[Camps-users] [Camps-commits] [SCM] DevCamps development environments branch, master, updated. 4c506f3ce3c269ec5ac7776865df2599454ccaca
Brian J. Miller
brian at endpoint.com
Tue Jan 19 17:47:11 UTC 2010
Jon Jensen wrote:
> On Tue, 19 Jan 2010, I pulled Brian's commits:
>
>> commit 4c506f3ce3c269ec5ac7776865df2599454ccaca
>> Author: Brian J. Miller <brian at endpoint.com>
>> Date: Tue Jan 19 12:05:10 2010 -0500
>>
>> Add handling of git submodules during mkcamp
>>
>> * Includes refactoring vcs_local_revert to change directory to the full path of the file(s) that it is checking out, as opposed to the top of the camp
>>
>> commit 21b89205adfd537346e6377e9692df0a75fff7b9
>> Author: Brian J. Miller <brian at endpoint.com>
>> Date: Tue Jan 19 10:24:30 2010 -0500
>>
>> Add configuration options for enabling generation of SSL certificate for Postgres during initialization
>
> Thanks for those, Brian. I haven't tested them but since you said they've
> been in production for a while, we'll trust they're fairly safe. :)
>
> I have a couple of minor questions/comments:
>
>> + push @cmds, [
>> + 'cd %s && git submodule init; git submodule update',
>> + $conf->{path},
>> + ];
>
> I presume, and from simple tests, believe, that "git submodule" is a no-op
> when you have no .gitmodules file. Is that right? So it's safe to run the
> above command blindly?
>
Correct.
> Also, the precedence of the shell operators you used doesn't seem right if
> I understand your intentions correctly. It'll run "git submodule update"
> regardless. Try this:
>
> % cd HOTDOGS && echo first; echo second
> cd: no such file or directory: HOTDOGS
> second
>
> I'm guessing you wanted:
>
> cd %s && (git submodule init; git submodule update)
>
Yeah, that seems more correct. My shell foo is weak so I pretty much
copy and pasted from elsewhere as in the wild it was working.
>> + do_system("openssl genrsa -out $key_path");
>> + do_system("chmod 600 $key_path");
>> + do_system("openssl req -new -x509 -days 3650 -key $key_path -out $crt_path -config $tmpfile");
>> +
>> + unlink($tmpfile) or die "Error unlinking $tmpfile: $!\n";
>
> Any reason to spawn a subprocess for that chmod, instead of using Perl's
> native chmod as we normally do (and as you do for native unlink)?
>
Not really. The only other 'chmod' currently in there was done using
do_system so I mimicked it.
> Thanks,
> Jon
>
You want to make the changes, or want me to?
--
Brian J. Miller
End Point Corp.
brian at endpoint.com
More information about the Camps-users
mailing list