest Link.pm has ALERT
> > > messages accompanying the segfaults in error_log:
> > >
> > > ALERT: bad pipe signal received for /page.html
> > > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> > > Segmentation fault (11)
> > >
> > > [snip]
> > >
> > We have aborted this request because someone terminated it.
> > Please try again soon.
> >
> Perhaps the page request really was terminated.  Do you have any
> extra-slow pages or searches etc. that users might bored waiting for?
>
> I tend to do that myself;  If a page hasn't loaded within 10 seconds,
> I'll press "stop" and either attempt a reload, or just go elsewhere.

I would believe that, except it happens to me all the time when I'm
browsing my own site.  It looks like a 404 page in IE (via XP in
vmware) and in firefox it's just a blank page.

- Grant
From emailgrant at gmail.com  Sat Dec  9 19:26:06 2006
From: emailgrant at gmail.com (Grant)
Date: Sat Dec  9 19:26:21 2006
Subject: [ic] Updated Interchange::Link
Message-ID: <49bf44f10612091626l3d41ac8ct75f7db73022ae96d@mail.gmail.com>

If you're using Interchange::Link and you'd like:

[bounce href="http://www.domain.com" status="301"]

to return 301 instead of 302, make the following changes to your Link.pm:

- use Apache2::Const;
+ use Apache2::Const -compile => qw(DECLINED OK NOT_FOUND FORBIDDEN
REDIRECT HTTP_MOVED_PERMANENTLY);

close (SOCK) or die "close: $!\n";
+ return Apache2::Const::HTTP_MOVED_PERMANENTLY if $set_status eq '301';
return Apache2::Const::REDIRECT;

That's one removed line and two added.  It would be good if someone
checked this into CVS.

- Grant

P.S. This took me a hideously long time.
From kevin at cursor.biz  Sat Dec  9 20:02:21 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Sat Dec  9 20:02:44 2006
Subject: [ic] Updated Interchange::Link
In-Reply-To: <49bf44f10612091626l3d41ac8ct75f7db73022ae96d@mail.gmail.com>
References: <49bf44f10612091626l3d41ac8ct75f7db73022ae96d@mail.gmail.com>
Message-ID: <200612100102.22776.kevin@cursor.biz>

Grant <emailgrant@gmail.com> wrote:
> If you're using Interchange::Link and you'd like:
> 
> [bounce href="http://www.domain.com" status="301"]
> 
> to return 301 instead of 302, make the following changes to your Link.pm:
> 
> - use Apache2::Const;
> + use Apache2::Const -compile => qw(DECLINED OK NOT_FOUND FORBIDDEN
> REDIRECT HTTP_MOVED_PERMANENTLY);
> 
> close (SOCK) or die "close: $!\n";
> + return Apache2::Const::HTTP_MOVED_PERMANENTLY if $set_status eq '301';
> return Apache2::Const::REDIRECT;
> 
> That's one removed line and two added.  It would be good if someone
> checked this into CVS.
> 
No sooner said than done.  Thanks, Grant.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From bryangmyrek at yahoo.com  Sun Dec 10 12:56:13 2006
From: bryangmyrek at yahoo.com (Bryan Gmyrek)
Date: Sun Dec 10 12:56:27 2006
Subject: [ic] How to not log nsession in usertrack?
In-Reply-To: <20061129132325.3AF79830070@mailserver1.zolotek.net>
Message-ID: <556025.19655.qm@web31807.mail.mud.yahoo.com>

Hi,

I'm trying to optimize my server that IC is running on and noticed that a huge
usertrack file was slowing things down (I _think_ it was causing a 99% CPU bug
but maybe not, it's hard to tell if the 99% thing has gone away since I'm not
staring at top all the time).  In any case I tried switching to database
sessions and it seemed to work but usernames were not being created and the
referring url was not being shown in orders, etc, so I went back to using the
flat file.  
I'm planning on doing log-rotation of usertrack now; before I thought it might
be necessary for the functioning of IC somehow but just read in another post
that it isn't.  Even so it would be more optimal I think if nsession sessions
weren't logged in usertrack, saving the time of logging a line every time a bot
does something ... this is already logged by apache and is easier for me to
analyze.  
Come to think of it maybe I should turn of usertrack logging entirely.  Anyone
know how to do this and/or how to not log nsession?

Thanks!
Bryan

p.s. 
On a very loosely related note, Kevin, did anyone ever put together a
NonRobotUA directive?
http://www.icdevgroup.org/pipermail/interchange-users/2004-March/038319.html



New England Art Express - Art Prints, Posters and Framing
http://www.neartexpress.com/

-----------------------------------------------------------
---------------------Legal Notice--------------------------
-----------------------------------------------------------
This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. It may also be privileged or otherwise protected by work product immunity or other legal rules.  If you have received this email in error please notify the system manager. This message contains confidential information and is intended only for the individual named. If you are not the named addressee you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately by e-mail if you have received this e-mail by mistake and delete this e-mail from your system. If you are not the intended recipient you are notified that disclosing, copying, distributing or taking any action in reliance on the contents of this information is strictly prohibited.


 
____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com
From kevin at cursor.biz  Sun Dec 10 14:55:11 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Sun Dec 10 14:55:47 2006
Subject: [ic] How to not log nsession in usertrack?
In-Reply-To: <556025.19655.qm@web31807.mail.mud.yahoo.com>
References: <556025.19655.qm@web31807.mail.mud.yahoo.com>
Message-ID: <200612101955.12375.kevin@cursor.biz>

Bryan Gmyrek <bryangmyrek@yahoo.com> wrote:
> I'm trying to optimize my server that IC is running on and noticed that a huge
> usertrack file was slowing things down (I _think_ it was causing a 99% CPU bug
> but maybe not, it's hard to tell if the 99% thing has gone away since I'm not
> staring at top all the time).  In any case I tried switching to database
> sessions and it seemed to work but usernames were not being created and the
> referring url was not being shown in orders, etc, so I went back to using the
> flat file.  
>
> I'm planning on doing log-rotation of usertrack now; before I thought it might
> be necessary for the functioning of IC somehow but just read in another post
> that it isn't.  Even so it would be more optimal I think if nsession sessions
> weren't logged in usertrack, saving the time of logging a line every time a bot
> does something ... this is already logged by apache and is easier for me to
> analyze.  
>
> Come to think of it maybe I should turn of usertrack logging entirely.  Anyone
> know how to do this and/or how to not log nsession?
> 
Usertrack won't have any effect on performance, as it's just a file
that Interchange appends to.  If the file is big then you could
rotate it and, perhaps, delete old rotated files to save disk space.
Appending to the usertrack file will have no noticeable effect on
your CPU usage.

I imagine your problem could be something to do with session expiry.
Are you regularly deleting expired session files?  I suspect that you
are not.  A huge sessions directory will have a direct impact on your
website's performance, so it must be regularly cleaned.  Deleting old
temporary files in your ScratchDir directory (defaults to "tmp") is
also a good idea.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From bryangmyrek at yahoo.com  Mon Dec 11 00:20:06 2006
From: bryangmyrek at yahoo.com (Bryan Gmyrek)
Date: Mon Dec 11 00:20:18 2006
Subject: [ic] How to not log nsession in usertrack?
In-Reply-To: <200612101955.12375.kevin@cursor.biz>
Message-ID: <282491.96006.qm@web31813.mail.mud.yahoo.com>


> >
> > Come to think of it maybe I should turn of usertrack logging entirely. 
> Anyone
> > know how to do this and/or how to not log nsession?
> > 
> Usertrack won't have any effect on performance, as it's just a file
> that Interchange appends to.  If the file is big then you could
OK, I'm not a file-access guru but couldn't that affect performance if the
usertrack file is on the order of a GB?

> rotate it and, perhaps, delete old rotated files to save disk space.
Will do this.

> Appending to the usertrack file will have no noticeable effect on
> your CPU usage.
> 
> I imagine your problem could be something to do with session expiry.
> Are you regularly deleting expired session files?  I suspect that you
> are not.  A huge sessions directory will have a direct impact on your
> website's performance, so it must be regularly cleaned.  Deleting old
> temporary files in your ScratchDir directory (defaults to "tmp") is
> also a good idea.
> 
I do both of these nightly in a cron job, so you suspect wrong ;)  But thanks
for the tips.  I was only not rotating usertrack since I thought perhaps it was
used in some of the reporting and shouldn't be rotated.

So... it isn't possible to turn off logging to usertrack?  Any tips on where I
could look in the code to comment out the line/routine that does the logging?

Thanks,
Bryan

New England Art Express - Art Prints, Posters and Framing
http://www.neartexpress.com/


 
____________________________________________________________________________________
Do you Yahoo!?
Everyone is raving about the all-new Yahoo! Mail beta.
http://new.mail.yahoo.com
From paul at gishnetwork.com  Mon Dec 11 01:18:31 2006
From: paul at gishnetwork.com (Paul Jordan)
Date: Mon Dec 11 01:18:44 2006
Subject: [ic] How to not log nsession in usertrack?
In-Reply-To: <282491.96006.qm@web31813.mail.mud.yahoo.com>
Message-ID: <056301c71cec$2e347d80$0d01a8c0@workstation>

interchange-users-bounces@icdevgroup.org wrote:
[snip]
>> I imagine your problem could be something to do with session expiry.
>> Are you regularly deleting expired session files?  I suspect that you
>> are not.  A huge sessions directory will have a direct impact on your
>> website's performance, so it must be regularly cleaned. Deleting old
>> temporary files in your ScratchDir directory (defaults to "tmp") is
>> also a good idea. 
>> 
> I do both of these nightly in a cron job, so you suspect
> wrong ;)  But thanks for the tips.  I was only not rotating
> usertrack since I thought perhaps it was used in some of the
> reporting and shouldn't be rotated.
> 
> So... it isn't possible to turn off logging to usertrack?
> Any tips on where I could look in the code to comment out the
> line/routine that does the logging?


Have you tried commenting out TrackFile?

http://www.interchange.rtfm.info/icdocs/config/TrackFile.html

# TrackFile  logs/usertrack  


Paul Jordan


From peter at pajamian.dhs.org  Mon Dec 11 01:31:58 2006
From: peter at pajamian.dhs.org (Peter)
Date: Mon Dec 11 01:32:27 2006
Subject: [ic] How to not log nsession in usertrack?
In-Reply-To: <282491.96006.qm@web31813.mail.mud.yahoo.com>
References: <282491.96006.qm@web31813.mail.mud.yahoo.com>
Message-ID: <457CFB5E.8090309@pajamian.dhs.org>

On 12/10/2006 09:20 PM, Bryan Gmyrek wrote:
> I was only not rotating usertrack since I thought perhaps it was
> used in some of the reporting and shouldn't be rotated.

It's used for the affiliate statistics report and probably a few other
reports.  It's not used for order or sales reports.

Peter

From kevin at cursor.biz  Mon Dec 11 07:20:45 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Mon Dec 11 07:21:26 2006
Subject: [ic] How to not log nsession in usertrack?
In-Reply-To: <056301c71cec$2e347d80$0d01a8c0@workstation>
References: <056301c71cec$2e347d80$0d01a8c0@workstation>
Message-ID: <200612111220.46240.kevin@cursor.biz>

"Paul Jordan" <paul@gishnetwork.com> wrote:
> interchange-users-bounces@icdevgroup.org wrote:
> > > I imagine your problem could be something to do with session expiry.
> > > Are you regularly deleting expired session files?  I suspect that you
> > > are not.  A huge sessions directory will have a direct impact on your
> > > website's performance, so it must be regularly cleaned. Deleting old
> > > temporary files in your ScratchDir directory (defaults to "tmp") is
> > > also a good idea. 
> > > 
> > I do both of these nightly in a cron job, so you suspect
> > wrong ;)  But thanks for the tips.  I was only not rotating
> > usertrack since I thought perhaps it was used in some of the
> > reporting and shouldn't be rotated.
> > 
> > So... it isn't possible to turn off logging to usertrack?
> > Any tips on where I could look in the code to comment out the
> > line/routine that does the logging?
> > 
> Have you tried commenting out TrackFile?
> 
> http://www.interchange.rtfm.info/icdocs/config/TrackFile.html
> 
> # TrackFile  logs/usertrack  
> 
That's the easiest way to check, yes.  I still think the problem will
be related to the session files, rather than the usertrack, but we'll
find out soon enough. :-)

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From emailgrant at gmail.com  Mon Dec 11 13:02:42 2006
From: emailgrant at gmail.com (Grant)
Date: Mon Dec 11 13:03:33 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
Message-ID: <49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>

> > Hello, I've been plagued by apache2 segfaults ever since I started
> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> > messages accompanying the segfaults in error_log:
> >
> > ALERT: bad pipe signal received for /page.html
> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> > Segmentation fault (11)
> >
> > Does anyone have any advice on solving this?  I'm using apache-2.0.58
> > and mod_perl-2.0.2 in Gentoo Linux.
>
> Also, here is the portion of Link.pm that the ALERT seems to come from:
>
> # Return this message to the browser when the server is not running.
> # Log an error log entry if set to notify
>
> sub die_page {
>
>     my $r = shift;
>     my $msg;
>
>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
>
>     $r->content_type ("text/html");
>     $r->print (<<EOF);
> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> <BODY BGCOLOR="#FFFFFF">
> <H3>Someone pressed stop...</H3>
> <P>
> We have aborted this request because someone terminated it.
> Please try again soon.
> </BODY></HTML>
> EOF
>
> }
>
> Please let me know if you have any ideas.

The segfaults are eliminated by commenting out the $r stuff in the
die_page sub.  I still get the ALERTs though.  Does anyone have any
advice on figuring out why I'm having the bad pipe problem?  Is there
an easy way to add extra debugging info to the sub?

Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs 50 fold.

- Grant
From josh at myprivacy.ca  Mon Dec 11 13:30:20 2006
From: josh at myprivacy.ca (Josh Lavin)
Date: Mon Dec 11 13:31:04 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
Message-ID: <183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>

On Dec 11, 2006, at 12:02 PM, Grant wrote:

>> > Hello, I've been plagued by apache2 segfaults ever since I started
>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>> > messages accompanying the segfaults in error_log:
>> >
>> > ALERT: bad pipe signal received for /page.html
>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>> > Segmentation fault (11)
>> >
>> > Does anyone have any advice on solving this?  I'm using  
>> apache-2.0.58
>> > and mod_perl-2.0.2 in Gentoo Linux.
>>
>> Also, here is the portion of Link.pm that the ALERT seems to come  
>> from:
>>
>> # Return this message to the browser when the server is not running.
>> # Log an error log entry if set to notify
>>
>> sub die_page {
>>
>>     my $r = shift;
>>     my $msg;
>>
>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
>>
>>     $r->content_type ("text/html");
>>     $r->print (<<EOF);
>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>> <BODY BGCOLOR="#FFFFFF">
>> <H3>Someone pressed stop...</H3>
>> <P>
>> We have aborted this request because someone terminated it.
>> Please try again soon.
>> </BODY></HTML>
>> EOF
>>
>> }
>>
>> Please let me know if you have any ideas.
>
> The segfaults are eliminated by commenting out the $r stuff in the
> die_page sub.  I still get the ALERTs though.  Does anyone have any
> advice on figuring out why I'm having the bad pipe problem?  Is there
> an easy way to add extra debugging info to the sub?
>
> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs  
> 50 fold.

I've been seeing this too, on my Apache 2 and latest Link.pm. I also  
had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.

The visible effect on the browser is that the page or image (which  
Link.pm apparently still has some part in delivering) does not load.  
I get them myself when browsing and testing my websites, and I have  
never stopped loading a page or had any other problems on non-IC  
sites I host.

I was told the problem stems from either the browser and a stop  
button or some other network fault. I may go back to Apache 1.3 to  
get around this.

--
Josh Lavin
Kingdom Design   http://www.kingdomdesign.com/

From ron at endpoint.com  Mon Dec 11 13:42:13 2006
From: ron at endpoint.com (Ron Phipps)
Date: Mon Dec 11 13:43:10 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
Message-ID: <457DA685.9000201@endpoint.com>

Josh Lavin wrote:
> On Dec 11, 2006, at 12:02 PM, Grant wrote:
> 
>>> > Hello, I've been plagued by apache2 segfaults ever since I started
>>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>>> > messages accompanying the segfaults in error_log:
>>> >
>>> > ALERT: bad pipe signal received for /page.html
>>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>>> > Segmentation fault (11)
>>> >
>>> > Does anyone have any advice on solving this?  I'm using apache-2.0.58
>>> > and mod_perl-2.0.2 in Gentoo Linux.
>>>
>>> Also, here is the portion of Link.pm that the ALERT seems to come from:
>>>
>>> # Return this message to the browser when the server is not running.
>>> # Log an error log entry if set to notify
>>>
>>> sub die_page {
>>>
>>>     my $r = shift;
>>>     my $msg;
>>>
>>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
>>>
>>>     $r->content_type ("text/html");
>>>     $r->print (<<EOF);
>>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>>> <BODY BGCOLOR="#FFFFFF">
>>> <H3>Someone pressed stop...</H3>
>>> <P>
>>> We have aborted this request because someone terminated it.
>>> Please try again soon.
>>> </BODY></HTML>
>>> EOF
>>>
>>> }
>>>
>>> Please let me know if you have any ideas.
>>
>> The segfaults are eliminated by commenting out the $r stuff in the
>> die_page sub.  I still get the ALERTs though.  Does anyone have any
>> advice on figuring out why I'm having the bad pipe problem?  Is there
>> an easy way to add extra debugging info to the sub?
>>
>> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs 50 
>> fold.
> 
> I've been seeing this too, on my Apache 2 and latest Link.pm. I also had 
> to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> 
> The visible effect on the browser is that the page or image (which 
> Link.pm apparently still has some part in delivering) does not load. I 
> get them myself when browsing and testing my websites, and I have never 
> stopped loading a page or had any other problems on non-IC sites I host.
> 
> I was told the problem stems from either the browser and a stop button 
> or some other network fault. I may go back to Apache 1.3 to get around 
> this.
> 

I saw this occur on two different installations about 4 months ago.  It was suggested that I abandon the use of Link.pm and go back to using the cgi method with URL rewrite rules as this was just as fast and proved stable over the years.

-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From emailgrant at gmail.com  Mon Dec 11 13:59:46 2006
From: emailgrant at gmail.com (Grant)
Date: Mon Dec 11 14:00:11 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
Message-ID: <49bf44f10612111059wc2a5032p3ff44eb5f1ebb3c9@mail.gmail.com>

> >> > Hello, I've been plagued by apache2 segfaults ever since I started
> >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> >> > messages accompanying the segfaults in error_log:
> >> >
> >> > ALERT: bad pipe signal received for /page.html
> >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> >> > Segmentation fault (11)
> >> >
> >> > Does anyone have any advice on solving this?  I'm using
> >> apache-2.0.58
> >> > and mod_perl-2.0.2 in Gentoo Linux.
> >>
> >> Also, here is the portion of Link.pm that the ALERT seems to come
> >> from:
> >>
> >> # Return this message to the browser when the server is not running.
> >> # Log an error log entry if set to notify
> >>
> >> sub die_page {
> >>
> >>     my $r = shift;
> >>     my $msg;
> >>
> >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> >>
> >>     $r->content_type ("text/html");
> >>     $r->print (<<EOF);
> >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> >> <BODY BGCOLOR="#FFFFFF">
> >> <H3>Someone pressed stop...</H3>
> >> <P>
> >> We have aborted this request because someone terminated it.
> >> Please try again soon.
> >> </BODY></HTML>
> >> EOF
> >>
> >> }
> >>
> >> Please let me know if you have any ideas.
> >
> > The segfaults are eliminated by commenting out the $r stuff in the
> > die_page sub.  I still get the ALERTs though.  Does anyone have any
> > advice on figuring out why I'm having the bad pipe problem?  Is there
> > an easy way to add extra debugging info to the sub?
> >
> > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
> > 50 fold.
>
> I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>
> The visible effect on the browser is that the page or image (which
> Link.pm apparently still has some part in delivering) does not load.
> I get them myself when browsing and testing my websites, and I have
> never stopped loading a page or had any other problems on non-IC
> sites I host.
>
> I was told the problem stems from either the browser and a stop
> button or some other network fault. I may go back to Apache 1.3 to
> get around this.

Sounds like the exact same problem I'm having.

I don't use the UI so I don't have an interchange-5/ folder.  If I
request an admin page, the html loads and my error_log is suddenly
filled with errors like this:

ALERT: bad pipe signal received for /interchange-5

It happens every single time.  After that, bad pipe alerts appear much
more frequently until apache2 is restarted.  I wonder if this is a
clue.

- Grant

> Josh Lavin
From emailgrant at gmail.com  Mon Dec 11 14:01:55 2006
From: emailgrant at gmail.com (Grant)
Date: Mon Dec 11 14:02:19 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <457DA685.9000201@endpoint.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<457DA685.9000201@endpoint.com>
Message-ID: <49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>

> >>> > Hello, I've been plagued by apache2 segfaults ever since I started
> >>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> >>> > messages accompanying the segfaults in error_log:
> >>> >
> >>> > ALERT: bad pipe signal received for /page.html
> >>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> >>> > Segmentation fault (11)
> >>> >
> >>> > Does anyone have any advice on solving this?  I'm using apache-2.0.58
> >>> > and mod_perl-2.0.2 in Gentoo Linux.
> >>>
> >>> Also, here is the portion of Link.pm that the ALERT seems to come from:
> >>>
> >>> # Return this message to the browser when the server is not running.
> >>> # Log an error log entry if set to notify
> >>>
> >>> sub die_page {
> >>>
> >>>     my $r = shift;
> >>>     my $msg;
> >>>
> >>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> >>>
> >>>     $r->content_type ("text/html");
> >>>     $r->print (<<EOF);
> >>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> >>> <BODY BGCOLOR="#FFFFFF">
> >>> <H3>Someone pressed stop...</H3>
> >>> <P>
> >>> We have aborted this request because someone terminated it.
> >>> Please try again soon.
> >>> </BODY></HTML>
> >>> EOF
> >>>
> >>> }
> >>>
> >>> Please let me know if you have any ideas.
> >>
> >> The segfaults are eliminated by commenting out the $r stuff in the
> >> die_page sub.  I still get the ALERTs though.  Does anyone have any
> >> advice on figuring out why I'm having the bad pipe problem?  Is there
> >> an easy way to add extra debugging info to the sub?
> >>
> >> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs 50
> >> fold.
> >
> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also had
> > to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> >
> > The visible effect on the browser is that the page or image (which
> > Link.pm apparently still has some part in delivering) does not load. I
> > get them myself when browsing and testing my websites, and I have never
> > stopped loading a page or had any other problems on non-IC sites I host.
> >
> > I was told the problem stems from either the browser and a stop button
> > or some other network fault. I may go back to Apache 1.3 to get around
> > this.
> >
>
> I saw this occur on two different installations about 4 months ago.  It was suggested that I abandon the use of Link.pm and go back to using the cgi method with URL rewrite rules as this was just as fast and proved stable over the years.

The problem with the cgi method is it requires your pages to have a
URL that includes the catalog name.

- Grant
From jon at endpoint.com  Mon Dec 11 14:13:34 2006
From: jon at endpoint.com (Jon Jensen)
Date: Mon Dec 11 14:14:09 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<457DA685.9000201@endpoint.com>
	<49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.0612111213100.10153@ybpnyubfg.ybpnyqbznva>

On Mon, 11 Dec 2006, Grant wrote:

>> I saw this occur on two different installations about 4 months ago. 
>> It was suggested that I abandon the use of Link.pm and go back to using 
>> the cgi method with URL rewrite rules as this was just as fast and 
>> proved stable over the years.
>
> The problem with the cgi method is it requires your pages to have a URL 
> that includes the catalog name.

mod_rewrite makes it easy to have URLs that look like you want.

Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From racke at linuxia.de  Mon Dec 11 14:28:12 2006
From: racke at linuxia.de (Stefan Hornburg (Racke))
Date: Mon Dec 11 14:28:51 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>	<457DA685.9000201@endpoint.com>
	<49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>
Message-ID: <457DB14C.2030206@linuxia.de>

Grant wrote:
>> >>> > Hello, I've been plagued by apache2 segfaults ever since I started
>> >>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>> >>> > messages accompanying the segfaults in error_log:
>> >>> >
>> >>> > ALERT: bad pipe signal received for /page.html
>> >>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>> >>> > Segmentation fault (11)
>> >>> >
>> >>> > Does anyone have any advice on solving this?  I'm using 
>> apache-2.0.58
>> >>> > and mod_perl-2.0.2 in Gentoo Linux.
>> >>>
>> >>> Also, here is the portion of Link.pm that the ALERT seems to come 
>> from:
>> >>>
>> >>> # Return this message to the browser when the server is not running.
>> >>> # Log an error log entry if set to notify
>> >>>
>> >>> sub die_page {
>> >>>
>> >>>     my $r = shift;
>> >>>     my $msg;
>> >>>
>> >>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
>> >>>
>> >>>     $r->content_type ("text/html");
>> >>>     $r->print (<<EOF);
>> >>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>> >>> <BODY BGCOLOR="#FFFFFF">
>> >>> <H3>Someone pressed stop...</H3>
>> >>> <P>
>> >>> We have aborted this request because someone terminated it.
>> >>> Please try again soon.
>> >>> </BODY></HTML>
>> >>> EOF
>> >>>
>> >>> }
>> >>>
>> >>> Please let me know if you have any ideas.
>> >>
>> >> The segfaults are eliminated by commenting out the $r stuff in the
>> >> die_page sub.  I still get the ALERTs though.  Does anyone have any
>> >> advice on figuring out why I'm having the bad pipe problem?  Is there
>> >> an easy way to add extra debugging info to the sub?
>> >>
>> >> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs 50
>> >> fold.
>> >
>> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also 
>> had
>> > to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>> >
>> > The visible effect on the browser is that the page or image (which
>> > Link.pm apparently still has some part in delivering) does not load. I
>> > get them myself when browsing and testing my websites, and I have never
>> > stopped loading a page or had any other problems on non-IC sites I 
>> host.
>> >
>> > I was told the problem stems from either the browser and a stop button
>> > or some other network fault. I may go back to Apache 1.3 to get around
>> > this.
>> >
>>
>> I saw this occur on two different installations about 4 months ago.  
>> It was suggested that I abandon the use of Link.pm and go back to 
>> using the cgi method with URL rewrite rules as this was just as fast 
>> and proved stable over the years.
> 
> The problem with the cgi method is it requires your pages to have a
> URL that includes the catalog name.

That is simply not true.

