PGnus 0.9{5,7} hangs after generating summary buffer

PGnus 0.9{5,7} hangs after generating summary buffer

[Thomas Löfgren]

I'm forwarding a mail which made two PGnusae (unpatched 0.95 and dito
0.97) hang after generating the group summary.  When I selected the
group with this article, the message buffer said
 "Generating summary...done"
The summary buffer was empty, and Emacs stopped responding.

I finally tried with this Gnus version, Gnus v5.7, and in the summary
buffer, the author appears to be '0'.

I can't see anything strange with this article (other than the fact
that someone sent an unsubscribe message to the bugtraq list, and the
moderator let it go through!), so I suggest someone more clueful than
myself when it comes to the standards takes a look at it.

Tom
-- 
T. Lofgren <lofgren@sics.se> - Wherever I lay my .emacs, that's my ${HOME}
I thought I was wrong once before, but I realized I was mistaken.

------- Start of forwarded message -------
Message-ID: <"/GUID:Q7+u6bcJm0xGIFQAIx5E5DA*/G=Tom/S=Bosscher/O=system-e/PRMD=vickers-systems/ADMD=attmail/C=us/"@MHS>
Date: 	Thu, 9 Sep 1999 10:40:57 -0400
Reply-To: Tom.Bosscher@VICKERS-SYSTEMS.COM
From: Tom Bosscher <Tom.Bosscher@VICKERS-SYSTEMS.COM>
Subject:      =?iso-8859-1?Q?RE:_[Linux]_glibc_2.1.x_/_wu-?=
              =?iso-8859-1?Q?ftpd_<=3d2.5_/_BeroFTPD_/_lynx_/_vlo?=
              =?iso-8859-1?Q?ck=0d=0a______________?=
              =?iso-8859-1?Q?/_mc_/_glibc_2.0.x?=
To: BUGTRAQ@SECURITYFOCUS.COM

unsubscribe

-----Original Message-----
From: BUGTRAQ(a)SECURITYFOCUS.COM [mailto:BUGTRAQ(a)SECURITYFOCUS.COM]
Sent: Sunday, September 05, 1999 2:48 PM
To: BUGTRAQ(a)SECURITYFOCUS.COM
Subject: Re: [Linux] glibc 2.1.x / wu-ftpd <=2.5 / BeroFTPD / lynx /
vlock / mc / glibc 2.0.x


On Mon, 30 Aug 1999, Norbert Warmuth wrote:
> Michal Zalewski writes:
> > Also, mc seems to have serious problems with directories containing
shell
> > commands enclosed in $(...) construction. Bad.
> What are you talking about? Please send details to
mc-bugs@nuclecu.unam.mx.

I haven't got any response from Michal and hasn't been able to
reproduce any problems with directories containing "$(...)" either.

Wojtek Pilorz reminded of the bash 1.14 vulnerability when PS1 contains
\w or \W. As MC doesn't touch PS1 blaming MC for this is more than
far-fetched.

Pavel Machek got the message 'Warning: Couldn't change to /tmp/$( ...
)'
because he triggered a MC bug which was fixed in March 1999 (release
4.5.27).

Last not least there was an issue with uncompressing files which was
fixed in 4.5.38.

The Midnight Commander bug's mentioned above are fixed in the latest
release which doesn't contain known security vulnerabilities. You can
get it from
     ftp://ftp.gnome.org/pub/GNOME/sources/mc

Please report bugs to mc-bugs@nuclecu.unam.mx, thanks.

Kind regards,
Norbert

------- End of forwarded message -------

Re: PGnus 0.9{5,7} hangs after generating summary buffer

[Lars Magne Ingebrigtsen]

lofgren@sics.se (Thomas Löfgren) writes:

> I'm forwarding a mail which made two PGnusae (unpatched 0.95 and dito
> 0.97) hang after generating the group summary.  When I selected the
> group with this article, the message buffer said
>  "Generating summary...done"
> The summary buffer was empty, and Emacs stopped responding.

Try to `(setq debug-on-quit t)' and then `C-g' when it hangs.  Mail me
the resulting backtrace.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: PGnus 0.9{5,7} hangs after generating summary buffer

[Thomas Lofgren]

And the same is true in PGnus 0.98 and 0.99.  When invoking
gnus-thread-hide-subtree in one of my groups, it goes into an infinite
loop.  I posted the mail that causes this on Sep 29, and I can repost
it if you wish.

