[Debian-l10n-devel] Failure of Spider

[Nicolas François]

On Wed, May 28, 2008 at 12:49:50PM -0300, Felipe Augusto van de Wiel (faw) wrote:
> 
> 	The last log with 'uninitialized' is this:
> 
> Spider.pm Continue spanish from message 2008/5/90
> Use of uninitialized value in split at lib/Debian/L10n/Spider.pm line 132.

This one is caused by (maybe) malformed mails, or mails with too long
subject. In that case, the subject looks like:
[TAF]	ddp://manuals/debian-faq/fr/{pkgtools,redist,software,support,uptoda
	te}.sgml Simon Paillard

And the spider cannot find the closing bracket.

The problem is that I'm not sure
[TAF]	foo
	bar

is equivalent to:

[TAF]	foobar

or:

[TAF]	foo bar


I found that the cron job erased the logs, so I made some changes in
/etc/cron.d/debian-i18n-robots to fix this.

I also prepared /org/dl10n-stuff/bin/spiderhourly2, which will work with
the svn version of dl10n.
It currently generates the files in my public_html because I would like to
test it more (and I will have to initialize the databases).

This script separates the errors from regular log. It generates a log and
err logfile, and output only the err logfile.

This should dramatically reduce the amount of log you receive, and should
be useful to point issues.

> 	I will create a local SVN just for the meta-data from the
> robots, we should work together to improve the log handling, right
> now it generates quite a few messages every day.

You can insert this to spidercvs (just look at the commented lines).

Cheers,
-- 
Nekral