Bye
	Racke

-- 
LinuXia Systems => http://www.linuxia.de/
Expert Interchange Consulting and System Administration
ICDEVGROUP => http://www.icdevgroup.org/
Interchange Development Team

From emailgrant at gmail.com  Mon Dec 11 14:48:00 2006
From: emailgrant at gmail.com (Grant)
Date: Mon Dec 11 14:48:11 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <457DB14C.2030206@linuxia.de>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<457DA685.9000201@endpoint.com>
	<49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>
	<457DB14C.2030206@linuxia.de>
Message-ID: <49bf44f10612111148g47de91edof2e5eb361d9ed4ea@mail.gmail.com>

> >> >>> > Hello, I've been plagued by apache2 segfaults ever since I started
> >> >>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> >> >>> > messages accompanying the segfaults in error_log:
> >> >>> >
> >> >>> > ALERT: bad pipe signal received for /page.html
> >> >>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> >> >>> > Segmentation fault (11)
> >> >>> >
> >> >>> > Does anyone have any advice on solving this?  I'm using
> >> apache-2.0.58
> >> >>> > and mod_perl-2.0.2 in Gentoo Linux.
> >> >>>
> >> >>> Also, here is the portion of Link.pm that the ALERT seems to come
> >> from:
> >> >>>
> >> >>> # Return this message to the browser when the server is not running.
> >> >>> # Log an error log entry if set to notify
> >> >>>
> >> >>> sub die_page {
> >> >>>
> >> >>>     my $r = shift;
> >> >>>     my $msg;
> >> >>>
> >> >>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> >> >>>
> >> >>>     $r->content_type ("text/html");
> >> >>>     $r->print (<<EOF);
> >> >>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> >> >>> <BODY BGCOLOR="#FFFFFF">
> >> >>> <H3>Someone pressed stop...</H3>
> >> >>> <P>
> >> >>> We have aborted this request because someone terminated it.
> >> >>> Please try again soon.
> >> >>> </BODY></HTML>
> >> >>> EOF
> >> >>>
> >> >>> }
> >> >>>
> >> >>> Please let me know if you have any ideas.
> >> >>
> >> >> The segfaults are eliminated by commenting out the $r stuff in the
> >> >> die_page sub.  I still get the ALERTs though.  Does anyone have any
> >> >> advice on figuring out why I'm having the bad pipe problem?  Is there
> >> >> an easy way to add extra debugging info to the sub?
> >> >>
> >> >> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs 50
> >> >> fold.
> >> >
> >> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> >> had
> >> > to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> >> >
> >> > The visible effect on the browser is that the page or image (which
> >> > Link.pm apparently still has some part in delivering) does not load. I
> >> > get them myself when browsing and testing my websites, and I have never
> >> > stopped loading a page or had any other problems on non-IC sites I
> >> host.
> >> >
> >> > I was told the problem stems from either the browser and a stop button
> >> > or some other network fault. I may go back to Apache 1.3 to get around
> >> > this.
> >> >
> >>
> >> I saw this occur on two different installations about 4 months ago.
> >> It was suggested that I abandon the use of Link.pm and go back to
> >> using the cgi method with URL rewrite rules as this was just as fast
> >> and proved stable over the years.
> >
> > The problem with the cgi method is it requires your pages to have a
> > URL that includes the catalog name.
>
> That is simply not true.
>
> Bye
>         Racke

Are you referring to mod_rewrite?  mod_rewrite will only change what
appears in the address bar right?  I need something that allows IC to
generate folder-free URLs.  Can IC be configured to generate and use
such URLs via the cgi method?

- Grant
From racke at linuxia.de  Mon Dec 11 15:11:09 2006
From: racke at linuxia.de (Stefan Hornburg (Racke))
Date: Mon Dec 11 15:11:42 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612111148g47de91edof2e5eb361d9ed4ea@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>	<457DA685.9000201@endpoint.com>	<49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>	<457DB14C.2030206@linuxia.de>
	<49bf44f10612111148g47de91edof2e5eb361d9ed4ea@mail.gmail.com>
Message-ID: <457DBB5D.7070901@linuxia.de>

Grant wrote:
>> >> >>> > Hello, I've been plagued by apache2 segfaults ever since I 
>> started
>> >> >>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>> >> >>> > messages accompanying the segfaults in error_log:
>> >> >>> >
>> >> >>> > ALERT: bad pipe signal received for /page.html
>> >> >>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>> >> >>> > Segmentation fault (11)
>> >> >>> >
>> >> >>> > Does anyone have any advice on solving this?  I'm using
>> >> apache-2.0.58
>> >> >>> > and mod_perl-2.0.2 in Gentoo Linux.
>> >> >>>
>> >> >>> Also, here is the portion of Link.pm that the ALERT seems to come
>> >> from:
>> >> >>>
>> >> >>> # Return this message to the browser when the server is not 
>> running.
>> >> >>> # Log an error log entry if set to notify
>> >> >>>
>> >> >>> sub die_page {
>> >> >>>
>> >> >>>     my $r = shift;
>> >> >>>     my $msg;
>> >> >>>
>> >> >>>     warn "ALERT: bad pipe signal received for 
>> $ENV{SCRIPT_NAME}\n";
>> >> >>>
>> >> >>>     $r->content_type ("text/html");
>> >> >>>     $r->print (<<EOF);
>> >> >>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>> >> >>> <BODY BGCOLOR="#FFFFFF">
>> >> >>> <H3>Someone pressed stop...</H3>
>> >> >>> <P>
>> >> >>> We have aborted this request because someone terminated it.
>> >> >>> Please try again soon.
>> >> >>> </BODY></HTML>
>> >> >>> EOF
>> >> >>>
>> >> >>> }
>> >> >>>
>> >> >>> Please let me know if you have any ideas.
>> >> >>
>> >> >> The segfaults are eliminated by commenting out the $r stuff in the
>> >> >> die_page sub.  I still get the ALERTs though.  Does anyone have any
>> >> >> advice on figuring out why I'm having the bad pipe problem?  Is 
>> there
>> >> >> an easy way to add extra debugging info to the sub?
>> >> >>
>> >> >> Also, restarting IC with PERL_SIGNALS=unsafe increases the 
>> ALERTs 50
>> >> >> fold.
>> >> >
>> >> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also
>> >> had
>> >> > to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>> >> >
>> >> > The visible effect on the browser is that the page or image (which
>> >> > Link.pm apparently still has some part in delivering) does not 
>> load. I
>> >> > get them myself when browsing and testing my websites, and I have 
>> never
>> >> > stopped loading a page or had any other problems on non-IC sites I
>> >> host.
>> >> >
>> >> > I was told the problem stems from either the browser and a stop 
>> button
>> >> > or some other network fault. I may go back to Apache 1.3 to get 
>> around
>> >> > this.
>> >> >
>> >>
>> >> I saw this occur on two different installations about 4 months ago.
>> >> It was suggested that I abandon the use of Link.pm and go back to
>> >> using the cgi method with URL rewrite rules as this was just as fast
>> >> and proved stable over the years.
>> >
>> > The problem with the cgi method is it requires your pages to have a
>> > URL that includes the catalog name.
>>
>> That is simply not true.
>>
>> Bye
>>         Racke
> 
> Are you referring to mod_rewrite?  mod_rewrite will only change what
> appears in the address bar right?  I need something that allows IC to
> generate folder-free URLs.  Can IC be configured to generate and use
> such URLs via the cgi method?

Look into the FullURL directive. This allows you to specify your Catalog like that:

Catalog linuxia /home/racke/linuxia www.linuxia.de www.linuxia.de:443

Now you have only to ensure that the calls from the clients are passed on to
IC, e.g.:

Alias /images/ /home/racke/linuxia/html/images/
ScriptAliasMatch ^/(.*)$ /usr/lib/cgi-bin/ic/linuxia/$1

You can also use rewrite do achieve that.

Bye
	Racke

-- 
LinuXia Systems => http://www.linuxia.de/
Expert Interchange Consulting and System Administration
ICDEVGROUP => http://www.icdevgroup.org/
Interchange Development Team

From emailgrant at gmail.com  Mon Dec 11 18:14:02 2006
From: emailgrant at gmail.com (Grant)
Date: Mon Dec 11 18:14:15 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612111059wc2a5032p3ff44eb5f1ebb3c9@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<49bf44f10612111059wc2a5032p3ff44eb5f1ebb3c9@mail.gmail.com>
Message-ID: <49bf44f10612111514n43bdd241m26270f51302f4102@mail.gmail.com>

> > >> > Hello, I've been plagued by apache2 segfaults ever since I started
> > >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> > >> > messages accompanying the segfaults in error_log:
> > >> >
> > >> > ALERT: bad pipe signal received for /page.html
> > >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> > >> > Segmentation fault (11)
> > >> >
> > >> > Does anyone have any advice on solving this?  I'm using
> > >> apache-2.0.58
> > >> > and mod_perl-2.0.2 in Gentoo Linux.
> > >>
> > >> Also, here is the portion of Link.pm that the ALERT seems to come
> > >> from:
> > >>
> > >> # Return this message to the browser when the server is not running.
> > >> # Log an error log entry if set to notify
> > >>
> > >> sub die_page {
> > >>
> > >>     my $r = shift;
> > >>     my $msg;
> > >>
> > >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> > >>
> > >>     $r->content_type ("text/html");
> > >>     $r->print (<<EOF);
> > >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> > >> <BODY BGCOLOR="#FFFFFF">
> > >> <H3>Someone pressed stop...</H3>
> > >> <P>
> > >> We have aborted this request because someone terminated it.
> > >> Please try again soon.
> > >> </BODY></HTML>
> > >> EOF
> > >>
> > >> }
> > >>
> > >> Please let me know if you have any ideas.
> > >
> > > The segfaults are eliminated by commenting out the $r stuff in the
> > > die_page sub.  I still get the ALERTs though.  Does anyone have any
> > > advice on figuring out why I'm having the bad pipe problem?  Is there
> > > an easy way to add extra debugging info to the sub?
> > >
> > > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
> > > 50 fold.
> >
> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> > had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> >
> > The visible effect on the browser is that the page or image (which
> > Link.pm apparently still has some part in delivering) does not load.
> > I get them myself when browsing and testing my websites, and I have
> > never stopped loading a page or had any other problems on non-IC
> > sites I host.
> >
> > I was told the problem stems from either the browser and a stop
> > button or some other network fault. I may go back to Apache 1.3 to
> > get around this.
>
> Sounds like the exact same problem I'm having.
>
> I don't use the UI so I don't have an interchange-5/ folder.  If I
> request an admin page, the html loads and my error_log is suddenly
> filled with errors like this:
>
> ALERT: bad pipe signal received for /interchange-5
>
> It happens every single time.  After that, bad pipe alerts appear much
> more frequently until apache2 is restarted.  I wonder if this is a
> clue.

I want to refine this a little by saying that requesting any page over
and over rapidly via F5 leads to a bunch of ALERTs.  At that point,
ALERTs happen much more frequently until apache2 is restarted.

There is some kind of a cumulative effect here that I don't
understand.  Any ideas?  Can anyone verify this particular behavior?

- Grant
From emailgrant at gmail.com  Mon Dec 11 18:16:37 2006
From: emailgrant at gmail.com (Grant)
Date: Mon Dec 11 18:16:49 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <457DBB5D.7070901@linuxia.de>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<457DA685.9000201@endpoint.com>
	<49bf44f10612111101u8d2257dydb75e717cd1b47cc@mail.gmail.com>
	<457DB14C.2030206@linuxia.de>
	<49bf44f10612111148g47de91edof2e5eb361d9ed4ea@mail.gmail.com>
	<457DBB5D.7070901@linuxia.de>
Message-ID: <49bf44f10612111516o7e930932o5d32d4fb2c31326d@mail.gmail.com>

> >> >> >>> > Hello, I've been plagued by apache2 segfaults ever since I
> >> started
> >> >> >>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> >> >> >>> > messages accompanying the segfaults in error_log:
> >> >> >>> >
> >> >> >>> > ALERT: bad pipe signal received for /page.html
> >> >> >>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> >> >> >>> > Segmentation fault (11)
> >> >> >>> >
> >> >> >>> > Does anyone have any advice on solving this?  I'm using
> >> >> apache-2.0.58
> >> >> >>> > and mod_perl-2.0.2 in Gentoo Linux.
> >> >> >>>
> >> >> >>> Also, here is the portion of Link.pm that the ALERT seems to come
> >> >> from:
> >> >> >>>
> >> >> >>> # Return this message to the browser when the server is not
> >> running.
> >> >> >>> # Log an error log entry if set to notify
> >> >> >>>
> >> >> >>> sub die_page {
> >> >> >>>
> >> >> >>>     my $r = shift;
> >> >> >>>     my $msg;
> >> >> >>>
> >> >> >>>     warn "ALERT: bad pipe signal received for
> >> $ENV{SCRIPT_NAME}\n";
> >> >> >>>
> >> >> >>>     $r->content_type ("text/html");
> >> >> >>>     $r->print (<<EOF);
> >> >> >>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> >> >> >>> <BODY BGCOLOR="#FFFFFF">
> >> >> >>> <H3>Someone pressed stop...</H3>
> >> >> >>> <P>
> >> >> >>> We have aborted this request because someone terminated it.
> >> >> >>> Please try again soon.
> >> >> >>> </BODY></HTML>
> >> >> >>> EOF
> >> >> >>>
> >> >> >>> }
> >> >> >>>
> >> >> >>> Please let me know if you have any ideas.
> >> >> >>
> >> >> >> The segfaults are eliminated by commenting out the $r stuff in the
> >> >> >> die_page sub.  I still get the ALERTs though.  Does anyone have any
> >> >> >> advice on figuring out why I'm having the bad pipe problem?  Is
> >> there
> >> >> >> an easy way to add extra debugging info to the sub?
> >> >> >>
> >> >> >> Also, restarting IC with PERL_SIGNALS=unsafe increases the
> >> ALERTs 50
> >> >> >> fold.
> >> >> >
> >> >> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> >> >> had
> >> >> > to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> >> >> >
> >> >> > The visible effect on the browser is that the page or image (which
> >> >> > Link.pm apparently still has some part in delivering) does not
> >> load. I
> >> >> > get them myself when browsing and testing my websites, and I have
> >> never
> >> >> > stopped loading a page or had any other problems on non-IC sites I
> >> >> host.
> >> >> >
> >> >> > I was told the problem stems from either the browser and a stop
> >> button
> >> >> > or some other network fault. I may go back to Apache 1.3 to get
> >> around
> >> >> > this.
> >> >> >
> >> >>
> >> >> I saw this occur on two different installations about 4 months ago.
> >> >> It was suggested that I abandon the use of Link.pm and go back to
> >> >> using the cgi method with URL rewrite rules as this was just as fast
> >> >> and proved stable over the years.
> >> >
> >> > The problem with the cgi method is it requires your pages to have a
> >> > URL that includes the catalog name.
> >>
> >> That is simply not true.
> >>
> >> Bye
> >>         Racke
> >
> > Are you referring to mod_rewrite?  mod_rewrite will only change what
> > appears in the address bar right?  I need something that allows IC to
> > generate folder-free URLs.  Can IC be configured to generate and use
> > such URLs via the cgi method?
>
> Look into the FullURL directive. This allows you to specify your Catalog like that:
>
> Catalog linuxia /home/racke/linuxia www.linuxia.de www.linuxia.de:443
>
> Now you have only to ensure that the calls from the clients are passed on to
> IC, e.g.:
>
> Alias /images/ /home/racke/linuxia/html/images/
> ScriptAliasMatch ^/(.*)$ /usr/lib/cgi-bin/ic/linuxia/$1
>
> You can also use rewrite do achieve that.

Thank you Racke, I'll check that out.  Interchange::Link doesn't seem
to be curable.  Maybe it's a limitation of mod_perl?

- Grant
From josh at myprivacy.ca  Mon Dec 11 18:24:17 2006
From: josh at myprivacy.ca (Josh Lavin)
Date: Mon Dec 11 18:24:50 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <457DA685.9000201@endpoint.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<457DA685.9000201@endpoint.com>
Message-ID: <EC1A7BEC-D5BB-4717-92C0-2756784119ED@myprivacy.ca>

On Dec 11, 2006, at 12:42 PM, Ron Phipps wrote:
> Josh Lavin wrote:
>> On Dec 11, 2006, at 12:02 PM, Grant wrote:
>>>> > Hello, I've been plagued by apache2 segfaults ever since I  
>>>> started
>>>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>>>> > messages accompanying the segfaults in error_log:
>>>> >
>>>> > ALERT: bad pipe signal received for /page.html
>>>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>>>> > Segmentation fault (11)
>>>> >
>>>> > Does anyone have any advice on solving this?  I'm using  
>>>> apache-2.0.58
>>>> > and mod_perl-2.0.2 in Gentoo Linux.
>>>>
>>>> Also, here is the portion of Link.pm that the ALERT seems to  
>>>> come from:
>>>>
>>>> # Return this message to the browser when the server is not  
>>>> running.
>>>> # Log an error log entry if set to notify
>>>>
>>>> sub die_page {
>>>>
>>>>     my $r = shift;
>>>>     my $msg;
>>>>
>>>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
>>>>
>>>>     $r->content_type ("text/html");
>>>>     $r->print (<<EOF);
>>>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>>>> <BODY BGCOLOR="#FFFFFF">
>>>> <H3>Someone pressed stop...</H3>
>>>> <P>
>>>> We have aborted this request because someone terminated it.
>>>> Please try again soon.
>>>> </BODY></HTML>
>>>> EOF
>>>>
>>>> }
>>>>
>>>> Please let me know if you have any ideas.
>>>
>>> The segfaults are eliminated by commenting out the $r stuff in the
>>> die_page sub.  I still get the ALERTs though.  Does anyone have any
>>> advice on figuring out why I'm having the bad pipe problem?  Is  
>>> there
>>> an easy way to add extra debugging info to the sub?
>>>
>>> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs  
>>> 50 fold.
>> I've been seeing this too, on my Apache 2 and latest Link.pm. I  
>> also had to use PERL_SIGNALS=unsafe and so I get quite a lot of  
>> these.
>> The visible effect on the browser is that the page or image (which  
>> Link.pm apparently still has some part in delivering) does not  
>> load. I get them myself when browsing and testing my websites, and  
>> I have never stopped loading a page or had any other problems on  
>> non-IC sites I host.
>> I was told the problem stems from either the browser and a stop  
>> button or some other network fault. I may go back to Apache 1.3 to  
>> get around this.
>
> I saw this occur on two different installations about 4 months  
> ago.  It was suggested that I abandon the use of Link.pm and go  
> back to using the cgi method with URL rewrite rules as this was  
> just as fast and proved stable over the years.

Is plain CGI really as fast as mod_perl or mod_interchange? That'd be  
my only concern with switching back to CGI+rewrites.

Maybe it is ok -- see mod_interchange's README:

"The Interchange link protocol has been
implemented via an Apache module which
saves us the (small) overhead
of the execution of a CGI program."

--
Josh Lavin
Kingdom Design   http://www.kingdomdesign.com/

From ron at endpoint.com  Mon Dec 11 18:35:13 2006
From: ron at endpoint.com (Ron Phipps)
Date: Mon Dec 11 18:35:54 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <EC1A7BEC-D5BB-4717-92C0-2756784119ED@myprivacy.ca>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>	<457DA685.9000201@endpoint.com>
	<EC1A7BEC-D5BB-4717-92C0-2756784119ED@myprivacy.ca>
Message-ID: <457DEB31.7080208@endpoint.com>

Josh Lavin wrote:
> On Dec 11, 2006, at 12:42 PM, Ron Phipps wrote:
>> Josh Lavin wrote:
>>> On Dec 11, 2006, at 12:02 PM, Grant wrote:
>>>>> > Hello, I've been plagued by apache2 segfaults ever since I started
>>>>> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>>>>> > messages accompanying the segfaults in error_log:
>>>>> >
>>>>> > ALERT: bad pipe signal received for /page.html
>>>>> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>>>>> > Segmentation fault (11)
>>>>> >
>>>>> > Does anyone have any advice on solving this?  I'm using 
>>>>> apache-2.0.58
>>>>> > and mod_perl-2.0.2 in Gentoo Linux.
>>>>>
>>>>> Also, here is the portion of Link.pm that the ALERT seems to come 
>>>>> from:
>>>>>
>>>>> # Return this message to the browser when the server is not running.
>>>>> # Log an error log entry if set to notify
>>>>>
>>>>> sub die_page {
>>>>>
>>>>>     my $r = shift;
>>>>>     my $msg;
>>>>>
>>>>>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
>>>>>
>>>>>     $r->content_type ("text/html");
>>>>>     $r->print (<<EOF);
>>>>> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>>>>> <BODY BGCOLOR="#FFFFFF">
>>>>> <H3>Someone pressed stop...</H3>
>>>>> <P>
>>>>> We have aborted this request because someone terminated it.
>>>>> Please try again soon.
>>>>> </BODY></HTML>
>>>>> EOF
>>>>>
>>>>> }
>>>>>
>>>>> Please let me know if you have any ideas.
>>>>
>>>> The segfaults are eliminated by commenting out the $r stuff in the
>>>> die_page sub.  I still get the ALERTs though.  Does anyone have any
>>>> advice on figuring out why I'm having the bad pipe problem?  Is there
>>>> an easy way to add extra debugging info to the sub?
>>>>
>>>> Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs 50 
>>>> fold.
>>> I've been seeing this too, on my Apache 2 and latest Link.pm. I also 
>>> had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>>> The visible effect on the browser is that the page or image (which 
>>> Link.pm apparently still has some part in delivering) does not load. 
>>> I get them myself when browsing and testing my websites, and I have 
>>> never stopped loading a page or had any other problems on non-IC 
>>> sites I host.
>>> I was told the problem stems from either the browser and a stop 
>>> button or some other network fault. I may go back to Apache 1.3 to 
>>> get around this.
>>
>> I saw this occur on two different installations about 4 months ago.  
>> It was suggested that I abandon the use of Link.pm and go back to 
>> using the cgi method with URL rewrite rules as this was just as fast 
>> and proved stable over the years.
> 
> Is plain CGI really as fast as mod_perl or mod_interchange? That'd be my 
> only concern with switching back to CGI+rewrites.
> 
> Maybe it is ok -- see mod_interchange's README:
> 
> "The Interchange link protocol has been
> implemented via an Apache module which
> saves us the (small) overhead
> of the execution of a CGI program."
> 

Josh,

I have not done any a/b tests to determine which is faster, others have.

Vlink/tlink is currently being used without any speed issues during high traffic periods at:

www.bcstore.com
www.frozencpu.com
www.citypass.com
www.reliablemedical.com

Quite a bit of time was put in with BCStore to compare which method was stable and fast and it was determined that tlink/vlink was the way to go.

I thought that we needed to use mod_interchange/mod_perl for FrozenCPU, but quickly found out that under Apache 2.0/mod_perl the issues far outweighed any perceived speed up.  tlink/vlink are compiled c programs and should be very quick in execution.

At End Point we use rewrites with tlink/vlink pretty much exclusively.

Thanks,

-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From kevin at cursor.biz  Mon Dec 11 20:49:19 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Mon Dec 11 20:49:49 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <EC1A7BEC-D5BB-4717-92C0-2756784119ED@myprivacy.ca>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<457DA685.9000201@endpoint.com>
	<EC1A7BEC-D5BB-4717-92C0-2756784119ED@myprivacy.ca>
Message-ID: <200612120149.21199.kevin@cursor.biz>

Josh Lavin <josh@myprivacy.ca> wrote:
> Is plain CGI really as fast as mod_perl or mod_interchange? That'd be  
> my only concern with switching back to CGI+rewrites.
> 
> Maybe it is ok -- see mod_interchange's README:
> 
> "The Interchange link protocol has been
> implemented via an Apache module which
> saves us the (small) overhead
> of the execution of a CGI program."
> 
The overhead referred to is the time taken to fork and exec the CGI
executable.  The executable file itself will most likely be cached,
on even lightly loaded systems, so there's no great concern over the
time it takes to open and read the CGI program prior to execution.

I've never been a great fan of rewrite rules, and there's a small
overhead to be saved there too.

Apart from any perceived efficiency enhancements, and pretty URIs
without rewrites, mod_interchange provides a couple of facilities
such as connection retries, failover to a backup Interchange server
and a request drop list etc., that are not available to CGI link
programs.

Mod_interchange is only available for use with Apache 1.3, but that's
not a massive problem.  Gentoo, for example, will install and maintain
Apache 1.3 instead of 2.x if you ask it to.

There's never been any demand for mod_interchange on Apache 2.x, which
is the main reason why it's still only available on 1.3.  I maintain
mod_interchange because I find it useful.  If I need to use Apache 2.x
for some reason then I'll port it regardless of the demand, or lack
thereof.

I assumed that the lack of demand for mod_interchange on Apache 2.x
was down to the availability of the Interchange::Link mod_perl module.
Interchange::Link shares the same advantages as mod_interchange but
seems to have its quirks.  I'm sure those will be ironed out in time,
as long as the problems are reported.

Any overhead saved by using mod_interchange can very quickly be eaten
up with sloppy page code, so it's much more important to get your pages
to be as efficient as possible, rather than worry about the relative
merits of the various link facilities.

www.interchange.rtfm.info uses mod_interchange, by the way.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From music at labyrinth.net.au  Mon Dec 11 21:23:43 2006
From: music at labyrinth.net.au (Music)
Date: Mon Dec 11 21:24:04 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com><49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com><49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
Message-ID: <019001c71d94$8cb5d2c0$0200a8c0@IBMKZ894FB2E08>

> I've been seeing this too, on my Apache 2 and latest Link.pm. I also  had 
> to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>
> The visible effect on the browser is that the page or image (which 
> Link.pm apparently still has some part in delivering) does not load.  I 
> get them myself when browsing and testing my websites, and I have  never 
> stopped loading a page or had any other problems on non-IC  sites I host.
>
> I was told the problem stems from either the browser and a stop  button or 
> some other network fault. I may go back to Apache 1.3 to  get around this.

I have no idea if this is related in anyway however just in case there is 
any connection:

I find that IC 5.4.1 on Apache 2.0.46-61.ent.centos3 (using secondary 
unthreaded perl 5.8.0 and suexec) and the standard CGI link gets into a 
'faulty' state that only occurs after certain catalogs have recently had 
'Apply Changes' applied.

The result is an intermittent error where a bunch of IC tags are not parsed 
and dumped to the screen (can happen on any page)

Depending on the catalog the webpage output looks something like a broken 
page with something similar displayed:

[control-set search_box_small[/component] [/control-set] [calc 
$Variable->{MV_DHTML_BROWSER} ||= 
'MSIE.*[5-9]\..*Windows|Mozilla.*Gecko|Opera.*[7-9]\.'; 
$Scratch->{dhtml_browser} = $Session->{browser} =~ 
m{$Variable->{MV_DHTML_BROWSER}}; if($Scratch->{members_only} and ! 
$Session->{logged_in}) { $Scratch->{mv_successpage} = $Tag->var('MV_PAGE', 
1); $Tag->deliver({ location => $Tag->area('login')}); } return; [/calc]

To reproduce the error I just need to hit apply changes with a new STANDARD 
store, and then keep refreshing a page. (Hit F5 over and over again.)
The pages will render beautifully however eventually the broken page will 
appear. Another refresh and the problem disappears. (Very intermittent but 
consistent)

The bad thing is that the faulty condition affects all catalogues on the 
server and similar garble will intermittently appear on other catalog pages.

To fix the problem I need to restart Interchange.

I did some more experimenting today inspired by this thread however the 
problem persists.

- I have tried setting all catalogs envrionments to use mod_perl
- I have disabled all mod_perl
- I have restarting both with and without PERL_SIGNALS=unsafe in command 
line.
- I have only tried this using RPC mode with variations of:

ifdef TRAFFIC =~ /rpc/i
Message RPC traffic settings.
PreFork             Yes
StartServers        5
MaxServers          0
MaxRequestsPerChild 100
HouseKeeping        2
PIDcheck            120
ChildLife 30 minutes
endif

I also tried disabling ChildLife and PreFork and increasing start servers.

All with no joy.

'Applying Changes' to one catalogue puts the Interchange daemon into faulty 
state.

No errors are displayed in IC error.log, catalog error.log nor apache error 
logs.

Just thought I would document this on the list to see if any feedback etc.

















