Re: Hypermail: new translation + bug reports

From: <jose.kahan_at_w3.org_at_hypermail-project.org>
Date: Fri, 5 Nov 1999 14:29:05 +0100 (MET)
Message-Id: <199911051329.OAA11605_at_tuvalu.inrialpes.fr>

Hello Bjanri,

I've just finishing commiting your "is" language translation.

In our previous episode, Bjarni R. Einarsson said:
>
> Patch number two adds a "stripsubject" feature to hypermail, which
> makes it remove all occurances of a given word from the subject line,
> which is really useful for un-cluttering the subjects of some mailing
> lists (my lists all have their subjects prefixed with [listname]: -
> which is quite redundant in an archive). This was a trivial feature to
> implement, the patch for it is also attached to this message.

Good idea! I don't know if I'll have the time to do it myself today.. and as it seems I can only find one week per month to work on hypermail, maybe someone else could do it in the meantime?

> Patch number three (and ex-bug-report) implemented Jose Kahan's
> HTML-detection-disabling, which I read about just now and thouroghly
> approve of. :-) Without it some inlined HTML was causing the lower
> half of some of my messages to go missing because of the </HTML> (with
> some browsers). As you may guess from my signature, I'm glad he
> decided not to disable the signature detecting code. :-)
>

Yes, I disabled the code (with a comment for the moment), rather than adding an option for disabling it. If no one was objects, we'll erase it eventually.  

> Bug reports:
>
> I noticed a minor bug in the hypermail "usage" message, where the -M
> flag was printed out as one of the supported languages. This patch is
> corrected by my "stripsubject" patch.

I'm not sure I understand, but it seems to be OK now. Looks like your someone else made the same patch.

> The second bug I found had to do with MIME parsing, and I didn't have
> the energy to wade in and try and fix it.
>
> If a header contains something like this:
>
> Content-Type: Multipart/Mixed;
> boundary="Boundary-=_nWlrBbmQBhCDarzOwKkYHIDdqSCD"
>
> and if the body of the message is just plain text that doesn't
> actually have any MIME parts (the boundary string never occurs),
> then hypermail will search the rest of the file for the boundary,
> and since it probably won't find it (the boundary string is usually
> pretty random) it will conclude that there are no more messages. If
> on the other hand it found the same boundary string further down,
> something even stranger would probably occur. :-)

What kind of mail client is generating that broken message? Is it at least adding an end-boundary?
If this is the case, the patch is quite easy to do.

Could you mail a sample broken message?

> Since I'm not in a rush to put my archives online, and I'm regenerating
> archives containing thousands of messages many times a day already,
> would I be doing you guys a favour if I ran daily snapshots instead of
> 2b25?

For testing purposes, it could be useful to run against the cvs base. This will help detect bugs (or their absence :) faster.

-Jose Received on Fri 05 Nov 1999 03:29:12 PM GMT