Strange INN-current crash.

[Julien ÉLIE]

Hi Petr,

> This is the session:
> takethis <a2 at 33>
> Path: pety
> Date: Wed, 01 Apr 2009 02:12:01 -0500
> Subject: 010518
> From: <pety at pety>
> Distribution:
> Message-ID: <a2 at 33>
> Newsgroups: relcom.test
>
>
> Test
> .
> 439 <a2 at 33>
> check <a2 at 33>
> 238 <a2 at 33>

OK, I understand this one.

I obtain with a new patch I have just written:

Apr  2 12:21:30.921 - ? (null) 439 No colon-space in "Distribution:" header
Apr  2 12:43:47.455 - localhost <a2 at 33> 439 No colon-space in "Distribution:" header

depending on the Message-ID: header is received before or after the
Distribution: header.

The fact is that innd does a preliminary check on the syntax of headers fields.
439 is known to be returned even before the article has been totally received:  just
after having received "Distribution:", innd knows it will reject it.

Then, on later CHECK, it returns 238 (and not 438) because it hopes a well-formed
article will be sent for this article (it was maybe an error the first time).



> takethis <a2 at 33>
> Path: pety
> Date: Wed, 01 Apr 2009 02:12:01 -0500
> Subject: 010518
> From: <pety at pety>
> Message-ID: <a2 at 33>
> Distribution:
> Newsgroups: relcom.test
>
>
> Test
> .
> 439 <a2 at 33>
> check <a2 at 33>
> 438 <a2 at 33>

Should it be accepted instead?
With my new patch, I have 239 for TAKETHIS and:

Apr  2 12:47:06.327 + localhost <a2 at 33> 240 inpaths! source-archive!


Is it what you would like?  Accepting such articles with "Distribution:", "Distribution: "
or "Distribution:   "?

-- 
Julien ÉLIE

« Oublie les injures, n'oublie jamais les bienfaits. »