From ron at endpoint.com  Mon Dec 11 23:32:43 2006
From: ron at endpoint.com (Ron Phipps)
Date: Mon Dec 11 23:33:23 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <200612120149.21199.kevin@cursor.biz>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>	<457DA685.9000201@endpoint.com>	<EC1A7BEC-D5BB-4717-92C0-2756784119ED@myprivacy.ca>
	<200612120149.21199.kevin@cursor.biz>
Message-ID: <457E30EB.40809@endpoint.com>

Kevin Walsh wrote:
> Josh Lavin <josh@myprivacy.ca> wrote:
>> Is plain CGI really as fast as mod_perl or mod_interchange? That'd be  
>> my only concern with switching back to CGI+rewrites.
>>
>> Maybe it is ok -- see mod_interchange's README:
>>
>> "The Interchange link protocol has been
>> implemented via an Apache module which
>> saves us the (small) overhead
>> of the execution of a CGI program."
>>
> The overhead referred to is the time taken to fork and exec the CGI
> executable.  The executable file itself will most likely be cached,
> on even lightly loaded systems, so there's no great concern over the
> time it takes to open and read the CGI program prior to execution.
> 
> I've never been a great fan of rewrite rules, and there's a small
> overhead to be saved there too.
> 
> Apart from any perceived efficiency enhancements, and pretty URIs
> without rewrites, mod_interchange provides a couple of facilities
> such as connection retries, failover to a backup Interchange server
> and a request drop list etc., that are not available to CGI link
> programs.
> 
> Mod_interchange is only available for use with Apache 1.3, but that's
> not a massive problem.  Gentoo, for example, will install and maintain
> Apache 1.3 instead of 2.x if you ask it to.
> 
> There's never been any demand for mod_interchange on Apache 2.x, which
> is the main reason why it's still only available on 1.3.  I maintain
> mod_interchange because I find it useful.  If I need to use Apache 2.x
> for some reason then I'll port it regardless of the demand, or lack
> thereof.
> 
> I assumed that the lack of demand for mod_interchange on Apache 2.x
> was down to the availability of the Interchange::Link mod_perl module.
> Interchange::Link shares the same advantages as mod_interchange but
> seems to have its quirks.  I'm sure those will be ironed out in time,
> as long as the problems are reported.
> 
> Any overhead saved by using mod_interchange can very quickly be eaten
> up with sloppy page code, so it's much more important to get your pages
> to be as efficient as possible, rather than worry about the relative
> merits of the various link facilities.
> 
> www.interchange.rtfm.info uses mod_interchange, by the way.
> 

With all the being said, I used mod_interchange with Apache 1.3.x for years on FrozenCPU and it worked beautifully during that time and if I were to setup an IC site on Apache 1.3.x today I probably would still use mod_interchange due to the reasons Kevin mentioned.

One thing that was especially nice with mod_interchange is that I could easily add files to  DropRequestList, this helped when we were attacked by a worm that was trying to exploit a XMLRPC bug which had the side effect of hanging IC due to a parsing bug.  I wasn't sure how to fix the parsing bug, but added the filename to the drop list and IC no longer hung.


-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From kevin at cursor.biz  Mon Dec 11 23:51:31 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Mon Dec 11 23:52:57 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <457E30EB.40809@endpoint.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<200612120149.21199.kevin@cursor.biz> <457E30EB.40809@endpoint.com>
Message-ID: <200612120451.33054.kevin@cursor.biz>

Ron Phipps <ron@endpoint.com> wrote:
> One thing that was especially nice with mod_interchange is that I
> could easily add files to  DropRequestList, this helped when we
> were attacked by a worm that was trying to exploit a XMLRPC bug
> which had the side effect of hanging IC due to a parsing bug.  I
> wasn't sure how to fix the parsing bug, but added the filename to
> the drop list and IC no longer hung.     
> 
Just for the record, the DoS bug referenced above has been fixed now.
It was nice to be able to drop the crafty requests while the bug was
being investigated.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From kevin at cursor.biz  Tue Dec 12 00:12:24 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Tue Dec 12 00:12:36 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <019001c71d94$8cb5d2c0$0200a8c0@IBMKZ894FB2E08>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<019001c71d94$8cb5d2c0$0200a8c0@IBMKZ894FB2E08>
Message-ID: <200612120512.25362.kevin@cursor.biz>

"Music" <music@labyrinth.net.au> wrote:
> ifdef TRAFFIC =~ /rpc/i
> Message RPC traffic settings.
> PreFork             Yes
> StartServers        5
> MaxServers          0
> MaxRequestsPerChild 100
> HouseKeeping        2
> PIDcheck            120
> ChildLife 30 minutes
> endif
> 
> I also tried disabling ChildLife and PreFork and increasing start servers.
> 
> All with no joy.
> 
> 'Applying Changes' to one catalogue puts the Interchange daemon into faulty 
> state.
> 
> No errors are displayed in IC error.log, catalog error.log nor apache error 
> logs.
> 
> Just thought I would document this on the list to see if any feedback etc.
> 
Just as an experiment, could you reduce the ChildLife to 60 (1 minute),
restart Interchange and reproduce the problem.  I suspect that the problem
will clear itself within one minute.

If that happens then we know that the problem is probably in Interchange
itself, rather than than in Interchange::Link.

The thinking is that the "apply changes" (or whatever), causes one of
the Interchange children to throw its toys out of its pram.  The constant
refreshes then causes the naughty child to receive a request every now
and again, resulting in a moody page.

One thing that may debunk my theory is your assertion that the problem
still occurs when PreFork is switched off.  Can you check that again
for me.

I'm afraid I don't have an Interchange::Link setup to play with.  If
someone wants to allow me some time on their machine, and a guaranteed
way to reproduce the problem, I'm sure I could track it down reasonably
quickly.

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From cplists at princeservices.com  Tue Dec 12 00:15:07 2006
From: cplists at princeservices.com (Cameron B. Prince)
Date: Tue Dec 12 00:15:48 2006
Subject: [ic] Admin New Entry Problem
Message-ID: <00bc01c71dac$860603b0$0401a8c0@PSLAPTOP1A>

Hey guys,

I'm having a problem with the Admin's New Entry function. This is using a
default standard catalog on IC v5.4.1 with MySQL. When you edit a table that
uses the autonumber counter files as a key, such as products, you can
duplicate the problem by clicking New Entry. Leave the pre-filled key intact
and click Ok to create the new entry. You will see that you are returned to
the flex_editor but instead of it having the next key pre-filled, all the
fields are empty.

You can click New Entry again to create another item, but you will find that
the counter has been incremented twice. I updated the meta for the New Entry
menu item and added a ui_return_to=admin/flex_select to cause the Ok to take
you back to the edit page rather than having the empty flex_editor but was
surprised to find the counter was still incremented twice after creating a
new entry.

So there are really two issues:

1) Is there a logical reason that I'm not realizing for incrementing the
counter twice? If not, how can this be prevented?

2) I understand this will not be ideal in every case, but I'd like the New
Entry function to return to the flex_editor to edit the item I just created.
Can you tell me if this is possible?

Thanks for your help and Merry Christmas to everyone,
Cameron

From kevin at cursor.biz  Tue Dec 12 00:15:51 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Tue Dec 12 00:16:01 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <457DBB5D.7070901@linuxia.de>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612111148g47de91edof2e5eb361d9ed4ea@mail.gmail.com>
	<457DBB5D.7070901@linuxia.de>
Message-ID: <200612120515.52518.kevin@cursor.biz>

"Stefan Hornburg (Racke)" <racke@linuxia.de> wrote:
> Grant wrote:
> > Are you referring to mod_rewrite?  mod_rewrite will only change what
> > appears in the address bar right?  I need something that allows IC to
> > generate folder-free URLs.  Can IC be configured to generate and use
> > such URLs via the cgi method?
> >
> Look into the FullURL directive. This allows you to specify your Catalog like that:
> 
> Catalog linuxia /home/racke/linuxia www.linuxia.de www.linuxia.de:443
> 
You can do that without FullURL too - as long as you only have one
website per Interchange instance:

    Catalog  linuxia  /home/racke/linuxia  /

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From kevin at cursor.biz  Tue Dec 12 00:36:44 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Tue Dec 12 00:36:52 2006
Subject: [ic] Admin New Entry Problem
In-Reply-To: <00bc01c71dac$860603b0$0401a8c0@PSLAPTOP1A>
References: <00bc01c71dac$860603b0$0401a8c0@PSLAPTOP1A>
Message-ID: <200612120536.45737.kevin@cursor.biz>

"Cameron B. Prince" <cplists@princeservices.com> wrote:
> I'm having a problem with the Admin's New Entry function. This is using a
> default standard catalog on IC v5.4.1 with MySQL. When you edit a table that
> uses the autonumber counter files as a key, such as products... [snip]
>
I have no ideas for you at the moment.  I'm just wondering why your
products table has an auto-incrementing key.

Are you able to reproduce the problem on the online demo?

    http://demo.icdevgroup.org/

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From cplists at teslauniverse.com  Tue Dec 12 00:54:03 2006
From: cplists at teslauniverse.com (Cameron B. Prince)
Date: Tue Dec 12 00:54:23 2006
Subject: [ic] Admin New Entry Problem
In-Reply-To: <200612120536.45737.kevin@cursor.biz>
Message-ID: <00bd01c71db1$edfa76e0$0401a8c0@PSLAPTOP1A>

Hi Kevin,

> > I'm having a problem with the Admin's New Entry function. This is using
> a
> > default standard catalog on IC v5.4.1 with MySQL. When you edit a table
> that
> > uses the autonumber counter files as a key, such as products... [snip]
> >
> I have no ideas for you at the moment.  I'm just wondering why your
> products table has an auto-incrementing key.

Almost all, if not all, the default tables will be pre-filled with the
number from the corresponding products/.tablename.autonumber file when
creating a new entry.

> Are you able to reproduce the problem on the online demo?
> 
>     http://demo.icdevgroup.org/

Yes, give it a try.

Thanks,
Cameron

From cplists at princeservices.com  Tue Dec 12 00:55:16 2006
From: cplists at princeservices.com (Cameron B. Prince)
Date: Tue Dec 12 00:55:53 2006
Subject: [ic] Admin New Entry Problem
In-Reply-To: <200612120536.45737.kevin@cursor.biz>
Message-ID: <00be01c71db2$18f0b1c0$0401a8c0@PSLAPTOP1A>

Hi Kevin,

> > I'm having a problem with the Admin's New Entry function. This is using
> a
> > default standard catalog on IC v5.4.1 with MySQL. When you edit a table
> that
> > uses the autonumber counter files as a key, such as products... [snip]
> >
> I have no ideas for you at the moment.  I'm just wondering why your
> products table has an auto-incrementing key.

Almost all, if not all, the default tables will be pre-filled with the
number from the corresponding products/.tablename.autonumber file when
creating a new entry.

> Are you able to reproduce the problem on the online demo?
> 
>     http://demo.icdevgroup.org/

Yes, give it a try.

Thanks,
Cameron


From music at labyrinth.net.au  Tue Dec 12 01:33:55 2006
From: music at labyrinth.net.au (Music)
Date: Tue Dec 12 01:34:23 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com><183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca><019001c71d94$8cb5d2c0$0200a8c0@IBMKZ894FB2E08>
	<200612120512.25362.kevin@cursor.biz>
Message-ID: <03e501c71db7$84e8abd0$0200a8c0@IBMKZ894FB2E08>


----- Original Message ----- 
From: "Kevin Walsh" <kevin@cursor.biz>
To: <interchange-users@icdevgroup.org>
Sent: Tuesday, December 12, 2006 4:12 PM
Subject: Re: [ic] Re: ALERT: bad pipe signal received for /page.html


> "Music" <music@labyrinth.net.au> wrote:
>> ifdef TRAFFIC =~ /rpc/i
>> Message RPC traffic settings.
>> PreFork             Yes
>> StartServers        5
>> MaxServers          0
>> MaxRequestsPerChild 100
>> HouseKeeping        2
>> PIDcheck            120
>> ChildLife 30 minutes
>> endif
>>
>> I also tried disabling ChildLife and PreFork and increasing start 
>> servers.
>>
>> All with no joy.
>>
>> 'Applying Changes' to one catalogue puts the Interchange daemon into 
>> faulty
>> state.
>>
>> No errors are displayed in IC error.log, catalog error.log nor apache 
>> error
>> logs.
>>
>> Just thought I would document this on the list to see if any feedback 
>> etc.
>>
> Just as an experiment, could you reduce the ChildLife to 60 (1 minute),
> restart Interchange and reproduce the problem.  I suspect that the problem
> will clear itself within one minute.

Yes, spot on. The problem pages disappeared after the 60 seconds!

>
> If that happens then we know that the problem is probably in Interchange
> itself, rather than than in Interchange::Link.

Not using Interchange:Link on this server - just regular tlink CGI running 
in Inet Mode.

>
> The thinking is that the "apply changes" (or whatever), causes one of
> the Interchange children to throw its toys out of its pram.  The constant
> refreshes then causes the naughty child to receive a request every now
> and again, resulting in a moody page.
>
> One thing that may debunk my theory is your assertion that the problem
> still occurs when PreFork is switched off.  Can you check that again
> for me.

Testing with Pre-Fork - Oh dear - it appears Pre-For is the problem.
I commented out Pre-Fork and ChildLife and the server is running very 
nicely.
I can't reporduce the problem. Now I have to try and remember why I added 
Pre-Fork and ChildLife; I hope it is not too dramatic a jolt!

I have dropped the PERL_SIGNALS=unsafe in restart command now also and all 
is running well.

>
> I'm afraid I don't have an Interchange::Link setup to play with.  If
> someone wants to allow me some time on their machine, and a guaranteed
> way to reproduce the problem, I'm sure I could track it down reasonably
> quickly.

You are more than welcome to test our Apache 2 setup however as I say not 
using Interchange:Link on this one as it doesn't play with the version of 
mod_perl (1.99_09-10.ent) available for centos3.

>There's never been any demand for mod_interchange on Apache 2.x

There is from some of us!! :-)



From peter at pajamian.dhs.org  Tue Dec 12 02:00:00 2006
From: peter at pajamian.dhs.org (Peter)
Date: Tue Dec 12 02:00:28 2006
Subject: [ic] Admin New Entry Problem
In-Reply-To: <00bc01c71dac$860603b0$0401a8c0@PSLAPTOP1A>
References: <00bc01c71dac$860603b0$0401a8c0@PSLAPTOP1A>
Message-ID: <457E5370.30302@pajamian.dhs.org>

On 12/11/2006 09:15 PM, Cameron B. Prince wrote:
> I'm having a problem with the Admin's New Entry function. This is using a
> default standard catalog on IC v5.4.1 with MySQL. When you edit a table that
> uses the autonumber counter files as a key, such as products, you can
> duplicate the problem by clicking New Entry. Leave the pre-filled key intact
> and click Ok to create the new entry. You will see that you are returned to
> the flex_editor but instead of it having the next key pre-filled, all the
> fields are empty.
> 
> You can click New Entry again to create another item, but you will find that
> the counter has been incremented twice. I updated the meta for the New Entry
> menu item and added a ui_return_to=admin/flex_select to cause the Ok to take
> you back to the edit page rather than having the empty flex_editor but was
> surprised to find the counter was still incremented twice after creating a
> new entry.

I've seen this exact issue and never had the time to look into it.  If I
get the time I'll be happy to look into it, or if someone either (1)
finds the problem themselves and gives me an idea where to look so I
don't have to spend all day and a half tracking it down or (2) pays me
to do it then I'll be likely to commit a fix for it in the near future,
otherwise may not happen for a while (I'm pretty swamped with other
things that take priority).  That's speaking for me, someone else may
fix it.

FWIW this problem goes back at least as far as 5.2.0 and is not specific
to MySQL (I get the same problem with PostgreSQL).  It's not specific to
standard, either, I see it in Foundation-based sites.

> 1) Is there a logical reason that I'm not realizing for incrementing the
> counter twice? If not, how can this be prevented?

Probably not, it's just a minor bug.

> 2) I understand this will not be ideal in every case, but I'd like the New
> Entry function to return to the flex_editor to edit the item I just created.
> Can you tell me if this is possible?

That sounds like a good idea.  It should be possible.

Peter

From pienihetki at gmail.com  Tue Dec 12 07:55:40 2006
From: pienihetki at gmail.com (pienihetki)
Date: Tue Dec 12 07:56:07 2006
Subject: [ic] SOAP - login tag issue
Message-ID: <457EA6CC.306@gmail.com>

Trying to use the code below to prove I can login to my site using 
SOAP.  If I turn on debug in interchange I do see the login reported as 
successful in the SOAP log.  However, the return value from calling the 
tag is always undefined!  I am using IC 4.8.7 and SOAP::Lite 0.68.  Any 
tips on how to debug this?  Thanks.

use SOAP::Lite;
my $cat = 'catalog';
my $id = 'ABCDEFGH:soap';

my $Tag = SOAP::Lite
        -> uri('http://10.27.29.148/Vend/SOAP')
        -> proxy("http://10.27.92.148:7780/$cat/$id");

my $login_opt = {
        username => 'test',
        password => 'test',
        function => 'login'
        };
my $ok = $Tag->userdb( $login_opt );
if (defined($ok)) {
        print "OK defined\n";
} else {
        print "Ok not defined\n";
}
From ron at endpoint.com  Tue Dec 12 13:03:50 2006
From: ron at endpoint.com (Ron Phipps)
Date: Tue Dec 12 13:04:32 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <03e501c71db7$84e8abd0$0200a8c0@IBMKZ894FB2E08>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com><183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca><019001c71d94$8cb5d2c0$0200a8c0@IBMKZ894FB2E08>	<200612120512.25362.kevin@cursor.biz>
	<03e501c71db7$84e8abd0$0200a8c0@IBMKZ894FB2E08>
Message-ID: <457EEF06.2020607@endpoint.com>

Music wrote:
> 
> ----- Original Message ----- From: "Kevin Walsh" <kevin@cursor.biz>
> To: <interchange-users@icdevgroup.org>
> Sent: Tuesday, December 12, 2006 4:12 PM
> Subject: Re: [ic] Re: ALERT: bad pipe signal received for /page.html
> 
> 
>> "Music" <music@labyrinth.net.au> wrote:
>>> ifdef TRAFFIC =~ /rpc/i
>>> Message RPC traffic settings.
>>> PreFork             Yes
>>> StartServers        5
>>> MaxServers          0
>>> MaxRequestsPerChild 100
>>> HouseKeeping        2
>>> PIDcheck            120
>>> ChildLife 30 minutes
>>> endif
>>>
>>> I also tried disabling ChildLife and PreFork and increasing start 
>>> servers.
>>>
>>> All with no joy.
>>>
>>> 'Applying Changes' to one catalogue puts the Interchange daemon into 
>>> faulty
>>> state.
>>>
>>> No errors are displayed in IC error.log, catalog error.log nor apache 
>>> error
>>> logs.
>>>
>>> Just thought I would document this on the list to see if any feedback 
>>> etc.
>>>
>> Just as an experiment, could you reduce the ChildLife to 60 (1 minute),
>> restart Interchange and reproduce the problem.  I suspect that the 
>> problem
>> will clear itself within one minute.
> 
> Yes, spot on. The problem pages disappeared after the 60 seconds!
> 
>>
>> If that happens then we know that the problem is probably in Interchange
>> itself, rather than than in Interchange::Link.
> 
> Not using Interchange:Link on this server - just regular tlink CGI 
> running in Inet Mode.
> 
>>
>> The thinking is that the "apply changes" (or whatever), causes one of
>> the Interchange children to throw its toys out of its pram.  The constant
>> refreshes then causes the naughty child to receive a request every now
>> and again, resulting in a moody page.
>>
>> One thing that may debunk my theory is your assertion that the problem
>> still occurs when PreFork is switched off.  Can you check that again
>> for me.
> 
> Testing with Pre-Fork - Oh dear - it appears Pre-For is the problem.
> I commented out Pre-Fork and ChildLife and the server is running very 
> nicely.
> I can't reporduce the problem. Now I have to try and remember why I 
> added Pre-Fork and ChildLife; I hope it is not too dramatic a jolt!
> 
> I have dropped the PERL_SIGNALS=unsafe in restart command now also and 
> all is running well.

You may have been running Pre-Fork mode to help with issues where sendmail or your gateway charges were not working properly.  If that is the case make sure that MaxServers 0 is present in your high and low traffic areas of interchange.cfg, this is the new default.
 
-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From MichaelC at glcweb.co.uk  Tue Dec 12 18:12:45 2006
From: MichaelC at glcweb.co.uk (Michael Curtis)
Date: Tue Dec 12 18:13:11 2006
Subject: [ic] Protx issues
Message-ID: <A33C77E06C9E924F8E6D796CA3D635D10238B6@w2k3sbs.glcdomain.local>

Hi all

FreeBSD 6.0
IC 5.4
Mysql
Protx 1.0.9f

Configuring the Standard store for Protx Payment Gateway using Lynn StGeorges module and instructions

All went well until I tried to create the new transactions table

I copied the new transactions.txt to the correct location
I copied the new transactions.mysql to the correct location
I deleted .transactions.sql

Then restarted Interchange

Errors are that generated are:
Snip...................
> sync2 varchar(128),
> protx1 varchar(128),
> protx2 varchar(128),
> protx3 varchar(128)
> )
> 
- - - [11/December/2006:22:54:50 +0000] - - table transactions index failed: Table 'test_standard.transactions' doesn't exist
- - - [11/December/2006:22:54:50 +0000] - - DBI: Post creation query 'CREATE  INDEX transactions_order_number ON transactions (order_number)' failed: Table 'test_standard.transactions' doesn't exist
- - - [11/December/2006:22:54:50 +0000] - - table 'transactions' failed: list_fields execute on transactions: Table 'test_standard.transactions' doesn't exist at /usr/local/interchange/lib/Vend/Table/DBI.pm line 1752.

I have been over all file permissions etc and they are all o.k

If I roll back to the standard transaction files the standard transaction table is created and all is o.k

I could create the table by creating a .sql and importing that but I do not know if that may create other problems

TIA

Mike Curtis



-- 
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.15.16/582 - Release Date: 11/12/2006 16:32
 

From emailgrant at gmail.com  Tue Dec 12 18:43:29 2006
From: emailgrant at gmail.com (Grant)
Date: Tue Dec 12 18:43:45 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
Message-ID: <49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>

> >> > Hello, I've been plagued by apache2 segfaults ever since I started
> >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> >> > messages accompanying the segfaults in error_log:
> >> >
> >> > ALERT: bad pipe signal received for /page.html
> >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> >> > Segmentation fault (11)
> >> >
> >> > Does anyone have any advice on solving this?  I'm using
> >> apache-2.0.58
> >> > and mod_perl-2.0.2 in Gentoo Linux.
> >>
> >> Also, here is the portion of Link.pm that the ALERT seems to come
> >> from:
> >>
> >> # Return this message to the browser when the server is not running.
> >> # Log an error log entry if set to notify
> >>
> >> sub die_page {
> >>
> >>     my $r = shift;
> >>     my $msg;
> >>
> >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> >>
> >>     $r->content_type ("text/html");
> >>     $r->print (<<EOF);
> >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> >> <BODY BGCOLOR="#FFFFFF">
> >> <H3>Someone pressed stop...</H3>
> >> <P>
> >> We have aborted this request because someone terminated it.
> >> Please try again soon.
> >> </BODY></HTML>
> >> EOF
> >>
> >> }
> >>
> >> Please let me know if you have any ideas.
> >
> > The segfaults are eliminated by commenting out the $r stuff in the
> > die_page sub.  I still get the ALERTs though.  Does anyone have any
> > advice on figuring out why I'm having the bad pipe problem?  Is there
> > an easy way to add extra debugging info to the sub?
> >
> > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
> > 50 fold.
>
> I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>
> The visible effect on the browser is that the page or image (which
> Link.pm apparently still has some part in delivering) does not load.
> I get them myself when browsing and testing my websites, and I have
> never stopped loading a page or had any other problems on non-IC
> sites I host.
>
> I was told the problem stems from either the browser and a stop
> button or some other network fault. I may go back to Apache 1.3 to
> get around this.

I've been working on this all day and I think I may have a solution.

Incidentally, I want to mention that if I add $! to the Link.pm warn
line, I can see that there are two different types of ALERTs in
error_log:

Broken pipe
Inappropriate ioctl for device

Also incidentally, hitting F5 repeatedly always prints a few ALERTS in
error_log, but the number of ALERTs printed seems to be about 10 times
more for an http page than for the same page accessed via https.  I
checked and tested the settings in my *:80 and *:443 vhosts trying to
narrow that down, but I didn't come up with anything.

Now, as I mentioned a few posts back, commenting out the html delivery
stuff in the die_page sub of Link.pm eliminates the segfaults in
error_log but not the ALERTs.  Today I enabled the prefork mpm in
apache2 and tinkered with the Server/Child settings in interchange.cfg
and httpd.conf and that did seem to reduce the ALERTs somewhat.

After all of this, I've been browsing around my site and I haven't
seen a single image or page fail to load like it used to.  The thing
is, ALERTs still show up in error_log even for requests I'm sure I
created myself AND WERE SERVED SUCCESSFULLY.  I hypothesize that the
failed requests were because a $SIG{PIPE} was generated for whatever
reason (although not because the user clicked the stop button), the
html was delivered, it caused a segfault, and the request failed.  I
should mention though, that if the failing request syndrome is fixed,
it could also be from the apache2, httpd.conf, or interchange.cfg
modifications mentioned above.  I doubt it though.

Incidentally, I'd like to mention that the ALERTs can be generated
pretty reliably by hitting the browser's stop button in the middle of
a request, clicking on the same link again or a different link in the
middle of a request (my Mom probably still double clicks links), or
just refreshing the page in the middle of a request.

If the failed requests do stop at this point, I think the die_page
section of Link.pm is doing more harm than good.  Is there anything
wrong with not handling $SIG{PIPE} at all?

- Grant
From lyn at zolotek.net  Wed Dec 13 06:19:32 2006
From: lyn at zolotek.net (Lyn St George)
Date: Wed Dec 13 06:18:13 2006
Subject: [ic] Protx issues
In-Reply-To: <A33C77E06C9E924F8E6D796CA3D635D10238B6@w2k3sbs.glcdomain.local>
Message-ID: <20061213111202.F1904830073@mailserver1.zolotek.net>

On Tue, 12 Dec 2006 23:12:45 -0000, Michael Curtis wrote:

>Hi all
>
>FreeBSD 6.0
>IC 5.4
>Mysql
>Protx 1.0.9f
>
>Configuring the Standard store for Protx Payment Gateway using Lynn StGeorges 
module and instructions
>
>All went well until I tried to create the new transactions table
>
>I copied the new transactions.txt to the correct location
>I copied the new transactions.mysql to the correct location
>I deleted .transactions.sql
>
>Then restarted Interchange
>
>Errors are that generated are:
>Snip...................
>> sync2 varchar(128),
>> protx1 varchar(128),
>> protx2 varchar(128),
>> protx3 varchar(128)
>> )

You could probably delete all of those unused fields - they were
put there to future-proof it for an update to the virtual terminal, but
as I've split the terminal out these are now superfluous. 

Though keep the ones that are listed in log_transaction. 
 
>- - - [11/December/2006:22:54:50 +0000] - - table transactions index failed: Table 
'test_standard.transactions' doesn't exist
>- - - [11/December/2006:22:54:50 +0000] - - DBI: Post creation query 'CREATE  INDEX 
transactions_order_number ON transactions (order_number)' failed: Table 
'test_standard.transactions' doesn't exist
>- - - [11/December/2006:22:54:50 +0000] - - table 'transactions' failed: list_fields execute 
on transactions: Table 'test_standard.transactions' doesn't exist at 
/usr/local/interchange/lib/Vend/Table/DBI.pm line 1752.
>
>I have been over all file permissions etc and they are all o.k
>
>If I roll back to the standard transaction files the standard transaction table is created and 
all is o.k
>
>I could create the table by creating a .sql and importing that but I do not know if that may 
create other problems

I don't know if BSD is different to Linux, but maybe you need to
explicitly specify these fields in dbconf/mysql/transactions.mysql?

Have you checked that the files are in Unix line endings? Having
them in DOS line-endings can cause similar problems. 

>TIA
>
>Mike Curtis
>

-
Lyn

From emailgrant at gmail.com  Wed Dec 13 10:07:57 2006
From: emailgrant at gmail.com (Grant)
Date: Wed Dec 13 10:08:37 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>
Message-ID: <49bf44f10612130707m6b693904waac5228452fb609b@mail.gmail.com>

> > >> > Hello, I've been plagued by apache2 segfaults ever since I started
> > >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> > >> > messages accompanying the segfaults in error_log:
> > >> >
> > >> > ALERT: bad pipe signal received for /page.html
> > >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> > >> > Segmentation fault (11)
> > >> >
> > >> > Does anyone have any advice on solving this?  I'm using
> > >> apache-2.0.58
> > >> > and mod_perl-2.0.2 in Gentoo Linux.
> > >>
> > >> Also, here is the portion of Link.pm that the ALERT seems to come
> > >> from:
> > >>
> > >> # Return this message to the browser when the server is not running.
> > >> # Log an error log entry if set to notify
> > >>
> > >> sub die_page {
> > >>
> > >>     my $r = shift;
> > >>     my $msg;
> > >>
> > >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> > >>
> > >>     $r->content_type ("text/html");
> > >>     $r->print (<<EOF);
> > >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> > >> <BODY BGCOLOR="#FFFFFF">
> > >> <H3>Someone pressed stop...</H3>
> > >> <P>
> > >> We have aborted this request because someone terminated it.
> > >> Please try again soon.
> > >> </BODY></HTML>
> > >> EOF
> > >>
> > >> }
> > >>
> > >> Please let me know if you have any ideas.
> > >
> > > The segfaults are eliminated by commenting out the $r stuff in the
> > > die_page sub.  I still get the ALERTs though.  Does anyone have any
> > > advice on figuring out why I'm having the bad pipe problem?  Is there
> > > an easy way to add extra debugging info to the sub?
> > >
> > > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
> > > 50 fold.
> >
> > I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> > had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> >
> > The visible effect on the browser is that the page or image (which
> > Link.pm apparently still has some part in delivering) does not load.
> > I get them myself when browsing and testing my websites, and I have
> > never stopped loading a page or had any other problems on non-IC
> > sites I host.
> >
> > I was told the problem stems from either the browser and a stop
> > button or some other network fault. I may go back to Apache 1.3 to
> > get around this.
>
> I've been working on this all day and I think I may have a solution.
>
> Incidentally, I want to mention that if I add $! to the Link.pm warn
> line, I can see that there are two different types of ALERTs in
> error_log:
>
> Broken pipe
> Inappropriate ioctl for device
>
> Also incidentally, hitting F5 repeatedly always prints a few ALERTS in
> error_log, but the number of ALERTs printed seems to be about 10 times
> more for an http page than for the same page accessed via https.  I
> checked and tested the settings in my *:80 and *:443 vhosts trying to
> narrow that down, but I didn't come up with anything.
>
> Now, as I mentioned a few posts back, commenting out the html delivery
> stuff in the die_page sub of Link.pm eliminates the segfaults in
> error_log but not the ALERTs.  Today I enabled the prefork mpm in
> apache2 and tinkered with the Server/Child settings in interchange.cfg
> and httpd.conf and that did seem to reduce the ALERTs somewhat.
>
> After all of this, I've been browsing around my site and I haven't
> seen a single image or page fail to load like it used to.  The thing
> is, ALERTs still show up in error_log even for requests I'm sure I
> created myself AND WERE SERVED SUCCESSFULLY.  I hypothesize that the
> failed requests were because a $SIG{PIPE} was generated for whatever
> reason (although not because the user clicked the stop button), the
> html was delivered, it caused a segfault, and the request failed.  I
> should mention though, that if the failing request syndrome is fixed,
> it could also be from the apache2, httpd.conf, or interchange.cfg
> modifications mentioned above.  I doubt it though.
>
> Incidentally, I'd like to mention that the ALERTs can be generated
> pretty reliably by hitting the browser's stop button in the middle of
> a request, clicking on the same link again or a different link in the
> middle of a request (my Mom probably still double clicks links), or
> just refreshing the page in the middle of a request.
>
> If the failed requests do stop at this point, I think the die_page
> section of Link.pm is doing more harm than good.  Is there anything
> wrong with not handling $SIG{PIPE} at all?
>
> - Grant

I still have yet to see a single failed image or page request in the
browser since the above changes, although ALERTs are still generated
in error_log.  Can anyone verify this?

- Grant
From emailgrant at gmail.com  Wed Dec 13 10:37:17 2006
From: emailgrant at gmail.com (Grant)
Date: Wed Dec 13 10:37:31 2006
Subject: [ic] PIDcheck vs. ChildLife
Message-ID: <49bf44f10612130737g23f3a423v7f5084851f43cbcb@mail.gmail.com>

What is the difference between PIDcheck and ChildLife?  It sounds like
they do the same thing, unless a PID and child aren't the same thing.

Will setting one or both of these directives interrupt a job's execution?

- Grant
From jon at endpoint.com  Wed Dec 13 15:18:54 2006
From: jon at endpoint.com (Jon Jensen)
Date: Wed Dec 13 15:19:17 2006
Subject: [ic] PIDcheck vs. ChildLife
In-Reply-To: <49bf44f10612130737g23f3a423v7f5084851f43cbcb@mail.gmail.com>
References: <49bf44f10612130737g23f3a423v7f5084851f43cbcb@mail.gmail.com>
Message-ID: <Pine.LNX.4.64.0612131314110.10153@ybpnyubfg.ybpnyqbznva>

On Wed, 13 Dec 2006, Grant wrote:

> What is the difference between PIDcheck and ChildLife?  It sounds like 
> they do the same thing, unless a PID and child aren't the same thing.
>
> Will setting one or both of these directives interrupt a job's 
> execution?

PIDcheck says that the housekeeping routine should look for out-of-control 
or dead Interchange child processes and kill them if they've been running 
more than that many seconds:

http://www.icdevgroup.org/docs/confs/PIDcheck.html

ChildLife limits how long a child is allowed to live, primarily for 
PreFork mode persistent children. For example, you may want to say that a 
child can only live for 10 minutes total, regardless of whether it's met 
its MaxRequestsPerChild limit yet:

http://www.icdevgroup.org/docs/confs/ChildLife.html

Most likely you do not need ChildLife, but you should almost always have 
PIDcheck set to a couple of minutes, so you don't have runaway child 
processes going longer than that.

Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From DB at M-and-D.com  Wed Dec 13 19:53:02 2006
From: DB at M-and-D.com (DB)
Date: Wed Dec 13 19:53:16 2006
Subject: [ic] removing .../cgi-bin/catname/... from URLs
Message-ID: <4580A06E.3090408@M-and-D.com>

Hi - I have an interchange store where a tag such as

[page home]home</a>

will when clicked send me to the url

http://domain1.com/home.html

Now I've set up another machine and another store, but on this new one

[page home]home</a>

sends me to

http://domain2.com/cgi-bin/catname/home.html

I can't for the life of me remember how I got the simplified urls to be
the default. I have the rewrite rules in place on the new store, but
what's the magic I need to have IC use the simpler URLs by default? Can
anyone jog my memory?

DB
From jon at endpoint.com  Wed Dec 13 20:11:01 2006
From: jon at endpoint.com (Jon Jensen)
Date: Wed Dec 13 20:11:22 2006
Subject: [ic] removing .../cgi-bin/catname/... from URLs
In-Reply-To: <4580A06E.3090408@M-and-D.com>
References: <4580A06E.3090408@M-and-D.com>
Message-ID: <Pine.LNX.4.64.0612131805050.10153@ybpnyubfg.ybpnyqbznva>

On Wed, 13 Dec 2006, DB wrote:

> Now I've set up another machine and another store, but on this new one
>
> [page home]home</a>
>
> sends me to
>
> http://domain2.com/cgi-bin/catname/home.html
>
> I can't for the life of me remember how I got the simplified urls to be
> the default. I have the rewrite rules in place on the new store, but
> what's the magic I need to have IC use the simpler URLs by default? Can
> anyone jog my memory?

In your interchange.cfg Catalog definition, the last path prefix alias is 
the one that will be used for writing URLs. E.g.:

   Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /

[area home] -> http://domain2.com/home.html

Whereas:

   Catalog  mycat  /home/someuser/catalogs/mycat  /  /cgi-bin/mycat

[area home] -> http://domain2.com/cgi-bin/mycat/home.html

Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From emailgrant at gmail.com  Wed Dec 13 20:23:55 2006
From: emailgrant at gmail.com (Grant)
Date: Wed Dec 13 20:24:17 2006
Subject: [ic] PIDcheck vs. ChildLife
In-Reply-To: <Pine.LNX.4.64.0612131314110.10153@ybpnyubfg.ybpnyqbznva>
References: <49bf44f10612130737g23f3a423v7f5084851f43cbcb@mail.gmail.com>
	<Pine.LNX.4.64.0612131314110.10153@ybpnyubfg.ybpnyqbznva>
Message-ID: <49bf44f10612131723k49f8c073x970594385bc7da37@mail.gmail.com>

> > What is the difference between PIDcheck and ChildLife?  It sounds like
> > they do the same thing, unless a PID and child aren't the same thing.
> >
> > Will setting one or both of these directives interrupt a job's
> > execution?
>
> PIDcheck says that the housekeeping routine should look for out-of-control
> or dead Interchange child processes and kill them if they've been running
> more than that many seconds:
>
> http://www.icdevgroup.org/docs/confs/PIDcheck.html
>
> ChildLife limits how long a child is allowed to live, primarily for
> PreFork mode persistent children. For example, you may want to say that a
> child can only live for 10 minutes total, regardless of whether it's met
> its MaxRequestsPerChild limit yet:
>
> http://www.icdevgroup.org/docs/confs/ChildLife.html
>
> Most likely you do not need ChildLife, but you should almost always have
> PIDcheck set to a couple of minutes, so you don't have runaway child
> processes going longer than that.
>
> Jon

Thanks Jon.  I have jobs that execute for over 2 minutes, but PIDcheck
is set to 120.  Shouldn't PIDcheck be killing those jobs off at 2
minutes?

- Grant
From DB at M-and-D.com  Wed Dec 13 20:49:32 2006
From: DB at M-and-D.com (DB)
Date: Wed Dec 13 20:49:43 2006
Subject: [ic] removing .../cgi-bin/catname/... from URLs
References: 4580A06E.3090408@M-and-D.com
Message-ID: <4580ADAC.8050603@M-and-D.com>

>> Now I've set up another machine and another store, but on this new one
>>
>> [page home]home</a>
>>
>> sends me to
>>
>> http://domain2.com/cgi-bin/catname/home.html
>>
>> I can't for the life of me remember how I got the simplified urls to be
>> the default. I have the rewrite rules in place on the new store, but
>> what's the magic I need to have IC use the simpler URLs by default? Can
>> anyone jog my memory?
> 
> In your interchange.cfg Catalog definition, the last path prefix alias is 
> the one that will be used for writing URLs. E.g.:
> 
>    Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /
> 
> [area home] -> http://domain2.com/home.html
> 
> Whereas:
> 
>    Catalog  mycat  /home/someuser/catalogs/mycat  /  /cgi-bin/mycat
> 
> [area home] -> http://domain2.com/cgi-bin/mycat/home.html
> 
> Jon


Hi - I've verified that Catalog definition indeed has the format:

Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /

yet IC is still including the .../cgi-bin/mycat/... in the URLs. I
restarted IC just for kicks yet still no luck. Can you suggest a way to
troubleshoot?

I've noticed that the misbehaving store has a CGI_URL defined in the
Prefernces tab of the Admin GUI, but the behaving store does not. Could
this be the cause?


DB
From Mutilovic at notyourtypicalrealtor.com  Wed Dec 13 23:52:21 2006
From: Mutilovic at notyourtypicalrealtor.com (Huerta Andrey)
Date: Wed Dec 13 23:59:59 2006
Subject: [ic] Make it larger!
Message-ID: <463501c71f3b$02864ef4$c9edd43e@[62.212.237.201]>

Hei buddy
 
You Love Big tits? But Girls love big schlong! 

If you don't have one - GET ONE!

Not only a larger one-eyed monster will make you feel better,
it will make you look better!

Get a months supply and see the difference!

No Pumps! No Surgery! No Exercises!

100% Guaranteed Safe Results Or Your Money Back! 

Take a look at how it works: http://

-- 
cmndfjgdfgjwhewdqpqqdc gmjojukljlkljljtjljljtjuflkoktjqksjimgkqkfkfkfjfkfjpkljujqghkhjrkn

' if you ask me you're throwing that money away.'
     'It'll come back,' replied Rimsky quietly, ' and then he'll pay dearly
for this little picnic.' And pointing at Varenukha's briefcase he said :
     'Go on, Ivan Savyelich, don't waste any time.' Varenukha picked up his
briefcase and trotted off. He went down to the ground floor, saw a very long
queue outside  the  box  office and heard  from  the  cashier that  she  was
expecting to have  to put up the ' House Full' notices that evening  because
they  were  being positively  overwhelmed  since the  special  bill had been
posted up.  Varenukha told her  to be sure not to sell the thirty best seats
in the boxes and stalls,  then rushed out of  the box office, fought off the


From peter at pajamian.dhs.org  Thu Dec 14 00:13:21 2006
From: peter at pajamian.dhs.org (Peter)
Date: Thu Dec 14 00:13:51 2006
Subject: [ic] removing .../cgi-bin/catname/... from URLs
In-Reply-To: <4580ADAC.8050603@M-and-D.com>
References: 4580A06E.3090408@M-and-D.com <4580ADAC.8050603@M-and-D.com>
Message-ID: <4580DD71.6020109@pajamian.dhs.org>

On 12/13/2006 05:49 PM, DB wrote:
>>> Now I've set up another machine and another store, but on this new one
>>>
>>> [page home]home</a>
>>>
>>> sends me to
>>>
>>> http://domain2.com/cgi-bin/catname/home.html
>>>
>>> I can't for the life of me remember how I got the simplified urls to be
>>> the default. I have the rewrite rules in place on the new store, but
>>> what's the magic I need to have IC use the simpler URLs by default? Can
>>> anyone jog my memory?
>> In your interchange.cfg Catalog definition, the last path prefix alias is 
>> the one that will be used for writing URLs. E.g.:
>>
>>    Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /
>>
>> [area home] -> http://domain2.com/home.html
>>
>> Whereas:
>>
>>    Catalog  mycat  /home/someuser/catalogs/mycat  /  /cgi-bin/mycat
>>
>> [area home] -> http://domain2.com/cgi-bin/mycat/home.html
>>
>> Jon
> 
> 
> Hi - I've verified that Catalog definition indeed has the format:
> 
> Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /
> 
> yet IC is still including the .../cgi-bin/mycat/... in the URLs. I
> restarted IC just for kicks yet still no luck. Can you suggest a way to
> troubleshoot?
> 
> I've noticed that the misbehaving store has a CGI_URL defined in the
> Prefernces tab of the Admin GUI, but the behaving store does not. Could
> this be the cause?

Indirectly, yes.  Look for VendURL and SecureURL in catalog.cfg.  Also see:
<http://www.interchange.rtfm.info/icdocs/config/VendURL.html>
<http://www.interchange.rtfm.info/icdocs/config/SecureURL.html>

Peter

From jon at endpoint.com  Thu Dec 14 00:54:06 2006
From: jon at endpoint.com (Jon Jensen)
Date: Thu Dec 14 00:54:38 2006
Subject: [ic] removing .../cgi-bin/catname/... from URLs
In-Reply-To: <4580DD71.6020109@pajamian.dhs.org>
References: 4580A06E.3090408@M-and-D.com <4580ADAC.8050603@M-and-D.com>
	<4580DD71.6020109@pajamian.dhs.org>
Message-ID: <Pine.LNX.4.64.0612132241430.10153@ybpnyubfg.ybpnyqbznva>

On Wed, 13 Dec 2006, Peter wrote:

>> I've verified that Catalog definition indeed has the format:
>>
>> Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /
>>
>> yet IC is still including the .../cgi-bin/mycat/... in the URLs. I
>> restarted IC just for kicks yet still no luck. Can you suggest a way to
>> troubleshoot?
>>
>> I've noticed that the misbehaving store has a CGI_URL defined in the
>> Prefernces tab of the Admin GUI, but the behaving store does not. Could
>> this be the cause?
>
> Indirectly, yes.  Look for VendURL and SecureURL in catalog.cfg.

DB,

I need to explain my earlier reply better. What Peter says about VendURL 
is in general correct. However, when you have any URL prefix aliases 
defined for the catalog, if a request comes in through one of them, that 
will override VendURL for all future URLs written by [area] etc. The 
manual explains this:

http://www.icdevgroup.org/docs/confs/Catalog.html

So my example only makes sense if taken in the context of an Apache setup 
for making pretty catalog URLs, such as this:

     ScriptAlias /cgi-bin/ /path/to/your/cgi-bin/dir/
     RewriteEngine on
     RewriteRule ^/store/(.*) /cgi-bin/standard/$1 [PT,L]

Then you'd set up interchange.cfg like this:

     Catalog  standard  /path/to/catalogs/standard  /cgi-bin/standard  /store

Thus, users would go to this URL to enter the site:

     http://your.domain/store/

But Interchange would see you coming in through:

     http://your.domain/cgi-bin/standard/

because of the Apache RewriteRule.

The first URL prefix is considered the "main" one, not an alias, and does 
not take over URL generation as an alias does, so the Catalog definition 
above doesn't lead to aliases URL generation; it just stays normal.

But in the end it turns out that the /store alias I showed in my example 
isn't even necessary; Interchange never uses it or needs it to find the 
catalog, because it's seeing /cgi-bin/standard as the path prefix after 
URL rewriting.

So, to sum up, I think your problem may be fixed with this Catalog 
directive:

     Catalog  mycat  /home/someuser/catalogs/mycat  /

assuming you have in Apache something like:

     RewriteRule ^/(.*) /cgi-bin/mycat/$1 [PT,L]

Though to take over the whole URL space like that, you'd need to have some 
previous RewriteRules that would make exceptions to statically serve 
images, CSS, JavaScript, robots.txt, favicon.ico, etc. directly from 
Apache, not passing them to Interchange.

You'll get a lot more wasted requests passed through to Interchange by 
hostile bots, etc., which wouldn't be if you use a prefix like /store/ or 
/ic/ or /app/ or something like that. You can still make the home page use 
the root URL http://your.domain/ with another RewriteRule like:

     RewriteRule ^/(index.html?)?$ /cgi-bin/standard/index [PT,L]

HTH,
Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From frank at goldissue.com  Thu Dec 14 01:39:25 2006
From: frank at goldissue.com (Frank Reitzenstein)
Date: Thu Dec 14 01:39:42 2006
Subject: [ic] removing .../cgi-bin/catname/... from URLs
In-Reply-To: <Pine.LNX.4.64.0612132241430.10153@ybpnyubfg.ybpnyqbznva>
References: 4580A06E.3090408@M-and-D.com
	<4580ADAC.8050603@M-and-D.com>	<4580DD71.6020109@pajamian.dhs.org>
	<Pine.LNX.4.64.0612132241430.10153@ybpnyubfg.ybpnyqbznva>
Message-ID: <4580F19D.8070906@goldissue.com>

Jon Jensen wrote:

> On Wed, 13 Dec 2006, Peter wrote:
>
>>> I've verified that Catalog definition indeed has the format:
>>>
>>> Catalog  mycat  /home/someuser/catalogs/mycat  /cgi-bin/mycat  /
>>>
>>> yet IC is still including the .../cgi-bin/mycat/... in the URLs. I
>>> restarted IC just for kicks yet still no luck. Can you suggest a way to
>>> troubleshoot?
>>>
>>> I've noticed that the misbehaving store has a CGI_URL defined in the
>>> Prefernces tab of the Admin GUI, but the behaving store does not. Could
>>> this be the cause?
>>
>>
>> Indirectly, yes.  Look for VendURL and SecureURL in catalog.cfg.
>
>
> DB,
>
> I need to explain my earlier reply better. What Peter says about
> VendURL is in general correct. However, when you have any URL prefix
> aliases defined for the catalog, if a request comes in through one of
> them, that will override VendURL for all future URLs written by [area]
> etc. The manual explains this:
>
> http://www.icdevgroup.org/docs/confs/Catalog.html
>
> So my example only makes sense if taken in the context of an Apache
> setup for making pretty catalog URLs, such as this:
>
>     ScriptAlias /cgi-bin/ /path/to/your/cgi-bin/dir/
>     RewriteEngine on
>     RewriteRule ^/store/(.*) /cgi-bin/standard/$1 [PT,L]
>
> Then you'd set up interchange.cfg like this:
>
>     Catalog  standard  /path/to/catalogs/standard  /cgi-bin/standard 
> /store
>
> Thus, users would go to this URL to enter the site:
>
>     http://your.domain/store/
>
> But Interchange would see you coming in through:
>
>     http://your.domain/cgi-bin/standard/
>
> because of the Apache RewriteRule.
>
> The first URL prefix is considered the "main" one, not an alias, and
> does not take over URL generation as an alias does, so the Catalog
> definition above doesn't lead to aliases URL generation; it just stays
> normal.
>
> But in the end it turns out that the /store alias I showed in my
> example isn't even necessary; Interchange never uses it or needs it to
> find the catalog, because it's seeing /cgi-bin/standard as the path
> prefix after URL rewriting.
>
> So, to sum up, I think your problem may be fixed with this Catalog
> directive:
>
>     Catalog  mycat  /home/someuser/catalogs/mycat  /
>
> assuming you have in Apache something like:
>
>     RewriteRule ^/(.*) /cgi-bin/mycat/$1 [PT,L]
>
> Though to take over the whole URL space like that, you'd need to have
> some previous RewriteRules that would make exceptions to statically
> serve images, CSS, JavaScript, robots.txt, favicon.ico, etc. directly
> from Apache, not passing them to Interchange.
>
> You'll get a lot more wasted requests passed through to Interchange by
> hostile bots, etc., which wouldn't be if you use a prefix like /store/
> or /ic/ or /app/ or something like that. You can still make the home
> page use the root URL http://your.domain/ with another RewriteRule like:
>
>     RewriteRule ^/(index.html?)?$ /cgi-bin/standard/index [PT,L]
>
> HTH,
> Jon
>
Hello Peter,

Below is an apache virtual host which does what you are seeking.

To get a seamless catalog at "/" all we would need to do is change
products/variable.txt as follows:

CGI_URL    /shop/example (ie /cgi-bin/example)

to

CGI_URL    /


Unfortunately we can't. We still need to check out at
https://www.goldissue.com

I tried using ports 444, 445, 446 for additional secure servers, but it
worked most of the time, then people would ring in and say that they
couldn't check out. Our server is on a very fast adsl connection, but
the isp is too inflexibble to supply the class C subnets we would like.

In other words urls like https://www.theyoungjerk.com:444 caused us some
grief. So what happens instead is that the visitor arrives at
http://www.example.com ok, but when he clicks home on the nav bar he
goes to www.example.com/shop/example

Here is a live example. You will need to replace /shop/example with
/cgi-bin/example:

                 <VirtualHost *:80>
                 DocumentRoot "/var/www/html/example"
                 ServerName www.example.com

SSLProxyEngine on
RewriteEngine on

RewriteRule ^\/Brahmi\/$  "http://www.example.com/PW-012.html" [P,L]
RewriteRule ^\/Acne\/$  "http://www.example.com/ET-036.html" [P,L]
RewriteRule ^\/Ginkgo\/$  "http://www.example.com/NA-010" [P,L]
RewriteRule ^\/Tryptophan\/$  "http://www.example.com/MM-001" [P,L]
RewriteRule ^\/Cod-Liver-Oil\/$  "http://www.example.com/TL-168" [P,L]
RewriteRule ^\/Tea-Tree-Oil\/$  "http://www.example.com/DE-007" [P,L]

RewriteRule ^\/Health\sConcerns\/(.*)$ 
"http://www.example.com/Health_Concerns/$1" [L]

RewriteRule ^\/Health_Concerns\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=5/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Weight\sLoss\/(.*)$ 
"http://www.example.com/Weight_Loss/$1" [L]


RewriteRule ^\/Weight_Loss\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=13/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Amino\sAcids\/(.*)$ 
"http://www.example.com/Amino_Acids/$1" [L]

RewriteRule ^\/Amino_Acids\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=1/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Fitness\sAccessories\/(.*)$ 
"http://www.example.com/Fitness_Accessories/$1" [L]

RewriteRule ^\/Fitness_Accessories\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=4/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Shakes\s&\sDrinks\/(.*)$ 
"http://www.example.com/Shakes_&_Drinks/$1" [L]

RewriteRule ^\/Shakes_&_Drinks\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=10/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Herbs\s&\sAlternative\sBlends\/(.*)$ 
"http://www.example.com/Herbs_&_Alternative_Blends/$1" [L]

RewriteRule ^\/Herbs_&_Alternative_Blends\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=7/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/EFAs\sand\sDietary\sOils\/(.*)$ 
"http://www.example.com/EFAs_and_Dietary_Oils/$1" [L]

RewriteRule ^\/EFAs_and_Dietary_Oils\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=3/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Sports\sNutrition\/(.*)$ 
"http://www.example.com/Sports_Nutrition/$1" [L]

RewriteRule ^\/Sports_Nutrition\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=11/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Vitamins\s&\sMinerals\/(.*)$ 
"http://www.example.com/Vitamins_&_Minerals/$1" [L]

RewriteRule ^\/Vitamins_&_Minerals\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=12/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Protein\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=9/va=banner_image=/va=banner_text=.html"
[P,L]


RewriteRule ^\/Creatine\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=2/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/Nutrition\sBars\/(.*)$ 
"http://www.example.com/Nutrition_Bars/$1" [L]

RewriteRule ^\/Nutrition_Bars\/$ 
"http://www.example.com/shop/example/scan/st=db/fi=cat/sp=imagecat/co=yes/sf=sel/ml=100/se=8/va=banner_image=/va=banner_text=.html"
[P,L]

RewriteRule ^\/([A-Z].*)\/(.*)\s(.*)\s(.*)\s(.*)\/$ 
"http://www.example.com/$1/$2_$3_$4_$5/" [L]
RewriteRule ^\/[A-Z].*\/([A-Z].*)_(.*)_(.*)_(.*)\/ 
"http://www.example.com/scan/st=db/fi=products/ml=100/sp=results/co=yes/sf=cat/se=$1
$2 $3 $4/va=banner_image=/va=banner_text=.html" [P,L]



RewriteRule ^\/([A-Z].*)\/(.*)\s(.*)\s(.*)\/$ 
"http://www.example.com/$1/$2_$3_$4/" [L]
RewriteRule ^\/[A-Z].*\/([A-Z].*)_(.*)_(.*)\/ 
"http://www.example.com/scan/st=db/fi=products/ml=100/sp=results/co=yes/sf=cat/se=$1
$2 $3/va=banner_image=/va=banner_text=.html" [P,L]


RewriteRule ^\/([A-Z].*)\/(.*)\s(.*)\/$ 
"http://www.example.com/$1/$2_$3/" [L]
RewriteRule ^\/[A-Z].*\/([A-Z].*)_(.*)\/ 
"http://www.example.com/scan/st=db/fi=products/ml=100/sp=results/co=yes/sf=cat/se=$1
$2/va=banner_image=/va=banner_text=.html" [P,L]

RewriteRule ^\/[A-Z].*\/([A-Z].*)\/ 
"http://www.example.com/scan/st=db/fi=products/ml=100/sp=results/co=yes/sf=cat/se=$1/va=banner_image=/va=banner_text=.html"
[P,L]

########################################################
RewriteRule ^$  /shop/example/index.html          [PT,L]
RewriteRule ^/$ /shop/example/index.html          [PT,L]
RewriteRule ^/index\.html$ /shop/example/index.htm [PT,L]
RewriteRule ^/shop/example/.* -                   [PT,L]
RewriteRule ^/.*images/.* -                            [PT,L]
RewriteRule ^/(.*) /shop/example/$1               [PT,L]

RewriteRule ^/admin$  /shop/example/admin/index.html          [PT,L]
RewriteRule ^/admin/$ /shop/example/admin/index.html          [PT,L]
RewriteRule ^/admin/index\.html$ /shop/example/admin/index.htm [PT,L]
RewriteRule ^/shop/example/admin/.* -                   [PT,L]
RewriteRule ^/admin/(.*) /shop/example/admin/$1               [PT,L]
         CustomLog /etc/httpd/logs/www.example.com/access_log
combined               
Redirect permanent /shop/example/arginine.html
"http://www.example.com/l_arginine_powder"
Redirect permanent /shop/example/arginine
"http://www.example.com/l_arginine_powder"


              </VirtualHost>

Regards,

Frank.
From racke at linuxia.de  Thu Dec 14 05:22:24 2006
From: racke at linuxia.de (Stefan Hornburg (Racke))
Date: Thu Dec 14 05:23:03 2006
Subject: [ic] Admin New Entry Problem
In-Reply-To: <457E5370.30302@pajamian.dhs.org>
References: <00bc01c71dac$860603b0$0401a8c0@PSLAPTOP1A>
	<457E5370.30302@pajamian.dhs.org>
Message-ID: <458125E0.6070202@linuxia.de>

Peter wrote:
> On 12/11/2006 09:15 PM, Cameron B. Prince wrote:
>> I'm having a problem with the Admin's New Entry function. This is using a
>> default standard catalog on IC v5.4.1 with MySQL. When you edit a table that
>> uses the autonumber counter files as a key, such as products, you can
>> duplicate the problem by clicking New Entry. Leave the pre-filled key intact
>> and click Ok to create the new entry. You will see that you are returned to
>> the flex_editor but instead of it having the next key pre-filled, all the
>> fields are empty.
>>
>> You can click New Entry again to create another item, but you will find that
>> the counter has been incremented twice. I updated the meta for the New Entry
>> menu item and added a ui_return_to=admin/flex_select to cause the Ok to take
>> you back to the edit page rather than having the empty flex_editor but was
>> surprised to find the counter was still incremented twice after creating a
>> new entry.
> 
> I've seen this exact issue and never had the time to look into it.  If I
> get the time I'll be happy to look into it, or if someone either (1)
> finds the problem themselves and gives me an idea where to look so I
> don't have to spend all day and a half tracking it down or (2) pays me
> to do it then I'll be likely to commit a fix for it in the near future,
> otherwise may not happen for a while (I'm pretty swamped with other
> things that take priority).  That's speaking for me, someone else may
> fix it.
> 
> FWIW this problem goes back at least as far as 5.2.0 and is not specific
> to MySQL (I get the same problem with PostgreSQL).  It's not specific to
> standard, either, I see it in Foundation-based sites.
> 
>> 1) Is there a logical reason that I'm not realizing for incrementing the
>> counter twice? If not, how can this be prevented?
> 
> Probably not, it's just a minor bug.

The counter is really incremented twice:

1. flex_editor

                 unless ($db->config('AUTO_SEQUENCE')) {
			$db->config('AUTO_NUMBER', '000001')
				unless $db->config('AUTO_NUMBER');
			$CGI->{item_id} = $db->autonumber();
		}

2. [table-editor]

	if($opt->{ui_new_item} and ! $opt->{notable}) {
		if( ! $db->config('_Auto_number') and ! $db->config('AUTO_SEQUENCE')) {
			$db->config('AUTO_NUMBER', '000001');
			$key = $db->autonumber($key);
		}
		else {
			$key = '';
			$opt->{mv_data_auto_number} = 1;
			$key_message = errmsg('(new key will be assigned if left blank)');
		}
	}

The code looks slightly different, so I wonder whether we can just drop the
code in flex_editor without regressions. It appears superfluous to me.

Bye
	Racke

-- 
LinuXia Systems => http://www.linuxia.de/
Expert Interchange Consulting and System Administration
ICDEVGROUP => http://www.icdevgroup.org/
Interchange Development Team

From emailgrant at gmail.com  Thu Dec 14 11:17:58 2006
From: emailgrant at gmail.com (Grant)
Date: Thu Dec 14 11:18:18 2006
Subject: [ic] Bad HouseKeeping
Message-ID: <49bf44f10612140817v6c6db45cn657eb07fb81f5c56@mail.gmail.com>

I'm using IC 5.2 and ps -ef shows three interchange processes that
started some time yesterday and two that started an hour and a half
ago which is the last time I restarted IC.  Here are my
interchange.cfg settings:

PreFork                 Yes
StartServers            5
MaxServers              50
MaxRequestsPerChild     100
ChildLife               600
HouseKeeping            2
PIDcheck                120

Shouldn't housekeeping clean up the above-mentioned five processes?
There are currently no jobs running.

Should each IC process be using about 30MB?

- Grant
From barb at emeraldcitymusic.com  Thu Dec 14 13:54:43 2006
From: barb at emeraldcitymusic.com (Barb Sorensen)
Date: Thu Dec 14 13:53:38 2006
Subject: [ic] Need help with install please. Newbie here...
Message-ID: <C1A6FA13.38E9D%barb@emeraldcitymusic.com>

HI, 
I am not a programmer or a geek guru but just hack my way through things
85% of the time I know what I"m doing  but this telnet stuff is new to
me.....

I downloaded, unzipped and uploaded the interchange package to my server.
 I am on 1and1.com and have a Home account (no ssh and no telnet apparently
as far as I can tell - I did not know I needed it to install interchange and
I just switched form my old host to 1and1.com and am beginning to think I
made am istake and hoping to salvage it somehow ..)

When I opened the read me file it says to run Makefile.PL..  Which I did and
it says it's installing Interchange

 but then the next instructions say to run "make" and some other commands
but that is where I get stuck..
 I know .. I haven't even left the starting gate and I"m excited about
getting there already LOL!

When I run "make" I get an error 404: not found message

Also.. The pages look like text pages not a pretty front. Is that supposed
to be that way?  Just wanting to be sure I am doing this correctly..

So then somewhere I stumbled across something that says I need telnet to
install it.. 
Is there a way to install without telnet?

My "home" package with 1and1.com says that instead of using tenet they use
ssh. And if I want to use the SSH I have to upgrade to the next bigger
"business" package  which is what I am trying to avoid doing as well.

Has anyone here on this list installed in a Home package on 1and1.com or
know how to go about this?

Thanks
Regards
Barb
http://www.barbsorensen.com

From racke at linuxia.de  Thu Dec 14 14:56:55 2006
From: racke at linuxia.de (Stefan Hornburg (Racke))
Date: Thu Dec 14 14:57:40 2006
Subject: [ic] Need help with install please. Newbie here...
In-Reply-To: <C1A6FA13.38E9D%barb@emeraldcitymusic.com>
References: <C1A6FA13.38E9D%barb@emeraldcitymusic.com>
Message-ID: <4581AC87.4090402@linuxia.de>

Barb Sorensen wrote:
> HI, 
> I am not a programmer or a geek guru but just hack my way through things
> 85% of the time I know what I"m doing  but this telnet stuff is new to
> me.....
> 
> I downloaded, unzipped and uploaded the interchange package to my server.
>  I am on 1and1.com and have a Home account (no ssh and no telnet apparently
> as far as I can tell - I did not know I needed it to install interchange and
> I just switched form my old host to 1and1.com and am beginning to think I
> made am istake and hoping to salvage it somehow ..)

First of all check with your ISP if you allowed to run a daemon (server
program) on your account.

Second, you really need a shell access to the server.

> 
> When I opened the read me file it says to run Makefile.PL..  Which I did and
> it says it's installing Interchange
> 
>  but then the next instructions say to run "make" and some other commands
> but that is where I get stuck..
>  I know .. I haven't even left the starting gate and I"m excited about
> getting there already LOL!
> 
> When I run "make" I get an error 404: not found message

You need to run these commands on as shell commands.

Bye
	Racke

-- 
LinuXia Systems => http://www.linuxia.de/
Expert Interchange Consulting and System Administration
ICDEVGROUP => http://www.icdevgroup.org/
Interchange Development Team

From barb at emeraldcitymusic.com  Fri Dec 15 00:57:13 2006
From: barb at emeraldcitymusic.com (Barb Sorensen)
Date: Fri Dec 15 00:55:53 2006
Subject: [ic] FW: Need help with install please. Newbie here...
In-Reply-To: <C1A6FA13.38E9D%barb@emeraldcitymusic.com>
Message-ID: <C1A79559.3D3F2%barb@emeraldcitymusic.com>

HI Stefan.  Thanks
 OK I broke down and upgraded my server to "business" so I could have the
ssh access..
Firs thing I did was go to the ftp address.
 it asked me for a password and I entered it and then it gave me a list of
all the files in the directory for interchange. :)
Looks Like I have access
 then I clicked on the configure file since the README instructions day to
do that first
  then I clicked on Makefile.pl as per the instructions after the
./configure file ran it's course.
and it again asked me for user and password and then it too ran it's course,
installed to /interchange and said "done"

 then I wasn't sure where to go but I remembered that after I ran the
configure file that the instructions at that point said to also run
Makefuile/.pl ( which I did ) and then run "make" and then run "make test"
and then run " make install"

 so I type in /interchange/make and it began to churn away .... And it's
been going for a little while.. Should it take this long, or is it waiting
for another command?  I ddid not see a "make" file in the folder.. And I"m
not that familiar with ssh commands and perl... So not sure if this is right
or did Miss something?:
Where to now? It's still on "make".. Toiling away ...

Thanks
Regards
Barb

------ Forwarded Message
From: Barb Sorensen <barb@emeraldcitymusic.com>
Date: Thu, 14 Dec 2006 12:54:43 -0600
To: <interchange-users@icdevgroup.org>
Subject: Need help with install please. Newbie here...

HI, 
I am not a programmer or a geek guru but just hack my way through things
85% of the time I know what I"m doing  but this telnet stuff is new to
me.....

I downloaded, unzipped and uploaded the interchange package to my server.
 I am on 1and1.com and have a Home account (no ssh and no telnet apparently
as far as I can tell - I did not know I needed it to install interchange and
I just switched form my old host to 1and1.com and am beginning to think I
made am istake and hoping to salvage it somehow ..)

When I opened the read me file it says to run Makefile.PL..  Which I did and
it says it's installing Interchange

 but then the next instructions say to run "make" and some other commands
but that is where I get stuck..
 I know .. I haven't even left the starting gate and I"m excited about
getting there already LOL!

When I run "make" I get an error 404: not found message

Also.. The pages look like text pages not a pretty front. Is that supposed
to be that way?  Just wanting to be sure I am doing this correctly..

So then somewhere I stumbled across something that says I need telnet to
install it.. 
Is there a way to install without telnet?

My "home" package with 1and1.com says that instead of using tenet they use
ssh. And if I want to use the SSH I have to upgrade to the next bigger
"business" package  which is what I am trying to avoid doing as well.

Has anyone here on this list installed in a Home package on 1and1.com or
know how to go about this?

Thanks
Regards
Barb
http://www.barbsorensen.com

------ End of Forwarded Message

From peter at pajamian.dhs.org  Fri Dec 15 04:27:38 2006
From: peter at pajamian.dhs.org (Peter)
Date: Fri Dec 15 04:28:13 2006
Subject: [ic] FW: Need help with install please. Newbie here...
In-Reply-To: <C1A79559.3D3F2%barb@emeraldcitymusic.com>
References: <C1A79559.3D3F2%barb@emeraldcitymusic.com>
Message-ID: <45826A8A.6040200@pajamian.dhs.org>

On 12/14/2006 09:57 PM, Barb Sorensen wrote:
> HI Stefan.  Thanks
>  OK I broke down and upgraded my server to "business" so I could have the
> ssh access..
> Firs thing I did was go to the ftp address.
>  it asked me for a password and I entered it and then it gave me a list of
> all the files in the directory for interchange. :)
> Looks Like I have access
>  then I clicked on the configure file since the README instructions day to
> do that first
>   then I clicked on Makefile.pl as per the instructions after the
> ./configure file ran it's course.
> and it again asked me for user and password and then it too ran it's course,
> installed to /interchange and said "done"
> 
>  then I wasn't sure where to go but I remembered that after I ran the
> configure file that the instructions at that point said to also run
> Makefuile/.pl ( which I did ) and then run "make" and then run "make test"
> and then run " make install"
> 
>  so I type in /interchange/make and it began to churn away .... And it's
> been going for a little while.. Should it take this long, or is it waiting
> for another command?  I ddid not see a "make" file in the folder.. And I"m
> not that familiar with ssh commands and perl... So not sure if this is right
> or did Miss something?:
> Where to now? It's still on "make".. Toiling away ...

Hi Barb,

I have recently written up a general overview about how to install
programs in Linux (and other Unixes) from source for a forum that I
frequent, it may help you to read through it.  You can find it at
<http://www.groklaw.net/comment.php?mode=display&sid=2006112417153768&title=Basics%20of%20compiling%20a%20program%20to%20run%20in%20Linux&type=article&order=&hideanonymous=0&pid=508554#c508658>.
 You may want to go back to the beginning of the thread at
<http://www.groklaw.net/comment.php?mode=display&sid=2006112417153768&title=OT%20Here&type=article&order=&hideanonymous=0&pid=508360#c508554>
and read the entire thread.

Installing and running Interchange generally requires at least a basic
knowledge of Unix command line shell concepts.  There are lots of
consultants who will gladly help you out for a modest fee if you find
that you can't handle it yourself.

Also, please don't top post per the rules and guidelines of this mailing
list, see <http://www.icdevgroup.org/mailman/listinfo/interchange-users>
for more info.

Regards,

Peter

From emailgrant at gmail.com  Fri Dec 15 11:02:28 2006
From: emailgrant at gmail.com (Grant)
Date: Fri Dec 15 11:03:05 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612130707m6b693904waac5228452fb609b@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>
	<49bf44f10612130707m6b693904waac5228452fb609b@mail.gmail.com>
Message-ID: <49bf44f10612150802s449fa7c3y7fbca8dbd00abc50@mail.gmail.com>

> > > >> > Hello, I've been plagued by apache2 segfaults ever since I started
> > > >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> > > >> > messages accompanying the segfaults in error_log:
> > > >> >
> > > >> > ALERT: bad pipe signal received for /page.html
> > > >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> > > >> > Segmentation fault (11)
> > > >> >
> > > >> > Does anyone have any advice on solving this?  I'm using
> > > >> apache-2.0.58
> > > >> > and mod_perl-2.0.2 in Gentoo Linux.
> > > >>
> > > >> Also, here is the portion of Link.pm that the ALERT seems to come
> > > >> from:
> > > >>
> > > >> # Return this message to the browser when the server is not running.
> > > >> # Log an error log entry if set to notify
> > > >>
> > > >> sub die_page {
> > > >>
> > > >>     my $r = shift;
> > > >>     my $msg;
> > > >>
> > > >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}\n";
> > > >>
> > > >>     $r->content_type ("text/html");
> > > >>     $r->print (<<EOF);
> > > >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> > > >> <BODY BGCOLOR="#FFFFFF">
> > > >> <H3>Someone pressed stop...</H3>
> > > >> <P>
> > > >> We have aborted this request because someone terminated it.
> > > >> Please try again soon.
> > > >> </BODY></HTML>
> > > >> EOF
> > > >>
> > > >> }
> > > >>
> > > >> Please let me know if you have any ideas.
> > > >
> > > > The segfaults are eliminated by commenting out the $r stuff in the
> > > > die_page sub.  I still get the ALERTs though.  Does anyone have any
> > > > advice on figuring out why I'm having the bad pipe problem?  Is there
> > > > an easy way to add extra debugging info to the sub?
> > > >
> > > > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
> > > > 50 fold.
> > >
> > > I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> > > had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> > >
> > > The visible effect on the browser is that the page or image (which
> > > Link.pm apparently still has some part in delivering) does not load.
> > > I get them myself when browsing and testing my websites, and I have
> > > never stopped loading a page or had any other problems on non-IC
> > > sites I host.
> > >
> > > I was told the problem stems from either the browser and a stop
> > > button or some other network fault. I may go back to Apache 1.3 to
> > > get around this.
> >
> > I've been working on this all day and I think I may have a solution.
> >
> > Incidentally, I want to mention that if I add $! to the Link.pm warn
> > line, I can see that there are two different types of ALERTs in
> > error_log:
> >
> > Broken pipe
> > Inappropriate ioctl for device
> >
> > Also incidentally, hitting F5 repeatedly always prints a few ALERTS in
> > error_log, but the number of ALERTs printed seems to be about 10 times
> > more for an http page than for the same page accessed via https.  I
> > checked and tested the settings in my *:80 and *:443 vhosts trying to
> > narrow that down, but I didn't come up with anything.
> >
> > Now, as I mentioned a few posts back, commenting out the html delivery
> > stuff in the die_page sub of Link.pm eliminates the segfaults in
> > error_log but not the ALERTs.  Today I enabled the prefork mpm in
> > apache2 and tinkered with the Server/Child settings in interchange.cfg
> > and httpd.conf and that did seem to reduce the ALERTs somewhat.
> >
> > After all of this, I've been browsing around my site and I haven't
> > seen a single image or page fail to load like it used to.  The thing
> > is, ALERTs still show up in error_log even for requests I'm sure I
> > created myself AND WERE SERVED SUCCESSFULLY.  I hypothesize that the
> > failed requests were because a $SIG{PIPE} was generated for whatever
> > reason (although not because the user clicked the stop button), the
> > html was delivered, it caused a segfault, and the request failed.  I
> > should mention though, that if the failing request syndrome is fixed,
> > it could also be from the apache2, httpd.conf, or interchange.cfg
> > modifications mentioned above.  I doubt it though.
> >
> > Incidentally, I'd like to mention that the ALERTs can be generated
> > pretty reliably by hitting the browser's stop button in the middle of
> > a request, clicking on the same link again or a different link in the
> > middle of a request (my Mom probably still double clicks links), or
> > just refreshing the page in the middle of a request.
> >
> > If the failed requests do stop at this point, I think the die_page
> > section of Link.pm is doing more harm than good.  Is there anything
> > wrong with not handling $SIG{PIPE} at all?
> >
> > - Grant
>
> I still have yet to see a single failed image or page request in the
> browser since the above changes, although ALERTs are still generated
> in error_log.  Can anyone verify this?
>
> - Grant

I'm sure this is fixed.  I propose the die_page call and sub are
removed from Link.pm in CVS.

- Grant
From josh at myprivacy.ca  Fri Dec 15 12:07:15 2006
From: josh at myprivacy.ca (Josh Lavin)
Date: Fri Dec 15 12:08:05 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>
Message-ID: <21375B2F-06D8-4614-98D5-1B28FCB26E47@myprivacy.ca>

On Dec 12, 2006, at 5:43 PM, Grant wrote:
>> >> > Hello, I've been plagued by apache2 segfaults ever since I  
>> started
>> >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
>> >> > messages accompanying the segfaults in error_log:
>> >> >
>> >> > ALERT: bad pipe signal received for /page.html
>> >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
>> >> > Segmentation fault (11)
>> >> >
>> >> > Does anyone have any advice on solving this?  I'm using
>> >> apache-2.0.58
>> >> > and mod_perl-2.0.2 in Gentoo Linux.
>> >>
>> >> Also, here is the portion of Link.pm that the ALERT seems to come
>> >> from:
>> >>
>> >> # Return this message to the browser when the server is not  
>> running.
>> >> # Log an error log entry if set to notify
>> >>
>> >> sub die_page {
>> >>
>> >>     my $r = shift;
>> >>     my $msg;
>> >>
>> >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME} 
>> \n";
>> >>
>> >>     $r->content_type ("text/html");
>> >>     $r->print (<<EOF);
>> >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
>> >> <BODY BGCOLOR="#FFFFFF">
>> >> <H3>Someone pressed stop...</H3>
>> >> <P>
>> >> We have aborted this request because someone terminated it.
>> >> Please try again soon.
>> >> </BODY></HTML>
>> >> EOF
>> >>
>> >> }
>> >>
>> >> Please let me know if you have any ideas.
>> >
>> > The segfaults are eliminated by commenting out the $r stuff in the
>> > die_page sub.  I still get the ALERTs though.  Does anyone have any
>> > advice on figuring out why I'm having the bad pipe problem?  Is  
>> there
>> > an easy way to add extra debugging info to the sub?
>> >
>> > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
>> > 50 fold.
>>
>> I've been seeing this too, on my Apache 2 and latest Link.pm. I also
>> had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
>>
>> The visible effect on the browser is that the page or image (which
>> Link.pm apparently still has some part in delivering) does not load.
>> I get them myself when browsing and testing my websites, and I have
>> never stopped loading a page or had any other problems on non-IC
>> sites I host.
>>
>> I was told the problem stems from either the browser and a stop
>> button or some other network fault. I may go back to Apache 1.3 to
>> get around this.
>
> I've been working on this all day and I think I may have a solution.
>
> Incidentally, I want to mention that if I add $! to the Link.pm warn
> line, I can see that there are two different types of ALERTs in
> error_log:
>
> Broken pipe
> Inappropriate ioctl for device
>
> Also incidentally, hitting F5 repeatedly always prints a few ALERTS in
> error_log, but the number of ALERTs printed seems to be about 10 times
> more for an http page than for the same page accessed via https.  I
> checked and tested the settings in my *:80 and *:443 vhosts trying to
> narrow that down, but I didn't come up with anything.
>
> Now, as I mentioned a few posts back, commenting out the html delivery
> stuff in the die_page sub of Link.pm eliminates the segfaults in
> error_log but not the ALERTs.  Today I enabled the prefork mpm in
> apache2 and tinkered with the Server/Child settings in interchange.cfg
> and httpd.conf and that did seem to reduce the ALERTs somewhat.
>
> After all of this, I've been browsing around my site and I haven't
> seen a single image or page fail to load like it used to.  The thing
> is, ALERTs still show up in error_log even for requests I'm sure I
> created myself AND WERE SERVED SUCCESSFULLY.  I hypothesize that the
> failed requests were because a $SIG{PIPE} was generated for whatever
> reason (although not because the user clicked the stop button), the
> html was delivered, it caused a segfault, and the request failed.  I
> should mention though, that if the failing request syndrome is fixed,
> it could also be from the apache2, httpd.conf, or interchange.cfg
> modifications mentioned above.  I doubt it though.
>
> Incidentally, I'd like to mention that the ALERTs can be generated
> pretty reliably by hitting the browser's stop button in the middle of
> a request, clicking on the same link again or a different link in the
> middle of a request (my Mom probably still double clicks links), or
> just refreshing the page in the middle of a request.
>
> If the failed requests do stop at this point, I think the die_page
> section of Link.pm is doing more harm than good.  Is there anything
> wrong with not handling $SIG{PIPE} at all?

Grant,

Shouldn't the $SIG{PIPE} be ignored at least? I don't know if there  
are implications for not handling it. Search perlipc manpage for PIPE.

--
Josh Lavin
Kingdom Design   http://www.kingdomdesign.com/

From emailgrant at gmail.com  Fri Dec 15 12:35:43 2006
From: emailgrant at gmail.com (Grant)
Date: Fri Dec 15 12:35:58 2006
Subject: [ic] Re: ALERT: bad pipe signal received for /page.html
In-Reply-To: <21375B2F-06D8-4614-98D5-1B28FCB26E47@myprivacy.ca>
References: <49bf44f10612091034n7e7a5ee7g34fad9eb5a1954a8@mail.gmail.com>
	<49bf44f10612091037hce3fa80m76077e0be4632510@mail.gmail.com>
	<49bf44f10612111002l231637f2r22b8370e5534457@mail.gmail.com>
	<183A6A5F-82D3-4C02-BCEA-6D645C96C2BB@myprivacy.ca>
	<49bf44f10612121543t348d3112i690232821392fed6@mail.gmail.com>
	<21375B2F-06D8-4614-98D5-1B28FCB26E47@myprivacy.ca>
Message-ID: <49bf44f10612150935xaa7e17bm6373cd84acb58635@mail.gmail.com>

> >> >> > Hello, I've been plagued by apache2 segfaults ever since I
> >> started
> >> >> > using Interchange::Link years ago.  The latest Link.pm has ALERT
> >> >> > messages accompanying the segfaults in error_log:
> >> >> >
> >> >> > ALERT: bad pipe signal received for /page.html
> >> >> > [Sat Dec 09 10:27:55 2006] [notice] child pid 21337 exit signal
> >> >> > Segmentation fault (11)
> >> >> >
> >> >> > Does anyone have any advice on solving this?  I'm using
> >> >> apache-2.0.58
> >> >> > and mod_perl-2.0.2 in Gentoo Linux.
> >> >>
> >> >> Also, here is the portion of Link.pm that the ALERT seems to come
> >> >> from:
> >> >>
> >> >> # Return this message to the browser when the server is not
> >> running.
> >> >> # Log an error log entry if set to notify
> >> >>
> >> >> sub die_page {
> >> >>
> >> >>     my $r = shift;
> >> >>     my $msg;
> >> >>
> >> >>     warn "ALERT: bad pipe signal received for $ENV{SCRIPT_NAME}
> >> \n";
> >> >>
> >> >>     $r->content_type ("text/html");
> >> >>     $r->print (<<EOF);
> >> >> <HTML><HEAD><TITLE>Interrupted</TITLE></HEAD>
> >> >> <BODY BGCOLOR="#FFFFFF">
> >> >> <H3>Someone pressed stop...</H3>
> >> >> <P>
> >> >> We have aborted this request because someone terminated it.
> >> >> Please try again soon.
> >> >> </BODY></HTML>
> >> >> EOF
> >> >>
> >> >> }
> >> >>
> >> >> Please let me know if you have any ideas.
> >> >
> >> > The segfaults are eliminated by commenting out the $r stuff in the
> >> > die_page sub.  I still get the ALERTs though.  Does anyone have any
> >> > advice on figuring out why I'm having the bad pipe problem?  Is
> >> there
> >> > an easy way to add extra debugging info to the sub?
> >> >
> >> > Also, restarting IC with PERL_SIGNALS=unsafe increases the ALERTs
> >> > 50 fold.
> >>
> >> I've been seeing this too, on my Apache 2 and latest Link.pm. I also
> >> had to use PERL_SIGNALS=unsafe and so I get quite a lot of these.
> >>
> >> The visible effect on the browser is that the page or image (which
> >> Link.pm apparently still has some part in delivering) does not load.
> >> I get them myself when browsing and testing my websites, and I have
> >> never stopped loading a page or had any other problems on non-IC
> >> sites I host.
> >>
> >> I was told the problem stems from either the browser and a stop
> >> button or some other network fault. I may go back to Apache 1.3 to
> >> get around this.
> >
> > I've been working on this all day and I think I may have a solution.
> >
> > Incidentally, I want to mention that if I add $! to the Link.pm warn
> > line, I can see that there are two different types of ALERTs in
> > error_log:
> >
> > Broken pipe
> > Inappropriate ioctl for device
> >
> > Also incidentally, hitting F5 repeatedly always prints a few ALERTS in
> > error_log, but the number of ALERTs printed seems to be about 10 times
> > more for an http page than for the same page accessed via https.  I
> > checked and tested the settings in my *:80 and *:443 vhosts trying to
> > narrow that down, but I didn't come up with anything.
> >
> > Now, as I mentioned a few posts back, commenting out the html delivery
> > stuff in the die_page sub of Link.pm eliminates the segfaults in
> > error_log but not the ALERTs.  Today I enabled the prefork mpm in
> > apache2 and tinkered with the Server/Child settings in interchange.cfg
> > and httpd.conf and that did seem to reduce the ALERTs somewhat.
> >
> > After all of this, I've been browsing around my site and I haven't
> > seen a single image or page fail to load like it used to.  The thing
> > is, ALERTs still show up in error_log even for requests I'm sure I
> > created myself AND WERE SERVED SUCCESSFULLY.  I hypothesize that the
> > failed requests were because a $SIG{PIPE} was generated for whatever
> > reason (although not because the user clicked the stop button), the
> > html was delivered, it caused a segfault, and the request failed.  I
> > should mention though, that if the failing request syndrome is fixed,
> > it could also be from the apache2, httpd.conf, or interchange.cfg
> > modifications mentioned above.  I doubt it though.
> >
> > Incidentally, I'd like to mention that the ALERTs can be generated
> > pretty reliably by hitting the browser's stop button in the middle of
> > a request, clicking on the same link again or a different link in the
> > middle of a request (my Mom probably still double clicks links), or
> > just refreshing the page in the middle of a request.
> >
> > If the failed requests do stop at this point, I think the die_page
> > section of Link.pm is doing more harm than good.  Is there anything
> > wrong with not handling $SIG{PIPE} at all?
>
> Grant,
>
> Shouldn't the $SIG{PIPE} be ignored at least? I don't know if there
> are implications for not handling it. Search perlipc manpage for PIPE.