I'm not sure how it happens, but the way it's displayed when I don't
hide the threads make me believe that its subject line confuses the
threading quite a bit.  It looks like this:

 Subject:      =?iso-8859-1?Q?RE:_[Linux]_glibc_2.1.x_/_wu-?=
               =?iso-8859-1?Q?ftpd_<=3d2.5_/_BeroFTPD_/_lynx_/_vlo?=
               =?iso-8859-1?Q?ck=0d=0a______________?=
               =?iso-8859-1?Q?/_mc_/_glibc_2.0.x?=

Now, the article is in a thread by itself, so one would think that
hiding the threads wouldn't touch it.  However, Gnus displays the
subject correctly (or at least I don't know it to be incorrect)
encoded in the summary buffer, but there is a newline in it, which
(from the font-lock) appears to make Gnus believe that the second line
is a reply to the first line.

I hope that this can be addressed before the release version is out,
since few people like their programs to hang. :)

Thanks,
Tom
-- 
T. Lofgren <lofgren@sics.se> - Wherever I lay my .emacs, that's my ${HOME}
"Do you have CD players in Sweden?" -The always well-informed David Letterman

Re: PGnus 0.9{5,7} hangs after generating summary buffer

[Lars Magne Ingebrigtsen]

Thomas Lofgren <lofgren@sics.se> writes:

> I'm not sure how it happens, but the way it's displayed when I don't
> hide the threads make me believe that its subject line confuses the
> threading quite a bit.  It looks like this:
> 
>  Subject:      =?iso-8859-1?Q?RE:_[Linux]_glibc_2.1.x_/_wu-?=
>                =?iso-8859-1?Q?ftpd_<=3d2.5_/_BeroFTPD_/_lynx_/_vlo?=
>                =?iso-8859-1?Q?ck=0d=0a______________?=
>                =?iso-8859-1?Q?/_mc_/_glibc_2.0.x?=

This looks like a normal subject line to me.  :-)

> Now, the article is in a thread by itself, so one would think that
> hiding the threads wouldn't touch it.  However, Gnus displays the
> subject correctly (or at least I don't know it to be incorrect)
> encoded in the summary buffer, but there is a newline in it, which
> (from the font-lock) appears to make Gnus believe that the second line
> is a reply to the first line.

Do you use font-lock in the Gnus buffers?

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: PGnus 0.9{5,7} hangs after generating summary buffer

[Thomas Lofgren]

>>>>> "Lars" == Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

    Lars> Thomas Lofgren <lofgren@sics.se> writes:

    >> Now, the article is in a thread by itself, so one would think
    >> that hiding the threads wouldn't touch it.  However, Gnus
    >> displays the subject correctly (or at least I don't know it to
    >> be incorrect) encoded in the summary buffer, but there is a
    >> newline in it, which (from the font-lock) appears to make Gnus
    >> believe that the second line is a reply to the first line.

    Lars> Do you use font-lock in the Gnus buffers?

I have global-font-lock-mode on; other than that I don't do anything
funky.  I also reproduced the bug with empty .gnus/.emacs files (apart
from setting select methods).  What do you need to reproduce the bug?
Should I MIME-encode the raw article and forward it to you again?

Tom
-- 
T. Lofgren <lofgren@sics.se> - Wherever I lay my .emacs, that's my ${HOME}
Stop Software Piracy -- Write Free Software

Re: PGnus 0.9{5,7} hangs after generating summary buffer

[Lars Magne Ingebrigtsen]

Thomas Lofgren <lofgren@sics.se> writes:

>     >> Now, the article is in a thread by itself, so one would think
>     >> that hiding the threads wouldn't touch it.  However, Gnus
>     >> displays the subject correctly (or at least I don't know it to
>     >> be incorrect) encoded in the summary buffer, but there is a
>     >> newline in it, which (from the font-lock) appears to make Gnus
>     >> believe that the second line is a reply to the first line.
> 
>     Lars> Do you use font-lock in the Gnus buffers?
> 
> I have global-font-lock-mode on; other than that I don't do anything
> funky.

Try switching that off and see whether it helps any.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen