[ic] Call for testers

[Gert van der Spoel]

> -----Original Message-----
> From: interchange-users-bounces at icdevgroup.org [mailto:interchange-
> users-bounces at icdevgroup.org] On Behalf Of David Christensen
> Sent: Tuesday, March 10, 2009 8:46 PM
> To: interchange-users at icdevgroup.org
> Subject: [ic] Call for testers
> 
> Greetings interchange enthusiasts!
> 
> As we all know, UTF-8 support in interchange has been a bit of a mixed
> bag.  We aim to rectify this issue by identifying and fixing these
> issues, but we need people who are willing to test the latest and
> greatest versions of the fixes.  I have set up a specific github
> repository to track the changes
> (http://github.com/machack666/interchange-utf8/tree/utf8
> ).  ***Please note that this code is considered *experimental* and
> should *not* be run on a production system at this point.***
> Nevertheless, if you have some spare hardware or spare cycles, or have
> a copy of your site that you can easily set up in a testing
> environment, we need your help.
> 
> Please note that all UTF-8-related development is occurring in the
> utf8 branch; `master' is just tracking interchange's existing CVS
> repository.
> 
> Acquire the code by issuing the following commands:
> 
>    $ cd where/you/want/your/repository
>    $ git clone git://github.com/machack666/interchange-utf8.git
>    $ cd interchange-utf8
>    $ git checkout utf8
> 
> The build/install process should be identical to Interchange's CVS
> build instructions, so if you are familiar with that, you're good to
> go.
> 
> Issues which I believe are resolved:
>   - the ever elusive utf8_safe_regex_workaround; aka Safe.pm require
> errors.  (And there was much rejoicing!)  I have tested the Safe fixes
> against perl 5.10, but we need to verify that the code functions in
> perl 5.8.8 and 5.8.9 as well.
> 
> Specific bugs/issues identified but still outstanding:
>   - Vend::Charset errors in [calc] blocks in INIT_PAGE.
>   - reports of PreFork fork bombs when MV_UTF8 is enabled.
>   - issues with Fedora's perl 5.10 version.
> 
> Additional consideration/candidates for increased UTF-8 support:
>   - GDBM support
>   - Outgoing email
> 
> Testing criteria:
>   - test all changes against perl 5.8.8, 5.8.9, 5.10, both threaded
> and non-threaded.
>   - test against catalogs with have MV_UTF8 enabled.
>   - just as important: we need to test against catalogs which use
> legacy encodings so we can verify that new changes do not break
> working code on your part.
> 
> Comments/questions/volunteers? :-)

Hi David,

Took me a while to get my development stuff in order, but I've started to
give it a try with the instructions above (and in the follow up messages
with regard to the git checkout procedures).

Not sure if I did all correctly, but I hope so ;)  ...

Anyway Debian etch, Perl 5.10.0 threaded
I am one using GDBM ... 

After the whole installation procedure and starting up I have the following
error:
10.0.0.152 ZNdyjm84:10.0.0.152 - [17/June/2009:13:25:51 +0200] acm /acm
Runtime error: Undefined subroutine
&Vend::CharSet::utf8_safe_regex_workaround called at
/opt/interchange-utf8/lib/Vend/Interpolate.pm line 1207.

Which seems to be correct because the sub looks like it does not exist ...
Or did I take the wrong branch to work on?

As known the GDBM stuff does not work ... I also think that the utf8 filter
story on the GDBM data is not needed or needs to be checked. I have
locale.txt  which is a text file which is having UTF8 encoded text ... if I
create the GDBM file from it there are 2 times utf8 filter applies .. But if
I understand correctly it will utf8 my utf8 and then utf8 that again on open
... So that does not seem like the right way to go ... 

But first I think the utf8_safe_regex_workaround question has to be looked
at .. is it still needed or not?

CU,

Gert