I see the section you're referring to in the perlipc manpage, but I
don't know enough about perl to interpret it.

       Be careful to check both the open() and the close() return values.  If
       you're writing to a pipe, you should also trap SIGPIPE.  Otherwise,
       think of what happens when you start up a pipe to a command that
       doesn't exist: the open() will in all likelihood succeed (it only
       reflects the fork()'s success), but then your output will fail--spec-
       tacularly.  Perl can't know whether the command worked because your
       command is actually running in a separate process whose exec() might
       have failed.  Therefore, while readers of bogus commands return just a
       quick end of file, writers to bogus command will trigger a signal
       they'd better be prepared to handle.  Consider:

           open(FH, "|bogus")  or die "can't fork: $!";
           print FH "bang\n"   or die "can't write: $!";
           close FH            or die "can't close: $!";

       That won't blow up until the close, and it will blow up with a SIGPIPE.
       To catch it, you could use this:

           $SIG{PIPE} = 'IGNORE';
           open(FH, "|bogus")  or die "can't fork: $!";
           print FH "bang\n"   or die "can't write: $!";
           close FH            or die "can't close: status=$?";

- Grant
From barb at emeraldcitymusic.com  Fri Dec 15 15:02:34 2006
From: barb at emeraldcitymusic.com (Barb Sorensen)
Date: Fri Dec 15 15:01:22 2006
Subject: [ic] Can't locate set cronjob (was - Need help with install please.
	Newbie here) 
In-Reply-To: <C1A6FA13.38E9D%barb@emeraldcitymusic.com>
Message-ID: <C1A85B7A.3D451%barb@emeraldcitymusic.com>

HI, Thanks..that helped tons.
 
1) I'm retitling this email subject line to be easier for other folks to
find their error message help .

2) I"m sorry for "Top posting" but I didn't know what that means. I have not
heard that phrase before But from the link you posted I"m guessing that
means not to repost the full emails I already wrote.. At the top of a new
post? 
So I won't. Thanks for alerting me to this.

3) OK . In re: to this install...

Thanks for the links.. They helped tons!
I recognized I have an Application on my MAC named Terminal.
And it offered ssh and so I entered my info and it seemed to connect to the
server :) 
I uploaded version 5.4.1 stable release.

Then I typed in ./configure and it ran ok
 then I typed in perl Makefile.PL and it ran and asked me where to install
it. I typed in the correct address and hit go.

It gave me this info next:

Writing Makefile for Interchange
~/websites/barbsboutique/interchange >

I then typed in "make" and hit go
And it seemed to run fine.

 then I typed in "make test && make install"
 
It began to run and then I got this
"Can't locate Set/Crontab.pm in @INC"
BEGIN failed--compilation aborted at
Compilation failed in require at
Compilation failed in require at blib/script/interchange line 249.
BEGIN failed--compilation aborted at blib/script/interchange line 249.
not ok 1
server/startup........-rw-r--r--



# When the above test fails, it may be due to your ISP or some other
# mechanism blocking port 8786.


SO I deleted everything and instead of me adjusting permissions myself I
uploaded the tarball.bz2 and ran it and yet by the time I get to " make
test" it stil gives me the same error.

Any ideas?

Thanks
Regards
Barb

From docelic at mail.inet.hr  Fri Dec 15 15:16:45 2006
From: docelic at mail.inet.hr (Davor Ocelic)
Date: Fri Dec 15 15:18:41 2006
Subject: [ic] Can't locate set cronjob (was - Need help with install
	please. Newbie here)
In-Reply-To: <C1A85B7A.3D451%barb@emeraldcitymusic.com>
References: <C1A6FA13.38E9D%barb@emeraldcitymusic.com>
	<C1A85B7A.3D451%barb@emeraldcitymusic.com>
Message-ID: <20061215211645.3d003003.docelic@mail.inet.hr>

On Fri, 15 Dec 2006 14:02:34 -0600
Barb Sorensen <barb@emeraldcitymusic.com> wrote:

> It began to run and then I got this
> "Can't locate Set/Crontab.pm in @INC"
> BEGIN failed--compilation aborted at
> Compilation failed in require at
> Compilation failed in require at blib/script/interchange line 249.
> BEGIN failed--compilation aborted at blib/script/interchange line 249.
> not ok 1
> server/startup........-rw-r--r--
> 
> 

Hey Barb. This means you need to install the Set::Crontab Perl module.
The common way to do this is to run 

	perl -MCPAN -e 'install "Set::Crontab"'

in your shell (at the place where you'll be running Interchange).

Since this will most likely be the first time you are invoking CPAN
there, CPAN will ask you a series of questions to configure itself.
You should find a way about it since just accepting the defaults
works most of the time.

Once CPAN is configured, it will download the said Perl module,
compile and install it. Then resume your installation procedure from
and including the command that gave the error onwards.

-doc
From interchange at hertell.com  Fri Dec 15 16:57:31 2006
From: interchange at hertell.com (Rene Hertell)
Date: Fri Dec 15 16:58:00 2006
Subject: [ic] Can't locate set cronjob (was - Need help with install
	please. Newbie here)
In-Reply-To: <C1A85B7A.3D451%barb@emeraldcitymusic.com>
References: <C1A85B7A.3D451%barb@emeraldcitymusic.com>
Message-ID: <45831A4B.4000006@hertell.com>

Barb Sorensen wrote:
> HI, Thanks..that helped tons.
>  
> 1) I'm retitling this email subject line to be easier for other folks to
> find their error message help .

Please start a new thread if you want to post a new message.. Jut simply
renaming a subject breaks the threads of messages

If your mail-client does not support threaded views, then check your
messages from the achieves and you'll see what i mean:

http://www.icdevgroup.org/pipermail/interchange-users/2006-December/thread.html#end


Ren?
From peter at pajamian.dhs.org  Fri Dec 15 17:16:32 2006
From: peter at pajamian.dhs.org (Peter)
Date: Fri Dec 15 17:17:03 2006
Subject: [ic] Can't locate set cronjob (was - Need help with install
	please. Newbie here)
In-Reply-To: <C1A85B7A.3D451%barb@emeraldcitymusic.com>
References: <C1A85B7A.3D451%barb@emeraldcitymusic.com>
Message-ID: <45831EC0.9090102@pajamian.dhs.org>

On 12/15/2006 12:02 PM, Barb Sorensen wrote:
> 2) I"m sorry for "Top posting" but I didn't know what that means. I have not
> heard that phrase before But from the link you posted I"m guessing that
> means not to repost the full emails I already wrote.. At the top of a new
> post? 

No, top posting means that you post your reply above the text you're
replying to.  This is considered bad form.  The proper way to reply is
to add your reply text in after each point made in the text as I'm doing
here.  That way when people read the email in the archives later on they
can easily follow each issue included there and the full conversation
leading to a resolution and don't have to try to match up which parts go
with what.

>  then I typed in "make test && make install"
>  
> It began to run and then I got this
> "Can't locate Set/Crontab.pm in @INC"

This means that your perl is missing a module that Interchange needs to
run.  There are several perl modules that usually need to be installed
prior to running Interchange on a server.  You should contact your
hosting provider and ask them to install the
"Bundle::InterchangeKitchenSink" bundle onto the server for you.  They
will probably do it for you if you ask.  if they won't then you'll have
to install them into your own space on the server as a non-root user,
google for "installing perl modules as non-root" for lots of web pages
telling you how to do so.

Peter

From peter at pajamian.dhs.org  Fri Dec 15 17:23:26 2006
From: peter at pajamian.dhs.org (Peter)
Date: Fri Dec 15 17:23:41 2006
Subject: [ic] Can't locate set cronjob (was - Need help with
	install	please. Newbie here)
In-Reply-To: <45831A4B.4000006@hertell.com>
References: <C1A85B7A.3D451%barb@emeraldcitymusic.com>
	<45831A4B.4000006@hertell.com>
Message-ID: <4583205E.8080200@pajamian.dhs.org>

On 12/15/2006 01:57 PM, Rene Hertell wrote:
> Please start a new thread if you want to post a new message.. Jut simply
> renaming a subject breaks the threads of messages
> 
> If your mail-client does not support threaded views, then check your
> messages from the achieves and you'll see what i mean:
> 
> http://www.icdevgroup.org/pipermail/interchange-users/2006-December/thread.html#end


What that is referring to is that a lot of people will "hijack" a
message thread to post their own question.  This might happen if, for
example someone hit the reply button on this message, deleted all the
quoted text, typed in their own problem and hit send.  Two things would
happen...

1.  Email clients keep track of what messages were replied to in the
headers of the email.  This allows a mail reader to organise the emails
by subject and show where in the "thread" that email occurred.  People
who use that feature will see a completely different email appearing
inside someone else's thread and it becomes annoying.

2.  If the subject isn't changed then it's also an additional level of
annoyance because the subject says one thing and the email is about
something completely different.

As a rule of thumb when you are starting a new discussion compose a new
email, don't use the reply button unless you are genuinely replying to
an existing email.

Peter
From list at candhdevelopment.com  Sat Dec 16 16:19:44 2006
From: list at candhdevelopment.com (list)
Date: Sat Dec 16 16:20:14 2006
Subject: [ic] (no subject)
Message-ID: <200612161619601.SM01052@c-e18fdc26944b4>

Hello, interchange-users,


Killing Interchange server 4315 with TERM.
Low traffic settings.
Calling UI......UI is loaded...
Interchange V5.4.1
Running with new signals, external programs could be unreliable.
Re-run with environment variable PERL_SIGNALS set to "unsafe" to change this.
Configuring catalog demo...demo config error: Please specify the VendURL directive in the configuration file 'catalog.cfg'



demo: error in configuration. Skipping.
demo: config error. Skipping.
Interchange server started in UNIX mode(s) (process id 4426)

-------------------

I guessed that it 'must' be ( /home/ddws/catalogs/demo/catalog.cgf ) that the script was talking about?


# The URLs which are written to refer back to our catalog.

<ParseVariables Yes>

VendURL    http://__SERVER_NAME____CGI_URL__
SecureURL  __SECURE_SERVER____CGI_URL__

ifndef SECURE_ENABLE
SecureURL  http://__SERVER_NAME____CGI_URL__
endif

# Set the image path for relative images
ImageDir          __IMAGE_DIR__/
ImageDirInternal  http://__SERVER_NAME____IMAGE_DIR__/

</ParseVariables>

 Could there be "something" not allowing _SERVER_NAME_CGI_URL_ to work?



Best regards. 

Don Carpenter
list@candhdevelopment.com
2006-12-16


From peter at pajamian.dhs.org  Sat Dec 16 18:29:34 2006
From: peter at pajamian.dhs.org (Peter)
Date: Sat Dec 16 18:30:16 2006
Subject: [ic] (no subject)
In-Reply-To: <200612161619601.SM01052@c-e18fdc26944b4>
References: <200612161619601.SM01052@c-e18fdc26944b4>
Message-ID: <4584815E.4040304@pajamian.dhs.org>

On 12/16/2006 01:19 PM, list wrote:
> demo: error in configuration. Skipping.
> demo: config error. Skipping.
> Interchange server started in UNIX mode(s) (process id 4426)

What do the error logs say?

Peter

From list at candhdevelopment.com  Sat Dec 16 19:31:27 2006
From: list at candhdevelopment.com (list)
Date: Sat Dec 16 19:31:51 2006
Subject: [ic] (no subject)
Message-ID: <200612161931711.SM01052@c-e18fdc26944b4>

Hello, Peter,

from;            /usr/local/interchange/error.log

6:13:34:49 -1000] - - STOP server (5152) on signal TERM
- - - [16/December/2006:13:34:53 -1000] - - Low traffic settings.
- - - [16/December/2006:13:34:54 -1000] - - ...UI is loaded...
- - - [16/December/2006:13:34:54 -1000] - - Interchange V5.4.1
- - - [16/December/2006:13:34:54 -1000] - - Running with new signals, external programs could be unreliable.
 Re-run with environment variable PERL_SIGNALS set to "unsafe" to change this.
- - - [16/December/2006:13:34:54 -1000] - - Config 'demo' at server startup
- - - [16/December/2006:13:34:54 -1000] - - demo: config error -- Couldn't change to /home/ddws/catalogs/demo: No such file or directory. Skipping.
- - - [16/December/2006:13:34:55 -1000] - - START server (5448) (UNIX)
- - - [16/December/2006:13:34:55 -1000] - - START server (5448) (UNIX)

When I check...

bash-3.1$ clear
bash-3.1$ cd /home/ddws/catalogs/demo/
bash-3.1$ ls
catalog.cfg  error.log  include  pages     session        tmp
config       etc        logs     products  special_pages  upload
dbconf       images     orders   README    templates      variables
bash-3.1$

It's there.

It seems to be something to do with VendURL?

Thanks!



======= At 2006-12-16, 15:29:00 you wrote: =======

>On 12/16/2006 01:19 PM, list wrote:
>> demo: error in configuration. Skipping.
>> demo: config error. Skipping.
>> Interchange server started in UNIX mode(s) (process id 4426)
>
>What do the error logs say?
>
>Peter
>
>_______________________________________________
>interchange-users mailing list
>interchange-users@icdevgroup.org
>http://www.icdevgroup.org/mailman/listinfo/interchange-users

= = = = = = = = = = = = = = = = = = = =
			

Best regards.				 
list
list@candhdevelopment.com
2006-12-16



From peter at pajamian.dhs.org  Sat Dec 16 19:46:24 2006
From: peter at pajamian.dhs.org (Peter)
Date: Sat Dec 16 19:46:45 2006
Subject: [ic] (no subject)
In-Reply-To: <200612161931711.SM01052@c-e18fdc26944b4>
References: <200612161931711.SM01052@c-e18fdc26944b4>
Message-ID: <45849360.4010700@pajamian.dhs.org>

On 12/16/2006 04:31 PM, list wrote:
> - - - [16/December/2006:13:34:54 -1000] - - demo: config error -- Couldn't change to /home/ddws/catalogs/demo: No such file or directory. Skipping.

> When I check...

> It's there.

Check the permissions on the directory and all directories above it,
make sure that Intechange has read/write/execute permissions on that
directory.

> It seems to be something to do with VendURL?

It doesn't even get to VendURL, in fact it can't even read the
catalog.cfg file because it can't access the directory.

Also please don't top post your replies.  See the list guidelines at
<http://www.icdevgroup.org/mailman/listinfo/interchange-users>

Peter

From list at candhdevelopment.com  Sat Dec 16 22:08:10 2006
From: list at candhdevelopment.com (list)
Date: Sat Dec 16 22:08:25 2006
Subject: [ic] (no subject)
Message-ID: <200612162208452.SM01052@c-e18fdc26944b4>

Hello, Peter,

That seemed to work...

It let me know that I needed

 DBD::mysql

error log

- - - [16/December/2006:16:31:49 -1000] - - demo config error: Required Perl module DBD::mysql not present. Aborting catalog.
 In line 2 of the configuration file 'dbconf/mysql/mysql.cfg':

- - - [16/December/2006:16:31:49 -1000] - - demo: config error. Skipping.
- - - [16/December/2006:16:31:50 -1000] - - START server (16889) (UNIX)
- - - [16/December/2006:16:31:50 -1000] - - START server (16889) (UNIX)

But when I try to intsall DBD::mysql it won't do it without force.

I installed a fresh install of perl wo/threads

..so I could be missing some modules...

Shoul I force DBD::mysql install

If so, how, I am new at this.
			

Best regards.				 
Don
list@candhdevelopment.com
2006-12-16



From peter at pajamian.dhs.org  Sat Dec 16 22:21:00 2006
From: peter at pajamian.dhs.org (Peter)
Date: Sat Dec 16 22:21:20 2006
Subject: [ic] (no subject)
In-Reply-To: <200612162208452.SM01052@c-e18fdc26944b4>
References: <200612162208452.SM01052@c-e18fdc26944b4>
Message-ID: <4584B79C.90809@pajamian.dhs.org>

On 12/16/2006 07:08 PM, list wrote:
> But when I try to intsall DBD::mysql it won't do it without force.

That's because the installer can't run the tests if mysql is password
protected so it aborts.

> Shoul I force DBD::mysql install

Yes.

> If so, how, I am new at this.

In the CPAN shell:

force install DBD::mysql

Peter

From oleg at eville.us  Sun Dec 17 15:09:09 2006
From: oleg at eville.us (Oleg Raskin)
Date: Sun Dec 17 15:09:26 2006
Subject: [ic] mv_nextpage URL
Message-ID: <1699.24.94.155.85.1166386149.squirrel@noc.eville.us>

Greetings.

This may seem like a very simple question, but I've not found an answer in
the mailing list as of yet. I have a form that goes to process.html and
performs some database operations in a procedure defined through mv_check.
That part works fine, however part of this procedure sets the mv_nextpage
depending on form values being submitted. The problem is that mv_nextpage
can only be an interchange page and I need it to go to a page outside of
the interchange catalog. Naturally, I am getting the "missing" page every
time. How can this be done?

Many thanks.
From kevin at cursor.biz  Sun Dec 17 17:40:52 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Sun Dec 17 17:41:11 2006
Subject: [ic] mv_nextpage URL
In-Reply-To: <1699.24.94.155.85.1166386149.squirrel@noc.eville.us>
References: <1699.24.94.155.85.1166386149.squirrel@noc.eville.us>
Message-ID: <200612172240.53283.kevin@cursor.biz>

"Oleg Raskin" <oleg@eville.us> wrote:
> This may seem like a very simple question, but I've not found an answer in
> the mailing list as of yet. I have a form that goes to process.html and
> performs some database operations in a procedure defined through mv_check.
> That part works fine, however part of this procedure sets the mv_nextpage
> depending on form values being submitted. The problem is that mv_nextpage
> can only be an interchange page and I need it to go to a page outside of
> the interchange catalog. Naturally, I am getting the "missing" page every
> time. How can this be done?
> 
Set mv_nextpage to a "bouncer" page and set another variable to the
target URI, such as "http://www.example.com/foo.html".

Your new "bouncer" page only needs to contain the following two lines:

    [bounce href="[cgi your_uri_target]" if="[cgi your_uri_target]"]
    [bounce page="index"]

See the [bounce] tag documentation for more information:

    http://www.interchange.rtfm.info/icdocs/tags/bounce.html

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From list at candhdevelopment.com  Mon Dec 18 13:22:57 2006
From: list at candhdevelopment.com (list)
Date: Mon Dec 18 13:23:24 2006
Subject: [ic] No css on public side
Message-ID: <200612181322851.SM01104@c-e18fdc26944b4>

Hello, interchange-users,

could it be  permissons issue?

>From the error log

[18/December/2006:07:59:57 -1000] standard /cgi-bin/standard/index.html CSS dir images does not exist.


Best regards. 

list
list@candhdevelopment.com
2006-12-18


From DDavenport at newagedigital.com  Mon Dec 18 14:06:13 2006
From: DDavenport at newagedigital.com (Daniel Davenport)
Date: Mon Dec 18 14:07:24 2006
Subject: [ic] No css on public side
Message-ID: <6646EB737DB9164B9A6992E17A47DC74210D89@mssbs01.NewAgeDigital.local>

 

> -----Original Message-----
> From: interchange-users-bounces@icdevgroup.org 
> [mailto:interchange-users-bounces@icdevgroup.org] On Behalf Of list
> Sent: 2006 December 18 -- Monday 1:23 PM
> To: interchange-users@icdevgroup.org
> Subject: [ic] No css on public side
> 
> Hello, interchange-users,
> 
> could it be  permissons issue?
> 
> >From the error log
> 
> [18/December/2006:07:59:57 -1000] standard 
> /cgi-bin/standard/index.html CSS dir images does not exist.

Possibly.  IC likes having write access to the images directory, as it
auto generates some files (among them, the CSS file for the default
catalog).  But also make sure there's a link in the catalog directory,
named "images", that points to the folder where you want IC to generate
the files.

--
Daniel Davenport
New Age Digital
http://www.newagedigital.com

From list at candhdevelopment.com  Mon Dec 18 14:48:39 2006
From: list at candhdevelopment.com (list)
Date: Mon Dec 18 14:49:02 2006
Subject: [ic] No css on public side
Message-ID: <200612181448675.SM01104@c-e18fdc26944b4>

Hello, Daniel Davenport,




>  But also make sure there's a link in the catalog directory,
>named "images", that points to the folder where you want IC to generate
>the files.

The link is there.  The owner is root, group is interch.

I tried to change owner to interch... (with webmin file editer) but it won't let me.
I assume it's because it is a link?

Actually, I have IC as the owner for everything 'even' public files. Could this be part of the problem?

I have tried a few different things :-)

>interchange-users@icdevgroup.org
>http://www.icdevgroup.org/mailman/listinfo/interchange-users

= = = = = = = = = = = = = = = = = = = =
			

Best regards.				 
list
list@candhdevelopment.com
2006-12-18



From DDavenport at newagedigital.com  Tue Dec 19 07:52:07 2006
From: DDavenport at newagedigital.com (Daniel Davenport)
Date: Tue Dec 19 07:52:28 2006
Subject: [ic] No css on public side
Message-ID: <6646EB737DB9164B9A6992E17A47DC74210DA2@mssbs01.NewAgeDigital.local>

 

> -----Original Message-----
> From: interchange-users-bounces@icdevgroup.org 
> [mailto:interchange-users-bounces@icdevgroup.org] On Behalf Of list
> Sent: 2006 December 18 -- Monday 2:49 PM
> To: interchange-users@icdevgroup.org
> Subject: Re: RE: [ic] No css on public side
> 
> Hello, Daniel Davenport,
> 
> >  But also make sure there's a link in the catalog directory,
> >named "images", that points to the folder where you want IC 
> > to generate the files.
> 
> The link is there.  The owner is root, group is interch.
> 
> I tried to change owner to interch... (with webmin file 
> editer) but it won't let me.
> I assume it's because it is a link?

Probably.  Actually, the permissions on the link generally don't matter
-- they're almost always 777.  It's where the link points that the
permissions are important.

> Actually, I have IC as the owner for everything 'even' public 
> files. Could this be part of the problem?

As long as the static files (images, css, etc) are readable by apache,
it generally doesn't matter who the owner is -- any files ic would
create, it'd own. 

Files in the catalog directory, though, should generally be owned by
either ic's user/group or the catalog's. 


From techlist at bnetmd.net  Wed Dec 20 09:31:35 2006
From: techlist at bnetmd.net (Glenn McCalley)
Date: Wed Dec 20 09:32:13 2006
Subject: [ic] admissions taxes based on item - not buyer
Message-ID: <008901c72443$91ea5210$6701a8c0@GLENN2>

Hi there,

We need to charge an admissions tax on sales of tickets based upon the state 
in which the -event- is located, NOT the state of the purchaser.  Needs to 
be based on an item attribute like "tix_tax" ( a percentage) and calculated 
on each item.  Can't apply to the total because an event in New Jersey has a 
different tax rate than an event in New York, and there could be items other 
than tickets in the order.

Looked at "Levies" and that seems to still work based on some constant that 
applies to the whole order, not individual items.

I can show a total Admissions Tax for the entire order or show separate tax 
items for each ticket, but it has to be based on the ticket not the 
purchaser.

Any thoughts on how to do this other than Perl to cycle through "if item = 
xxx" and check for a ticket?

THanks!
Glenn. 


From kevin at cursor.biz  Wed Dec 20 10:43:05 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Wed Dec 20 10:45:00 2006
Subject: [ic] admissions taxes based on item - not buyer
In-Reply-To: <008901c72443$91ea5210$6701a8c0@GLENN2>
References: <008901c72443$91ea5210$6701a8c0@GLENN2>
Message-ID: <200612201543.06226.kevin@cursor.biz>

"Glenn McCalley" <techlist@bnetmd.net> wrote:
> We need to charge an admissions tax on sales of tickets based upon the state 
> in which the -event- is located, NOT the state of the purchaser.  Needs to 
> be based on an item attribute like "tix_tax" ( a percentage) and calculated 
> on each item.  Can't apply to the total because an event in New Jersey has a 
> different tax rate than an event in New York, and there could be items other 
> than tickets in the order.
> 
> Looked at "Levies" and that seems to still work based on some constant that 
> applies to the whole order, not individual items.
> 
> I can show a total Admissions Tax for the entire order or show separate tax 
> items for each ticket, but it has to be based on the ticket not the 
> purchaser.
> 
> Any thoughts on how to do this other than Perl to cycle through "if item = 
> xxx" and check for a ticket?
> 
That's the way I'd probably do it.  I'd most likely create a UserTag
that loops through the cart and applies either an admissions tax or
a sales tax depending upon the current item in the loop.  The total tax
value would be returned for Interchange to use.

You'd presumably charge sales tax, rather than admissions tax, on
any shipping costs involved - or perhaps you hold tickets at the door,
rather than posting them at all.

Your shiny new UserTag could be used by creating a "salestax.asc" file
with one line, as follows:

    default [your-usertag-name]

See here for more tax-related information:

    http://www.interchange.rtfm.info/icdocs/Interchange_sales_tax_calculations.html

The only thing left to do is to work out how you want to look up the
sales and admission tax percentage values for each US location
(performance location for admissions and customer's location for other
sales), and code that into the UserTag.

I wouldn't store the percentages in item attributes, by the way;  I'd
store the performance location and code the UserTag to do the percentage
lookup.  That'd reduce the amount of maintenance required when the tax
rates change.  Whatever you do, make sure you do it securely - you don't
really want people to create their own item attributes and make up some
random sales tax amounts, do you?

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From cplists at princeservices.com  Wed Dec 20 15:25:10 2006
From: cplists at princeservices.com (Cameron B. Prince)
Date: Wed Dec 20 15:25:29 2006
Subject: [ic] Image Upload / Imagehelper Widget
Message-ID: <007501c72474$f2bce360$0401a8c0@PSLAPTOP1A>

Hey guys,

Is there a way to remove an image with the Imagehelper widget?

I received a call from a client who wanted to remove a large image, but
leave the standard and thumbnail images intact. She told me that before the
upgrade from IC v4.8 to v5.2 she could put a space in the field and it would
cause the image to be removed.

Now when you enter a space, the Ok button no longer functions until you
re-edit the item.

Is this a bug or are we overlooking something obvious?

Thanks,
Cameron

From jon at endpoint.com  Wed Dec 20 15:46:43 2006
From: jon at endpoint.com (Jon Jensen)
Date: Wed Dec 20 15:47:21 2006
Subject: [ic] Image Upload / Imagehelper Widget
In-Reply-To: <007501c72474$f2bce360$0401a8c0@PSLAPTOP1A>
References: <007501c72474$f2bce360$0401a8c0@PSLAPTOP1A>
Message-ID: <Pine.LNX.4.64.0612201345040.5268@ybpnyubfg.ybpnyqbznva>

On Wed, 20 Dec 2006, Cameron B. Prince wrote:

> Is there a way to remove an image with the Imagehelper widget?
>
> I received a call from a client who wanted to remove a large image, but 
> leave the standard and thumbnail images intact. She told me that before 
> the upgrade from IC v4.8 to v5.2 she could put a space in the field and 
> it would cause the image to be removed.
>
> Now when you enter a space, the Ok button no longer functions until you 
> re-edit the item.
>
> Is this a bug or are we overlooking something obvious?

