Re: New version of the coordination pages (pseudo-urls)

[Guillem Jover]

Hi,

On Fri, 2008-06-20 at 22:49:24 +0200, Nicolas François wrote:
> I've prepared a new script for the generation of the coordination pages
> (http://i18n.debian.net/debian-l10n/)

> The new pages are currently generated on:
> http://i18n.debian.net/~nekral/www-debian-l10n/

We also use our own bot at  (based on some
old version of Tim Dijkstra's bot), source code at
.

> I rebuilt the database by parsing the mailing list since their creation.
> 
> You will notice that there are some old messages. Those ones were not
> closed with some BTS or DONE tags.

> Could you review the new pages and check if using the 2006-01-01 threshold
> would be OK for all of you?
> (I can use a different threshold for each mailing lists.
> You may also prefer to use some DONE messages to close them.)

For the d-i manual pages I usually send a DONE message to the bot
after committing the changes to svn, althought I think Jordà at some
point was working on being able to automatically close the d-i manual
entries.

The bot is subscribed to the list and hooked through procmail, so we
can also easily remove and regenerate the sqlite db as we store the
mails in an mbox, this way we also avoid having to poll and parse the
mailing list web archives, which has always felt like a regression
from the initial bots.

regards,
guillem


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New version of the coordination pages (pseudo-urls)

[Nicolas François]

Hi,

On Tue, Jun 24, 2008 at 09:58:23AM +0300, Guillem Jover wrote:
> 
> We also use our own bot at  (based on some
> old version of Tim Dijkstra's bot), source code at
> .

Thanks for the pointer.

The main differences I could find:
 * procmail filter based
 * cgi script (the i18n.debian.net pages are statically generated every
   hour)
 * sqlite database (the i18n.debian.net database uses a Perl database
   similar to the i18n material)
 * The po files are extracted from the RFR/LCFC messages and available on
   the server

> The bot is subscribed to the list and hooked through procmail, so we
> can also easily remove and regenerate the sqlite db as we store the
> mails in an mbox, this way we also avoid having to poll and parse the
> mailing list web archives, which has always felt like a regression
> from the initial bots.

I will check with my i18n.debian.net admins to see how we can provide
procmail bots. That should not be an issue.

One advantage of parsing the web archives is that we can provide a link to
the thread. On the other hand, I can change my habits and use a date.

Kind Regards,
-- 
Nekral


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New version of the coordination pages (pseudo-urls)

[Guillem Jover]

On Sun, 2008-07-06 at 21:27:49 +0200, Nicolas François wrote:
> On Tue, Jun 24, 2008 at 09:58:23AM +0300, Guillem Jover wrote:
> > The bot is subscribed to the list and hooked through procmail, so we
> > can also easily remove and regenerate the sqlite db as we store the
> > mails in an mbox, this way we also avoid having to poll and parse the
> > mailing list web archives, which has always felt like a regression
> > from the initial bots.
> 
> I will check with my i18n.debian.net admins to see how we can provide
> procmail bots. That should not be an issue.
> 
> One advantage of parsing the web archives is that we can provide a link to
> the thread. On the other hand, I can change my habits and use a date.

Oh, but the bot could infer the right URL to the archive using:

  

regards,
guillem


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]

Re: New version of the coordination pages (pseudo-urls)

[Nicolas François]

On Mon, Jul 07, 2008 at 12:04:20AM +0300, Guillem Jover wrote:
> > 
> > One advantage of parsing the web archives is that we can provide a link to
> > the thread. On the other hand, I can change my habits and use a date.
> 
> Oh, but the bot could infer the right URL to the archive using:
> 
>   

Nice!
Thanks a lot.

Cheers,
-- 
Nekral


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]