I wonder if there is now a post-filter on that field in the metadata, 
stripping out the space. I thought I had seen some imagehelper setups that 
had a blank filename as the first entry, but apparently you're not getting 
that. Sorry I don't have more to suggest.

Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From cplists at princeservices.com  Wed Dec 20 16:33:51 2006
From: cplists at princeservices.com (Cameron B. Prince)
Date: Wed Dec 20 16:33:41 2006
Subject: [ic] Image Upload / Imagehelper Widget
In-Reply-To: <Pine.LNX.4.64.0612201345040.5268@ybpnyubfg.ybpnyqbznva>
Message-ID: <009101c7247e$8b4bfa40$0401a8c0@PSLAPTOP1A>


Hey Jon,

> I wonder if there is now a post-filter on that field in the metadata,
> stripping out the space. I thought I had seen some imagehelper setups that
> had a blank filename as the first entry, but apparently you're not getting
> that. Sorry I don't have more to suggest.

Thanks for replying... I checked the meta and there are no filters defined
for the field. This is in the client's catalog and also on the demo site.
It's really weird. What's happening is almost like some javascript function
is disabling the whole form. Even if you switch to the General tab, the Ok
button does nothing after putting a space in the image field. 

You know, Mike and I worked on what we called the "Ultimate Image Widget"
back when he and I were working on the Red Hat store together. It was
basically a combination of the imagelisting and imagehelper widgets that
allowed you to either select an existing image or use a file input to upload
a new one. I wonder whatever happened to this. I believe I recall seeing it
mentioned in a WHATSNEW file some time ago so I don't think I dreamed the
whole thing.

I may try to resurrect this for the client and it would probably be helpful
to check it into the source.

Thanks,
Cameron

From jon at endpoint.com  Wed Dec 20 16:51:58 2006
From: jon at endpoint.com (Jon Jensen)
Date: Wed Dec 20 16:52:24 2006
Subject: [ic] Image Upload / Imagehelper Widget
In-Reply-To: <009101c7247e$8b4bfa40$0401a8c0@PSLAPTOP1A>
References: <009101c7247e$8b4bfa40$0401a8c0@PSLAPTOP1A>
Message-ID: <Pine.LNX.4.64.0612201450340.5268@ybpnyubfg.ybpnyqbznva>

On Wed, 20 Dec 2006, Cameron B. Prince wrote:

> You know, Mike and I worked on what we called the "Ultimate Image 
> Widget" back when he and I were working on the Red Hat store together. 
> It was basically a combination of the imagelisting and imagehelper 
> widgets that allowed you to either select an existing image or use a 
> file input to upload a new one. I wonder whatever happened to this. I 
> believe I recall seeing it mentioned in a WHATSNEW file some time ago so 
> I don't think I dreamed the whole thing.

Ok, I just checked a site where we have this working. It's using the 
imagehelper widget with the standard nullselect filter, and that's it. The 
first option is "(none)", and I see that code in 
code/Widget/imagehelper.widget. Perhaps upgrading to the latest IC, or at 
least that widget, would work for you. Or even just backporting the 
"(none)" first option code.

Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From cplists at princeservices.com  Wed Dec 20 19:41:57 2006
From: cplists at princeservices.com (Cameron B. Prince)
Date: Wed Dec 20 19:41:33 2006
Subject: [ic] Image Upload / Imagehelper Widget
In-Reply-To: <Pine.LNX.4.64.0612201450340.5268@ybpnyubfg.ybpnyqbznva>
Message-ID: <00a001c72498$d1e78b30$0401a8c0@PSLAPTOP1A>

Hi Jon,

> Ok, I just checked a site where we have this working. It's using the
> imagehelper widget with the standard nullselect filter, and that's it. The
> first option is "(none)", and I see that code in
> code/Widget/imagehelper.widget. Perhaps upgrading to the latest IC, or at
> least that widget, would work for you. Or even just backporting the
> "(none)" first option code.

What do you have in the Directory meta field for the widget? It looks like
that is what is triggering the select box based on this:

if($path =~ s!/\*(?:\.([^/]+))?$!!) {

Thanks,
Cameron

From jon at endpoint.com  Wed Dec 20 22:17:59 2006
From: jon at endpoint.com (Jon Jensen)
Date: Wed Dec 20 22:18:19 2006
Subject: [ic] Image Upload / Imagehelper Widget
In-Reply-To: <00a001c72498$d1e78b30$0401a8c0@PSLAPTOP1A>
References: <00a001c72498$d1e78b30$0401a8c0@PSLAPTOP1A>
Message-ID: <Pine.LNX.4.64.0612202017301.5268@ybpnyubfg.ybpnyqbznva>

On Wed, 20 Dec 2006, Cameron B. Prince wrote:

>> Ok, I just checked a site where we have this working. It's using the 
>> imagehelper widget with the standard nullselect filter, and that's it. 
>> The first option is "(none)", and I see that code in 
>> code/Widget/imagehelper.widget. Perhaps upgrading to the latest IC, or 
>> at least that widget, would work for you. Or even just backporting the 
>> "(none)" first option code.
>
> What do you have in the Directory meta field for the widget? It looks 
> like that is what is triggering the select box based on this:
>
> if($path =~ s!/\*(?:\.([^/]+))?$!!) {

Nothing in a "Directory" field, but the "outboard" field is 
a simple "/path/to/images/*".

Jon

-- 
Jon Jensen
End Point Corporation
http://www.endpoint.com/
From Moskova at infotecharch.com  Thu Dec 21 03:46:40 2006
From: Moskova at infotecharch.com (Wareham Diego)
Date: Thu Dec 21 03:47:54 2006
Subject: [ic] + 5 inches or money back
Message-ID: <35ae01c724dc$0027f6f5$4b556bd0@Gene.midco.net>

YO sir

I don't care why your woody is so small, but 79% of women do.
They are pretty sure that bigger thing will make their desire
stronger. You have the chance to change your life.

Here http://www.mucous.us you can get the thing.

It will help you for sure.
The remedy can be sent worldwide.
If you wont be satisfied - we will return all you money.
No bullshit.

-- 
uhprpnqipsqqpopipspipoppuspjpgqlplplolqtqqqsqupsqqpuqgppofrunsogni
weiuvbmdngsdfwelkgjhhpjw
     'God knows, the blockheads! They  grabbed me,  tied  me  up  with some
filthy rags and dumped me in a lorry!'
     'May  I ask  why you  came  into  the restaurant in  nothing but your
underwear?'
     'There's nothing odd about it,' answered Ivan. '  I went for a swim in
the  Moscow  River  and  someone pinched my  clothes  and left me this  junk
instead! I  couldn't  walk round Moscow naked, could I? I had to put on what
there was, because I was in a hurry to get to the Griboyedov restaurant.'
     The doctor glanced questioningly at Ryukhin, who mumbled sulkily:
     'Yes, that's the name of the restaurant.'


From michael.curtis at citigroup.com  Thu Dec 21 05:13:18 2006
From: michael.curtis at citigroup.com (Curtis, Michael [CCC-OT_IT])
Date: Thu Dec 21 05:14:07 2006
Subject: [ic] Protx Payment Gateway
Message-ID: <948D099B6BEC8843BC4325E8F4A2C94604CE2A94@exukmb20.eur.nsroot.net>

Rather a lot to enquire about but as all seem to relate to Protx Payment Gateway issues then please bear with me

Original configuration
FreeBSD 6.0
Perl 5.8.8 Unthreaded
IC 5.4.1
Mysql 5.x

Configuration now
Fedora Core 6
Perl 5.8.8 Threaded
IC 5.4.1
Mysql 5.x

Standard Store was created and locale changed to UK with shipping configured to use Royal mail so test orders (From P.O's) could be created....they were all o.k.

Installed Protx 1.0.9f according to the doc's and now I have a few problems

1 using transactions.txt & transactions.mysql from the protx upgrade & then removing .transactions.sql
The table fails to build and the catalog fails to start, reinstating the original transactions.txt & transactions.mysql then the table will build and the catalog starts

2 Adding the columns with mysql admin using transactions.mysql to identify the additional columns allows the catalog to start but the 'key' column 'code' is now empty i.e. I did have entries such as 'TEST001' now there is no data in the table.column 'transactions.code' on a fresh order, transactions.code is not updated (There must be somewhere in UI to set the numbering system but I cannot find it!)

3 On trying to use a Credit Card transaction I get 
x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:42 +0000] standard4 /cgi-bin/standard4/process Non-existent order routing main, skipping.
x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:43 +0000] standard4 /cgi-bin/standard4/process Unknown charge type: protx
x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:43 +0000] standard4 /cgi-bin/standard4/process Safe: Real-time charge failed. Reason: 
> 
> 
> 		die errmsg(
> 				"Real-time charge failed. Reason: %s\n",
> 				errmsg($Session->{cybercash_error}),
> 			);
> 	
> 
x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:43 +0000] standard4 /cgi-bin/standard4/process Route log failed.
x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:43 +0000] standard4 /cgi-bin/standard4/process ERRORS on ORDER :
> Error during creation of order routing log:
		> Route log failed. at /usr/local/interchange/lib/Vend/Order.pm line 1835.

Your help in solving these issues will be most welcome

Regards

Michael


From lyn at zolotek.net  Thu Dec 21 11:15:52 2006
From: lyn at zolotek.net (Lyn St George)
Date: Thu Dec 21 11:14:51 2006
Subject: [ic] Protx Payment Gateway
In-Reply-To: <948D099B6BEC8843BC4325E8F4A2C94604CE2A94@exukmb20.eur.nsroot.net>
Message-ID: <20061221160758.03ADF830086@mailserver1.zolotek.net>

On Thu, 21 Dec 2006 10:13:18 -0000, Curtis, Michael [CCC-OT_IT] wrote:

>Rather a lot to enquire about but as all seem to relate to Protx Payment Gateway issues 
then please bear with me
>
>Original configuration
>FreeBSD 6.0
>Perl 5.8.8 Unthreaded
>IC 5.4.1
>Mysql 5.x
>
>Configuration now
>Fedora Core 6
>Perl 5.8.8 Threaded
>IC 5.4.1
>Mysql 5.x

Your threaded perl may be an issue elsewhere, but probably not in this
particular case.

>Standard Store was created and locale changed to UK with shipping configured to use 
Royal mail so test orders (From P.O's) could be created....they were all o.k.
>
>Installed Protx 1.0.9f according to the doc's and now I have a few problems
>
>1 using transactions.txt & transactions.mysql from the protx upgrade & then removing 
.transactions.sql
>The table fails to build and the catalog fails to start, reinstating the original transactions.txt 
& transactions.mysql then the table will build and the catalog starts

Are you sure that the files did not have DOS line endings when uploaded?
That can cause a variety of similar effects.


>2 Adding the columns with mysql admin using transactions.mysql to identify the additional 
columns allows the catalog to start but the 'key' column 'code' is now empty i.e. I did have 
entries such as 'TEST001' now there is no data in the table.column 'transactions.code' on a 
fresh order, transactions.code is not updated (There must be somewhere in UI to set the 
numbering system but I cannot find it!)

When you install a new, empty, transactions.txt and then remove the
.transactions.sql marker and restart Interchange, it will import the data
from the new transactions.txt into your MySQL table and overwrite
whatever is there.  The fact that your mysql admin panel could add
new columns to the table further suggests that the uploaded files
were faulty (in the Unix sense)

The counter is in etc/order.number. 


>3 On trying to use a Credit Card transaction I get 
>x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:42 +0000] standard4 
/cgi-bin/standard4/process Non-existent order routing main, skipping.
>x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:43 +0000] standard4 
/cgi-bin/standard4/process Unknown charge type: protx
>x.x.0.50 rqAKSV6h:x.x.0.50 - [16/December/2006:19:35:43 +0000] standard4 
/cgi-bin/standard4/process Safe: Real-time charge failed. Reason: 

Did the Protx module load when Interchange started? Probably not
or it would not be 'unknown'.

[snip]

>Your help in solving these issues will be most welcome
>
>Regards
>
>Michael
>

-
Lyn

From DB at M-and-D.com  Thu Dec 21 11:20:12 2006
From: DB at M-and-D.com (DB)
Date: Thu Dec 21 11:21:12 2006
Subject: [ic] change displayed URL of results page
Message-ID: <458AB43C.2070208@M-and-D.com>

Is there anyway to change the URL of a results page from something like:

http://domain.com/scan/se=AA/sf=category/st=db/sp=results/tf=sku/ml=20.html

to something like:

http://domain.com/somethingelse.html ?

Is this possible with any IC and/or Apache tricks?

DB



From MichaelC at glcweb.co.uk  Thu Dec 21 16:26:16 2006
From: MichaelC at glcweb.co.uk (Michael Curtis)
Date: Thu Dec 21 16:26:43 2006
Subject: [ic] Protx Payment Gateway
Message-ID: <A33C77E06C9E924F8E6D796CA3D635D10238B7@w2k3sbs.glcdomain.local>

> 
> On Thu, 21 Dec 2006 10:13:18 -0000, Curtis, Michael [CCC-OT_IT] wrote:
> 
> >Configuration now
> >Fedora Core 6
> >Perl 5.8.8 Threaded
> >IC 5.4.1
> >Mysql 5.x
> 
> >The table fails to build and the catalog fails to start, reinstating the
> original transactions.txt
> & transactions.mysql then the table will build and the catalog starts
> 
> Are you sure that the files did not have DOS line endings when uploaded?
> That can cause a variety of similar effects.
It must be along those lines, tried Vi, Nano to edit but all looks fine, tried cp the file from the untarred Protx directory i.e not opened in an editor, still no joy, tried retyping the transactions.txt and got to nine entries before the table rebuild failed
> 
> The counter is in etc/order.number.
That location confirms numbering starting at TEST00001 but still no order ref showing in 'transactions.code'
> 
> Did the Protx module load when Interchange started? Probably not
> or it would not be 'unknown'.
You were spot on with that, I put the 'include' statement in catalog.cfg Doh!! Further error messages with that now but I suspect Firewall issues and will have to read up on Protx requirements

Many thanks Lyn........
Your further help in solving these issues will be most welcome
> >
> >Regards
> >
> >Michael
> >
> 
> -
> Lyn
> 
> _______________________________________________
> interchange-users mailing list
> interchange-users@icdevgroup.org
> http://www.icdevgroup.org/mailman/listinfo/interchange-users
> 
> --
> This message has been scanned for viruses and
> dangerous content by IC-MailScanner, and is
> believed to be clean.
> 
> For queries or information please contact:-
> 
> =================================
> Internet Central Technical Support
> 
> 
>  http://www.netcentral.co.uk
> =================================
> 
> 
> --
> No virus found in this incoming message.
> Checked by AVG Free Edition.
> Version: 7.5.432 / Virus Database: 268.15.26/594 - Release Date:
> 20/12/2006 15:54
> 

-- 
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.15.26/594 - Release Date: 20/12/2006 15:54
 

From DDavenport at newagedigital.com  Thu Dec 21 17:23:05 2006
From: DDavenport at newagedigital.com (Daniel Davenport)
Date: Thu Dec 21 17:23:31 2006
Subject: [ic] change displayed URL of results page
Message-ID: <6646EB737DB9164B9A6992E17A47DC74210E4B@mssbs01.NewAgeDigital.local>

 

> -----Original Message-----
> From: interchange-users-bounces@icdevgroup.org 
> [mailto:interchange-users-bounces@icdevgroup.org] On Behalf Of DB
> Sent: 2006 December 21 -- Thursday 11:20 AM
> To: interchange-users@icdevgroup.org
> Subject: [ic] change displayed URL of results page
> 
> Is there anyway to change the URL of a results page from 
> something like:
> 
> http://domain.com/scan/se=AA/sf=category/st=db/sp=results/tf=s
> ku/ml=20.html
> 
> to something like:
> 
> http://domain.com/somethingelse.html ?
> 
> Is this possible with any IC and/or Apache tricks?
> 
> DB

Of course it's possible -- anything is, if you want it badly enough.  =)
Two ways i can think of off the top of my head are ActionMaps and a
little mod_rewrite magic.

Rewrites are pretty simple, but can be effective.  You could, for
example, just match (.*)/your_url(\.html)?$, and rewrite it to
$1/scan/se=category/se=AA/ml=20.  The only real problem is that
more-links on the results page still end up with long, ugly urls.  Only
way around that is to do your own more-links, afaik, and that could
complicate your rewrite rules a bit.

ActionMaps are a bit more involved.  The /scan/... urls are actually
done via an action map (internal to IC?), so of course it's possible to
do your own action that works like /scan.  But since you'd be pretty
much taking over the job that /scan does, you'll likely need to set a
lot of the cgi and search variables yourself (unless someone else knows
of an easier way...).

Another option, if you only care about the search being available via
your url, might be to make a page that just bounces to the /scan... url.
But that's ugly -- among other things, it replaces your nice short url
in the address bar with the /scan... one.
 
--
Daniel Davenport
New Age Digital
http://www.newagedigital.com

From kevin at cursor.biz  Thu Dec 21 20:15:05 2006
From: kevin at cursor.biz (Kevin Walsh)
Date: Thu Dec 21 20:16:36 2006
Subject: [ic] change displayed URL of results page
In-Reply-To: <6646EB737DB9164B9A6992E17A47DC74210E4B@mssbs01.NewAgeDigital.local>
References: <6646EB737DB9164B9A6992E17A47DC74210E4B@mssbs01.NewAgeDigital.local>
Message-ID: <200612220115.07451.kevin@cursor.biz>

"Daniel Davenport" <DDavenport@newagedigital.com> wrote:
> > Is there anyway to change the URL of a results page from 
> > something like:
> > 
> > http://domain.com/scan/se=AA/sf=category/st=db/sp=results/tf=s
> > ku/ml=20.html
> > 
> > to something like:
> > 
> > http://domain.com/somethingelse.html ?
> > 
> > Is this possible with any IC and/or Apache tricks?
> > 
> [snip]
>
> ActionMaps are a bit more involved.  The /scan/... urls are actually
> done via an action map (internal to IC?), so of course it's possible to
> do your own action that works like /scan.  But since you'd be pretty
> much taking over the job that /scan does, you'll likely need to set a
> lot of the cgi and search variables yourself (unless someone else knows
> of an easier way...).
> 
It's not too involved.  You could use something like the following:

    ActionMap cat <<EOA
    sub {
        my $cat = shift;
        $cat =~ s:^cat/+::i;

        $CGI->{mv_searchtype} = 'db';
        $CGI->{mv_search_field} = 'category';
        $CGI->{mv_searchspec} = $cat;
        $CGI->{mv_sort_field} = 'sku';
        $CGI->{mv_matchlimit} = 20;

        $CGI->{mv_todo} = 'search';
        $Tag->update('process');

        return 1;
    }
    EOA

Using the above local ActionMap example, you could search for all products
in the "AA" category using a URI like "http://www.example.com/cat/AA.html".

Substitute "AA" for "foobar", or whatever you'd like to look for in the
"products.category" column.

See also:

    http://www.interchange.rtfm.info/icdocs/config/ActionMap.html
    http://www.interchange.rtfm.info/icdocs/Search_parameters.html

-- 
   _/   _/  _/_/_/_/  _/    _/  _/_/_/  _/    _/
  _/_/_/   _/_/      _/    _/    _/    _/_/  _/   K e v i n   W a l s h
 _/ _/    _/          _/ _/     _/    _/  _/_/    kevin@cursor.biz
_/   _/  _/_/_/_/      _/    _/_/_/  _/    _/
From emailgrant at gmail.com  Thu Dec 21 20:56:23 2006
From: emailgrant at gmail.com (Grant)
Date: Thu Dec 21 20:56:48 2006
Subject: [ic] Payment.pm bug fixed
Message-ID: <49bf44f10612211756y146c0975y8890dea64eeadcd6@mail.gmail.com>

I've never been able to submit extra information along with my
Authorize.net transactions, except for the name and address.  The
city, state, zip code, and country would never show up on the other
end.  I'm finally able to submit all of the above information after
commenting out all of the sections in Payment.pm which contain
"billing_set".  Here is a link to Payment.pm:

http://www.icdevgroup.org/cgi-bin/cvsweb/~checkout~/interchange/lib/Vend/Payment.pm?rev=2.17;content-type=text%2Fplain;hideattic=0

Can anyone tell me what the billing_set stuff does?  I'm using IC 5.2.

- Grant
From michael.curtis at citigroup.com  Fri Dec 22 02:42:07 2006
From: michael.curtis at citigroup.com (Curtis, Michael [CCC-OT_IT])
Date: Fri Dec 22 02:42:29 2006
Subject: [ic] Protx Payment Gateway
Message-ID: <948D099B6BEC8843BC4325E8F4A2C94604CE2A99@exukmb20.eur.nsroot.net>


> 
> >The table fails to build and the catalog fails to start, reinstating the
> original transactions.txt
> & transactions.mysql then the table will build and the catalog starts
> 
The problem is with the column 'repeats'.  With this column in then the table will not build, not even if I edit the 'transactions.txt' file with vi/nano!
Leave it out and the table builds fine
 
> The counter is in etc/order.number.
That location confirms numbering starting at TEST00001 but still no order ref showing in 'transactions.code'
> 

Many thanks Lyn........
Your further help in solving these issues will be most welcome
> >
> >Regards
> >
> >Michael
> >
> 
> -
> Lyn
> 
> _______________________________________________
> interchange-users mailing list
> interchange-users@icdevgroup.org
> http://www.icdevgroup.org/mailman/listinfo/interchange-users
> 
> --
> This message has been scanned for viruses and
> dangerous content by IC-MailScanner, and is
> believed to be clean.
> 
> For queries or information please contact:-
> 
> =================================
> Internet Central Technical Support
> 
> 
>  http://www.netcentral.co.uk
> =================================
> 
> 
> --
> No virus found in this incoming message.
> Checked by AVG Free Edition.
> Version: 7.5.432 / Virus Database: 268.15.26/594 - Release Date:
> 20/12/2006 15:54
> 

-- 
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.15.26/594 - Release Date: 20/12/2006 15:54
 

_______________________________________________
interchange-users mailing list
interchange-users@icdevgroup.org
http://www.icdevgroup.org/mailman/listinfo/interchange-users

From lyn at zolotek.net  Fri Dec 22 05:21:10 2006
From: lyn at zolotek.net (Lyn St George)
Date: Fri Dec 22 05:19:21 2006
Subject: [ic] Protx Payment Gateway
In-Reply-To: <948D099B6BEC8843BC4325E8F4A2C94604CE2A99@exukmb20.eur.nsroot.net>
Message-ID: <20061222101313.3C950830087@mailserver1.zolotek.net>

On Fri, 22 Dec 2006 07:42:07 -0000, Curtis, Michael [CCC-OT_IT] wrote:

>
>> 
>> >The table fails to build and the catalog fails to start, reinstating the
>> original transactions.txt
>> & transactions.mysql then the table will build and the catalog starts
>> 
>The problem is with the column 'repeats'.  With this column in then the table will not build, 
not even if I edit the 'transactions.txt' file with vi/nano!
>Leave it out and the table builds fine

I've just looked through the docs for that, and it is mentioned as 'repeat',
ie not plural. However this is one of those that were put in as an attempt
at "future proofing", and as I mentioned before they can all be omitted.

The only ones you need are those listed for log_transaction.


I'm in the process of doing a new version, both for their new API and
also to handle their problem with duplicate orders (what happens is
that they send back an 'empty' response, which IC reads as failure
and so the customer resends the order. However Protx have taken
the money at their end, even though they don't say so, and have
no system for checking duplicate payments. The result is irate
customers ...). It also checks for their availability (they have their
share of outages) and optionally logs the transaction offline for
you to complete manually later - generally thought to be better
than telling the customer to come back later.  If you want to try
this one then I'll send it offlist.



>Many thanks Lyn........
>Your further help in solving these issues will be most welcome
>> >
>> >Regards
>> >
>> >Michael
>> >
>> 
>> -
>> Lyn


-
Lyn

From michael.curtis at citigroup.com  Fri Dec 22 05:54:59 2006
From: michael.curtis at citigroup.com (Curtis, Michael [CCC-OT_IT])
Date: Fri Dec 22 05:55:31 2006
Subject: [ic] Protx Payment Gateway
Message-ID: <948D099B6BEC8843BC4325E8F4A2C94604CE2A9A@exukmb20.eur.nsroot.net>


I'm in the process of doing a new version, both for their new API and
also to handle their problem with duplicate orders (what happens is
that they send back an 'empty' response, which IC reads as failure
and so the customer resends the order. However Protx have taken
the money at their end, even though they don't say so, and have
no system for checking duplicate payments. The result is irate
customers ...). It also checks for their availability (they have their
share of outages) and optionally logs the transaction offline for
you to complete manually later - generally thought to be better
than telling the customer to come back later.  If you want to try
this one then I'll send it offlist.

----------

Yes Please..I have just looked at 5.5 demo and I am impressed, not shure of what features of 5.5 I will end up with in 5.4.1 and so will try a two pronged approach (dev and er! superdev?) 
Can you please sent it to michael.curtis@glcweb.co.uk as this I can access from anywhere
As for 'repeat' in transactions.txt that was from the doc's, 'repeats' in the email was from memory so me wrong there

Rgds

Michael 



>Many thanks Lyn........
>Your further help in solving these issues will be most welcome
>> >
>> >Regards
>> >
>> >Michael
>> >
>> 
>> -
>> Lyn


-
Lyn

_______________________________________________
interchange-users mailing list
interchange-users@icdevgroup.org
http://www.icdevgroup.org/mailman/listinfo/interchange-users

From peter at pajamian.dhs.org  Fri Dec 22 06:48:52 2006
From: peter at pajamian.dhs.org (Peter)
Date: Fri Dec 22 06:49:30 2006
Subject: [ic] Protx Payment Gateway
In-Reply-To: <948D099B6BEC8843BC4325E8F4A2C94604CE2A99@exukmb20.eur.nsroot.net>
References: <948D099B6BEC8843BC4325E8F4A2C94604CE2A99@exukmb20.eur.nsroot.net>
Message-ID: <458BC624.8010603@pajamian.dhs.org>

On 12/21/2006 11:42 PM, Curtis, Michael [CCC-OT_IT] wrote:
> The problem is with the column 'repeats'.  With this column in then
> the table will not build, not even if I edit the 'transactions.txt'
> file with vi/nano! Leave it out and the table builds fine

It probably has something to do with the fact that REPEAT is a reserved
word in mysql.  <http://dev.mysql.com/doc/refman/5.1/en/reserved-words.html>

Peter

From emailgrant at gmail.com  Fri Dec 22 08:55:21 2006
From: emailgrant at gmail.com (Grant)
Date: Fri Dec 22 08:56:08 2006
Subject: [ic] Re: Payment.pm bug fixed
In-Reply-To: <49bf44f10612211756y146c0975y8890dea64eeadcd6@mail.gmail.com>
References: <49bf44f10612211756y146c0975y8890dea64eeadcd6@mail.gmail.com>
Message-ID: <49bf44f10612220555t41f69248i88ba2daad8e879f@mail.gmail.com>

> I've never been able to submit extra information along with my
> Authorize.net transactions, except for the name and address.  The
> city, state, zip code, and country would never show up on the other
> end.  I'm finally able to submit all of the above information after
> commenting out all of the sections in Payment.pm which contain
> "billing_set".  Here is a link to Payment.pm:
>
> http://www.icdevgroup.org/cgi-bin/cvsweb/~checkout~/interchange/lib/Vend/Payment.pm?rev=2.17;content-type=text%2Fplain;hideattic=0
>
> Can anyone tell me what the billing_set stuff does?  I'm using IC 5.2.
>
> - Grant

I wanted to add that this is important because it qualifies me for the
discounted rate with my processor which is about half.

- Grant
From zeolite61 at yahoo.com  Fri Dec 22 12:21:56 2006
From: zeolite61 at yahoo.com (Ken Douglas)
Date: Fri Dec 22 12:22:22 2006
Subject: [ic] SEO Module Community Project
Message-ID: <20061222172156.24705.qmail@web59101.mail.re1.yahoo.com>

Hi,

I have searched thru mailing-list archive (example
thread
http://www.icdevgroup.org/pipermail/interchange-users/2005-April/042831.html
) and found various partial and non-standard ways to
achieve some kind of SEO-optimized URLs. 

In the past Perusion, Mike Heins' company, has
Community Projects located in
http://www.perusion.com/c/public/communityprojects/projectlistings.html.
Logic is simple setup a public project with a cost
then interested parties pledge money. So Mike can
setup this and we can go on.

A similar example project is OpenSEF located in
http://projects.j-prosolution.com/en/projects/os-projects/project-opensef.html
. This a Joomla/Mambo CMS component and it lets you do
creative ways of URL rewriting. You can have:
http://www.domain.com/section/category/item.html
http://www.domain.com/section/some-prefix-item.html
http://www.domain.com/category/item-some-suffix.html
http://www.domain.com/section/category/item.asp
In short you can change everything after your
top-level domain address.

But, of course, this comes with a performance penalty.
This can be neutralized by using cache. One more point
is that you may only want to use it only on product
and category display pages not on account or order
pages. Otherwise it can be problematic.

These days you should utilize every chance to show up
high in Google and keywords in URLs help a lot. 

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
From zeolite61 at yahoo.com  Fri Dec 22 12:22:02 2006
From: zeolite61 at yahoo.com (Ken Douglas)
Date: Fri Dec 22 12:22:26 2006
Subject: [ic] SEO Module Community Project
Message-ID: <782136.28834.qm@web59106.mail.re1.yahoo.com>

Hi,

I have searched thru mailing-list archive (example
thread
http://www.icdevgroup.org/pipermail/interchange-users/2005-April/042831.html
) and found various partial and non-standard ways to
achieve some kind of SEO-optimized URLs. 

In the past Perusion, Mike Heins' company, has
Community Projects located in
http://www.perusion.com/c/public/communityprojects/projectlistings.html.
Logic is simple setup a public project with a cost
then interested parties pledge money. So Mike can
setup this and we can go on.

A similar example project is OpenSEF located in
http://projects.j-prosolution.com/en/projects/os-projects/project-opensef.html
. This a Joomla/Mambo CMS component and it lets you do
creative ways of URL rewriting. You can have:
http://www.domain.com/section/category/item.html
http://www.domain.com/section/some-prefix-item.html
http://www.domain.com/category/item-some-suffix.html
http://www.domain.com/section/category/item.asp
In short you can change everything after your
top-level domain address.

But, of course, this comes with a performance penalty.
This can be neutralized by using cache. One more point
is that you may only want to use it only on product
and category display pages not on account or order
pages. Otherwise it can be problematic.

These days you should utilize every chance to show up
high in Google and keywords in URLs help a lot. 

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
From cplists at princeservices.com  Sat Dec 23 03:03:48 2006
From: cplists at princeservices.com (Cameron B. Prince)
Date: Sat Dec 23 03:03:40 2006
Subject: [ic] Image Upload / Imagehelper Widget
In-Reply-To: <Pine.LNX.4.64.0612202017301.5268@ybpnyubfg.ybpnyqbznva>
Message-ID: <006b01c72668$e1ea4730$0401a8c0@PSLAPTOP1A>

Hi Jon,

> Nothing in a "Directory" field, but the "outboard" field is
> a simple "/path/to/images/*".

That was it... 

Thanks,
Cameron

From lyn at zolotek.net  Sat Dec 23 04:09:34 2006
From: lyn at zolotek.net (Lyn St George)
Date: Sat Dec 23 04:07:40 2006
Subject: [ic] Protx Payment Gateway
In-Reply-To: <948D099B6BEC8843BC4325E8F4A2C94604CE2A9A@exukmb20.eur.nsroot.net>
Message-ID: <20061223090134.21914830087@mailserver1.zolotek.net>

On Fri, 22 Dec 2006 10:54:59 -0000, Curtis, Michael [CCC-OT_IT] wrote:

[snip about new version]
>than telling the customer to come back later.  If you want to try
>this one then I'll send it offlist.
>
>----------
>
>Yes Please..I have just looked at 5.5 demo and I am impressed, not shure of what 
features of 5.5 I will end up with in 5.4.1 and so will try a two pronged approach (dev and er! 
superdev?) 
>Can you please sent it to michael.curtis@glcweb.co.uk as this I can access from 
anywhere

I'll send it after the 27th - Protx have disabled my test account
and as I haven't looked at this for a while I need to run at least
one test before sending it. (Protx are back on the 27th)

>Rgds
>
>Michael 


-
Lyn

From jonathan.millerobrq at viatel.com  Mon Dec 25 00:02:04 2006
From: jonathan.millerobrq at viatel.com (Brenton Small)
Date: Sun Dec 24 16:02:58 2006
Subject: [ic] demonstration  
Message-ID: <884027917264.035696461791@viatel.com>

Fantastis DVD safe quality exclusive generous videos

http://girlslovearnie.info

From bob at nleaudio.com  Tue Dec 26 11:17:58 2006
From: bob at nleaudio.com (Bob Puff)
Date: Tue Dec 26 11:09:35 2006
Subject: [ic] Dead 5.2 in RPC?
Message-ID: <45914B36.6090309@nleaudio.com>

Hi Gang,

I recently had some significant trouble with a 5.2 cart and 
authorize.net duplicating charges - it would tell the customer the card 
was declined, when the charge really did go through.  In searching the 
archives, the only solution I found was to use RPC mode (I have been 
running several carts on this same box in High setting for years, no 
problems).  Two days after I switched to RPC mode, interchange hangs 
up.  I had to kill -9 all the ic processes, and restart.  I saw no 
errors in any of the error.logs that warned of any problems.  It didn't 
appear that my box ran out of memory, as top reports about the same 
memory usage as before the switch.

Where can I look?  Obviously, a lockup like this is not a good thing!

Alternatively, if there were a fix for the authorizenet problem, I may 
want to implement that, and go back to High traffic settings.

Bob

From ron at endpoint.com  Tue Dec 26 11:20:39 2006
From: ron at endpoint.com (Ron Phipps)
Date: Tue Dec 26 11:21:32 2006
Subject: [ic] Dead 5.2 in RPC?
In-Reply-To: <45914B36.6090309@nleaudio.com>
References: <45914B36.6090309@nleaudio.com>
Message-ID: <45914BD7.2060601@endpoint.com>

Bob Puff wrote:
> Hi Gang,
> 
> I recently had some significant trouble with a 5.2 cart and 
> authorize.net duplicating charges - it would tell the customer the card 
> was declined, when the charge really did go through.  In searching the 
> archives, the only solution I found was to use RPC mode (I have been 
> running several carts on this same box in High setting for years, no 
> problems).  Two days after I switched to RPC mode, interchange hangs 
> up.  I had to kill -9 all the ic processes, and restart.  I saw no 
> errors in any of the error.logs that warned of any problems.  It didn't 
> appear that my box ran out of memory, as top reports about the same 
> memory usage as before the switch.
> 
> Where can I look?  Obviously, a lockup like this is not a good thing!
> 
> Alternatively, if there were a fix for the authorizenet problem, I may 
> want to implement that, and go back to High traffic settings.
> 
> Bob

Hello Bob,

Try editing your interchange.cfg file and in the high traffic area change the max servers setting to: MaxServers 0 and then switch back to high traffic.

-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From bob at nleaudio.com  Tue Dec 26 12:21:59 2006
From: bob at nleaudio.com (Bob Puff)
Date: Tue Dec 26 12:22:35 2006
Subject: [ic] Re: Dead 5.2 in RPC?
In-Reply-To: <200612261700.kBQH072p022729@icdevgroup.org>
References: <200612261700.kBQH072p022729@icdevgroup.org>
Message-ID: <20061226172033.M96687@nleaudio.com>


> Bob Puff wrote:
> > Hi Gang,
> > 
> > I recently had some significant trouble with a 5.2 cart and 
> > authorize.net duplicating charges - it would tell the customer the card 
> > was declined, when the charge really did go through.  In searching the 
> > archives, the only solution I found was to use RPC mode (I have been 
> > running several carts on this same box in High setting for years, no 
> > problems).  Two days after I switched to RPC mode, interchange hangs 
> > up.  I had to kill -9 all the ic processes, and restart.  I saw no 
> > errors in any of the error.logs that warned of any problems.  It didn't 
> > appear that my box ran out of memory, as top reports about the same 
> > memory usage as before the switch.
> > 
> > Where can I look?  Obviously, a lockup like this is not a good thing!
> > 
> > Alternatively, if there were a fix for the authorizenet problem, I may 
> > want to implement that, and go back to High traffic settings.
> > 
> > Bob
> 
> Hello Bob,
> 
> Try editing your interchange.cfg file and in the high traffic area 
> change the max servers setting to: MaxServers 0 and then switch back 
> to high traffic.
> 
> -- 
> Ron Phipps
> End Point Corporation
> ron@endpoint.com


Hi Ron,

Will that fix the authorizenet problems?  I'm a little gun-shy, after having
to do like 20 refunds due to this bug.  Thanks!

Bob

P.S. I am running Perl 5.8.4, non-threaded

From ryantxln at resourceskc.com  Tue Dec 26 20:32:42 2006
From: ryantxln at resourceskc.com (Tia Edwards)
Date: Tue Dec 26 12:37:09 2006
Subject: [ic] double-dyed  .
Message-ID: <616941937266.214014470611@resourceskc.com>

[Unknown Tag *$$dating* Please Fix]
http://supportotherwomen.info

From ron at endpoint.com  Tue Dec 26 12:41:21 2006
From: ron at endpoint.com (Ron Phipps)
Date: Tue Dec 26 12:42:13 2006
Subject: [ic] Re: Dead 5.2 in RPC?
In-Reply-To: <20061226172033.M96687@nleaudio.com>
References: <200612261700.kBQH072p022729@icdevgroup.org>
	<20061226172033.M96687@nleaudio.com>
Message-ID: <45915EC1.2000705@endpoint.com>

Bob Puff wrote:
>> Bob Puff wrote:
>>> Hi Gang,
>>>
>>> I recently had some significant trouble with a 5.2 cart and 
>>> authorize.net duplicating charges - it would tell the customer the card 
>>> was declined, when the charge really did go through.  In searching the 
>>> archives, the only solution I found was to use RPC mode (I have been 
>>> running several carts on this same box in High setting for years, no 
>>> problems).  Two days after I switched to RPC mode, interchange hangs 
>>> up.  I had to kill -9 all the ic processes, and restart.  I saw no 
>>> errors in any of the error.logs that warned of any problems.  It didn't 
>>> appear that my box ran out of memory, as top reports about the same 
>>> memory usage as before the switch.
>>>
>>> Where can I look?  Obviously, a lockup like this is not a good thing!
>>>
>>> Alternatively, if there were a fix for the authorizenet problem, I may 
>>> want to implement that, and go back to High traffic settings.
>>>
>>> Bob
>> Hello Bob,
>>
>> Try editing your interchange.cfg file and in the high traffic area 
>> change the max servers setting to: MaxServers 0 and then switch back 
>> to high traffic.
>>
>> -- 
>> Ron Phipps
>> End Point Corporation
>> ron@endpoint.com
> 
> 
> Hi Ron,
> 
> Will that fix the authorizenet problems?  I'm a little gun-shy, after having
> to do like 20 refunds due to this bug.  Thanks!
> 
> Bob
> 
> P.S. I am running Perl 5.8.4, non-threaded
> 

Hello Bob,

I can't guarantee it will fix the problem, but your symptoms all sound like 
things I have ran into in the past (if you search the archives you'll find a lot of info about "double triple charges with signio" and the fix was to run in RPC mode with MaxServers 0.  Recently I was instructed to setup all sites at End Point with MaxServers 0 in high traffic mode to get the stability of high traffic mode with the fix for perl signaling problems that cause issues when calling sendmail or external programs.

There was also a thread about authorizenet issues awhile ago and I think the solution may have been to use another method to contact authorizenet (using wget instead of lwp).

-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From bob at nleaudio.com  Wed Dec 27 02:03:20 2006
From: bob at nleaudio.com (Bob Puff)
Date: Wed Dec 27 01:54:25 2006
Subject: [ic] Re: Dead 5.2 in RPC?
Message-ID: <45921AB8.60308@nleaudio.com>

>> Hi Ron,
>> 
>> Will that fix the authorizenet problems?  I'm a little gun-shy, after having
>> to do like 20 refunds due to this bug.  Thanks!
>> 
>> Bob
>> 
>> P.S. I am running Perl 5.8.4, non-threaded
>> 
> 
> Hello Bob,
> 
> I can't guarantee it will fix the problem, but your symptoms all sound like 
> things I have ran into in the past (if you search the archives you'll find a lot of info 
> about "double triple charges with signio" and the fix was to run in RPC mode 
> with MaxServers 0.  Recently I was instructed to setup all sites at End Point with 
> MaxServers 0 in high traffic mode to get the stability of high traffic mode with the fix 
> for perl signaling problems that cause issues when calling sendmail or external 
> programs.
> 
> There was also a thread about authorizenet issues awhile ago and I think the solution may 
> have been to use another method to contact authorizenet (using wget instead of lwp).
> 
> -- 
> Ron Phipps
> End Point Corporation
> ron@endpoint.com

Hi Ron,

I thought the thing that fixed the perl signal problems was the RPC mode, not necessarily 
the MaxServers thing.  Has the high traffic setting been working ok for you?

I did a search in the archives on authorizenet and wget, and came up with nothing.  Any 
other suggestions for info on this?

Bob


From ron at endpoint.com  Wed Dec 27 13:11:09 2006
From: ron at endpoint.com (Ron Phipps)
Date: Wed Dec 27 13:12:03 2006
Subject: [ic] Re: Dead 5.2 in RPC?
In-Reply-To: <45921AB8.60308@nleaudio.com>
References: <45921AB8.60308@nleaudio.com>
Message-ID: <4592B73D.60109@endpoint.com>

Bob Puff wrote:
>>> Hi Ron,
>>>
>>> Will that fix the authorizenet problems?  I'm a little gun-shy, after 
>>> having
>>> to do like 20 refunds due to this bug.  Thanks!
>>>
>>> Bob
>>>
>>> P.S. I am running Perl 5.8.4, non-threaded
>>>
>>
>> Hello Bob,
>>
>> I can't guarantee it will fix the problem, but your symptoms all sound 
>> like things I have ran into in the past (if you search the archives 
>> you'll find a lot of info about "double triple charges with signio" 
>> and the fix was to run in RPC mode with MaxServers 0.  Recently I was 
>> instructed to setup all sites at End Point with MaxServers 0 in high 
>> traffic mode to get the stability of high traffic mode with the fix 
>> for perl signaling problems that cause issues when calling sendmail or 
>> external programs.
>>
>> There was also a thread about authorizenet issues awhile ago and I 
>> think the solution may have been to use another method to contact 
>> authorizenet (using wget instead of lwp).
>>
>> -- 
>> Ron Phipps
>> End Point Corporation
>> ron@endpoint.com
> 
> Hi Ron,
> 
> I thought the thing that fixed the perl signal problems was the RPC 
> mode, not necessarily the MaxServers thing.  Has the high traffic 
> setting been working ok for you?
> 
> I did a search in the archives on authorizenet and wget, and came up 
> with nothing.  Any other suggestions for info on this?
> 
> Bob

Hi Bob,

It's definitely MaxServers 0 that gets around the signaling issue when calling external programs.  High traffic with MaxServers 0 is working wonderfully for us.  It's used at FrozenCPU.com, Backcountry.com, Reliabemedical.com, any of the sites maintained by End Point are using that setup.saz

Here's a link to the commit about using wget:

http://www.icdevgroup.org/pipermail/interchange-cvs/2005-May/006341.html

I'm not sure where the main thread is about this problem, but it's in the archives.  Here is another post about the matter:

http://www.icdevgroup.org/pipermail/interchange-users/2005-June/043376.html

You may try searching for wget and Payment.pm or wget and Payment and see what turns up.

-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From tpruitt at pruittcom.com  Wed Dec 27 16:55:17 2006
From: tpruitt at pruittcom.com (Tommy Pruitt)
Date: Wed Dec 27 16:57:10 2006
Subject: [ic] error when order on admin Enter Order page
Message-ID: <E035AC17E02B774D8BFD71187ED937A00BECFA@pci12.pruittcom.local>

I get the following error when entering an order via the admin pages
using the "Enter Order" tab.

There were errors in your last entry. Pointers are shown in this color
below.

    * po_number: Please include your PO number

We have a PO entered in the PO box.

Thanks for any help.
Tommy


From icdev at mrlock.com  Wed Dec 27 17:14:49 2006
From: icdev at mrlock.com (Steve Graham)
Date: Wed Dec 27 17:15:07 2006
Subject: [ic] error when order on admin Enter Order page
In-Reply-To: <E035AC17E02B774D8BFD71187ED937A00BECFA@pci12.pruittcom.loc
 al>
References: <E035AC17E02B774D8BFD71187ED937A00BECFA@pci12.pruittcom.local>
Message-ID: <7.0.1.0.0.20061227161323.01c42fd0@mrlock.com>

At 03:55 PM 12/27/2006, you wrote:
>I get the following error when entering an order via the admin pages
>using the "Enter Order" tab.
>
>There were errors in your last entry. Pointers are shown in this color
>below.
>
>     * po_number: Please include your PO number
>
>We have a PO entered in the PO box.
>
>Thanks for any help.
>Tommy

I'm not sure I can offer any assistance, but you need to let everyone 
know what version of IC you are using.
Can you test and create this error in the demo version?
http://www.icdevgroup.org/i/dev/demo

Steve 

From tpruitt at pruittcom.com  Thu Dec 28 13:18:54 2006
From: tpruitt at pruittcom.com (Tommy Pruitt)
Date: Thu Dec 28 13:20:53 2006
Subject: [ic] error when order on admin Enter Order page
In-Reply-To: <7.0.1.0.0.20061227161323.01c42fd0@mrlock.com>
Message-ID: <E035AC17E02B774D8BFD71187ED937A00BED06@pci12.pruittcom.local>

>>At 03:55 PM 12/27/2006, you wrote:
>>I get the following error when entering an order via the admin pages
>>using the "Enter Order" tab.
>>
>>There were errors in your last entry. Pointers are shown in this color
>>below.
>>
>>     * po_number: Please include your PO number
>>
>>We have a PO entered in the PO box.
>>
>>Thanks for any help.
>>Tommy
>
>I'm not sure I can offer any assistance, but you need to let everyone 
>know what version of IC you are using.
>Can you test and create this error in the demo version?
>http://www.icdevgroup.org/i/dev/demo
>
>Steve 

I am running 5.4.1
On the demo store I get a different error message.  Do you think
upgrading to 5.5 will resolve the problem?

There were errors in your last entry. Pointers are shown in this color
below. 

prof_po_accepted: This site doesn't accept purchase orders. You should
not have been offered that option; please contact us.



_______________________________________________
interchange-users mailing list
interchange-users@icdevgroup.org
http://www.icdevgroup.org/mailman/listinfo/interchange-users

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


From icdev at mrlock.com  Thu Dec 28 14:00:38 2006
From: icdev at mrlock.com (Steve Graham)
Date: Thu Dec 28 14:00:59 2006
Subject: [ic] error when order on admin Enter Order page
In-Reply-To: <E035AC17E02B774D8BFD71187ED937A00BED06@pci12.pruittcom.loc
 al>
References: <7.0.1.0.0.20061227161323.01c42fd0@mrlock.com>
	<E035AC17E02B774D8BFD71187ED937A00BED06@pci12.pruittcom.local>
Message-ID: <7.0.1.0.0.20061228125648.01cbd870@mrlock.com>


>I am running 5.4.1
>On the demo store I get a different error message.  Do you think
>upgrading to 5.5 will resolve the problem?
>
>There were errors in your last entry. Pointers are shown in this color
>below.
>
>prof_po_accepted: This site doesn't accept purchase orders. You should
>not have been offered that option; please contact us.

I also run 5.4, but I don't use the admin order entry - we get our 
people to enter orders from the front end just like customers.
If I really needed it, I would try to debug the admin entry on the 
back end to fit my needs. I can't say about upgrading to 5.5, because 
I haven't looked to see if there are any database changes - proceed 
carefully on that option.

You might be able to copy the needed (newer admin) file from 5.5 and 
see if it works on your 5.4 backend - I would try this first (easiest 
:) - make sure you backup your old file.

Others on this list might have a little more insight, but I think 
most are out on vacation.

Good luck

-Steve 

From josh at myprivacy.ca  Thu Dec 28 14:53:47 2006
From: josh at myprivacy.ca (Josh Lavin)
Date: Thu Dec 28 14:54:18 2006
Subject: [ic] error when order on admin Enter Order page
In-Reply-To: <E035AC17E02B774D8BFD71187ED937A00BECFA@pci12.pruittcom.local>
References: <E035AC17E02B774D8BFD71187ED937A00BECFA@pci12.pruittcom.local>
Message-ID: <0525A7EC-67C7-41F6-B35A-81FA84566DF2@myprivacy.ca>

On Dec 27, 2006, at 3:55 PM, Tommy Pruitt wrote:

> I get the following error when entering an order via the admin pages
> using the "Enter Order" tab.
>
> There were errors in your last entry. Pointers are shown in this color
> below.
>
>     * po_number: Please include your PO number
>
> We have a PO entered in the PO box.

Upgrading might help, if there's some unknown problem, but that could  
be a lot of work.

With that error, the first thing I would check is to make sure the PO  
number field is named the same as the label it has. The error makes  
it seem that the field should be called "po_number".

That's my meager contribution...

--
Josh Lavin
Kingdom Design   http://www.kingdomdesign.com/

From josh at myprivacy.ca  Thu Dec 28 17:21:13 2006
From: josh at myprivacy.ca (Josh Lavin)
Date: Thu Dec 28 17:21:45 2006
Subject: [ic] Email as username with 5.4.0
Message-ID: <0643C9D8-C12C-493A-9363-9FD027250D2A@myprivacy.ca>

Hello list,

Just wanted to make sure I do this right. I am making a new site  
require everyone to login, using their email address as the username.

It seems I need to take these steps to get this to work with IC 5.4.0:

1. Set "UserDB  indirect_login  email" in catalog.cfg
2. Make sure the email column is a UNIQUE index
3. Upgrade UserDB.pm to latest CVS
4. Set "UserDB  default  username_email  1" and
    "UserDB  default  ignore_case  1" in catalog.cfg

Does that look right? I want to be sure before upgrading UserDB.pm,  
because I have other catalogs on this installation which use the  
standard username method.

Side note: according to:
http://www.icdevgroup.org/pipermail/interchange-cvs/2006-July/ 
006875.html
it looks like users can't change their username (email). Is that  
still the case?

Thanks!
--
Josh Lavin
Kingdom Design   http://www.kingdomdesign.com/

From bob at nleaudio.com  Fri Dec 29 13:11:26 2006
From: bob at nleaudio.com (Bob Puff@NLE)
Date: Fri Dec 29 13:37:51 2006
Subject: [ic] Re: Dead 5.2 in RPC?
In-Reply-To: <200612281700.kBSH0K2p020756@icdevgroup.org>
References: <200612281700.kBSH0K2p020756@icdevgroup.org>
Message-ID: <45955A4E.1030105@nleaudio.com>


> Hi Bob,
> 
> It's definitely MaxServers 0 that gets around the signaling issue when calling external programs.  High traffic with MaxServers 0 is working wonderfully for us.  It's used at FrozenCPU.com, Backcountry.com, Reliabemedical.com, any of the sites maintained by End Point are using that setup.saz
> 
> Here's a link to the commit about using wget:
> 
> http://www.icdevgroup.org/pipermail/interchange-cvs/2005-May/006341.html
> 
> I'm not sure where the main thread is about this problem, but it's in the archives.  Here is another post about the matter:
> 
> http://www.icdevgroup.org/pipermail/interchange-users/2005-June/043376.html
> 
> You may try searching for wget and Payment.pm or wget and Payment and see what turns up.
> 

Hi Ron,

Thanks for the leads!  I've implemented the Payment.pm and AuthorizeNet.pm from a 5.4.1 on my 5.2
box, and set it back to high traffic with Maxservers set to 0.  So far, so good.

This was a big, ugly bug that put a lot of egg on my face.  Too bad this wasn't addressed a long
time ago.

Bob

From ron at endpoint.com  Fri Dec 29 13:50:44 2006
From: ron at endpoint.com (Ron Phipps)
Date: Fri Dec 29 14:19:06 2006
Subject: [ic] Re: Dead 5.2 in RPC?
In-Reply-To: <45955A4E.1030105@nleaudio.com>
References: <200612281700.kBSH0K2p020756@icdevgroup.org>
	<45955A4E.1030105@nleaudio.com>
Message-ID: <45956384.5000500@endpoint.com>

Bob Puff@NLE wrote:
>> Hi Bob,
>>
>> It's definitely MaxServers 0 that gets around the signaling issue when calling external programs.  High traffic with MaxServers 0 is working wonderfully for us.  It's used at FrozenCPU.com, Backcountry.com, Reliabemedical.com, any of the sites maintained by End Point are using that setup.saz
>>
>> Here's a link to the commit about using wget:
>>
>> http://www.icdevgroup.org/pipermail/interchange-cvs/2005-May/006341.html
>>
>> I'm not sure where the main thread is about this problem, but it's in the archives.  Here is another post about the matter:
>>
>> http://www.icdevgroup.org/pipermail/interchange-users/2005-June/043376.html
>>
>> You may try searching for wget and Payment.pm or wget and Payment and see what turns up.
>>
> 
> Hi Ron,
> 
> Thanks for the leads!  I've implemented the Payment.pm and AuthorizeNet.pm from a 5.4.1 on my 5.2
> box, and set it back to high traffic with Maxservers set to 0.  So far, so good.
> 
> This was a big, ugly bug that put a lot of egg on my face.  Too bad this wasn't addressed a long
> time ago.
> 
> Bob

No problem.

I believe this was addressed as soon as reports turned up that it was an issue.  I know the RPC/MaxServers 0 solution was addressed back in 2002.

RPC/MaxServers 0 is stable in most installations, however it appears you have one of those installs where it acts up.

The Payment.pm/wget fix was done in May 2005 shortly after people started to see the issue.

Trust me I know how you feel, the mess I had to clean up after the Verisign double/triple charges issue was a pain, but in the end it helped solidify IC even more :)

-- 
Ron Phipps
End Point Corporation
ron@endpoint.com
From emailgrant at gmail.com  Sun Dec 31 12:05:55 2006
From: emailgrant at gmail.com (Grant)
Date: Sun Dec 31 12:06:12 2006
Subject: [ic] Re: Payment.pm bug fixed
In-Reply-To: <49bf44f10612211756y146c0975y8890dea64eeadcd6@mail.gmail.com>
References: <49bf44f10612211756y146c0975y8890dea64eeadcd6@mail.gmail.com>
Message-ID: <49bf44f10612310905t731d9633s21196e15e25f82a@mail.gmail.com>

> I've never been able to submit extra information along with my
> Authorize.net transactions, except for the name and address.  The
> city, state, zip code, and country would never show up on the other
> end.  I'm finally able to submit all of the above information after
> commenting out all of the sections in Payment.pm which contain
> "billing_set".  Here is a link to Payment.pm:
>
> http://www.icdevgroup.org/cgi-bin/cvsweb/~checkout~/interchange/lib/Vend/Payment.pm?rev=2.17;content-type=text%2Fplain;hideattic=0
>
> Can anyone tell me what the billing_set stuff does?  I'm using IC 5.2.

Is anyone able to submit city, state, zip code, and country to
Authorize.net with an unmodified Payment.pm?

- Grant