"End of file during parsing" in Gcc handling

"End of file during parsing" in Gcc handling

[Andreas Fuchs]

[-- Attachment #1: Type: text/plain, Size: 1116 bytes --]

Hi,

I just tried to send a mail in recent CVS gnus, and a semi-recent CVS
emacs (2002-10-06), to discover that when Gcc handling takes place, I
get an error that is not traceable. toggle-debug-on-error did not
reveal a backtrace. Stepping with edebug shows some _very_ weird
behaviour:

My edebugging usually follows this algorithm:
1. Try to discover which function throws the error
2. Step through to see where in the function it happens.
3. If it is in a (stepped-over) defun, instrument it and continue at 1.

Now, stepping through `gnus-inews-do-gcc', shows me that the error
happens somewhere inside or after a call to
`gnus-request-accept-article'. When I instrument this defun and step
through it, nothing happens until after emacs jumps out of it, when I
get this error: "End of file during parsing" (But no backtrace, grrrr.)

I really don't know what to do to track this bug down, seems like it is
related to Khepris or another bug deity. (-:

-- 
Andreas Fuchs, <asf@acm.org>, asf@jabber.at, antifuchs
Was I helpful?  Let others know:
http://svcs.affero.net/rm.php?r=antifuchs

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]

Re: "End of file during parsing" in Gcc handling

[Simon Josefsson]

Andreas Fuchs <asf@void.at> writes:

> Hi,
>
> I just tried to send a mail in recent CVS gnus, and a semi-recent CVS
> emacs (2002-10-06), to discover that when Gcc handling takes place, I
> get an error that is not traceable. toggle-debug-on-error did not
> reveal a backtrace. Stepping with edebug shows some _very_ weird
> behaviour:

Perhaps (setq debug-on-signal t) gives a backtrace?

Re: "End of file during parsing" in Gcc handling

[Andreas Fuchs]

[-- Attachment #1: Type: text/plain, Size: 883 bytes --]

Today, Simon Josefsson <jas@extundo.com> wrote:
>> I just tried to send a mail in recent CVS gnus, and a semi-recent CVS
>> emacs (2002-10-06), to discover that when Gcc handling takes place, I
>> get an error that is not traceable. toggle-debug-on-error did not
>> reveal a backtrace. Stepping with edebug shows some _very_ weird
>> behaviour:
> 
> Perhaps (setq debug-on-signal t) gives a backtrace?

Nope, and neither does -on-quit. I fear emacs is out of its mind. Or I
am.

I tried this with an older emacs (debian's) with current CVS gnus, and
there the problem is present, too (but only when I load my gnus config
)-:).

The message doesn't seem to come from any of the gnus sources. WTF?

Thanks for your suggestions,
-- 
Andreas Fuchs, <asf@acm.org>, asf@jabber.at, antifuchs
Was I helpful?  Let others know:
http://svcs.affero.net/rm.php?r=antifuchs

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]

Re: "End of file during parsing" in Gcc handling

[Paul Jarc]

Andreas Fuchs <asf@void.at> wrote:
> Now, stepping through `gnus-inews-do-gcc', shows me that the error
> happens somewhere inside or after a call to
> `gnus-request-accept-article'. When I instrument this defun and step
> through it, nothing happens until after emacs jumps out of it, when I
> get this error: "End of file during parsing" (But no backtrace, grrrr.)

Random guess: do you use any advice for that function?


paul

Re: "End of file during parsing" in Gcc handling

[Andreas Fuchs]

[-- Attachment #1: Type: text/plain, Size: 999 bytes --]

Today, Paul Jarc <prj@po.cwru.edu> wrote:
>> Now, stepping through `gnus-inews-do-gcc', shows me that the error
>> happens somewhere inside or after a call to
>> `gnus-request-accept-article'. When I instrument this defun and step
>> through it, nothing happens until after emacs jumps out of it, when I
>> get this error: "End of file during parsing" (But no backtrace,
>> grrrr.)
> 
> Random guess: do you use any advice for that function?

Seems not to be so.

,----[ `gnus-request-accept-article' function documentation ]
| gnus-request-accept-article is a Lisp function in `gnus-int'.
| (gnus-request-accept-article GROUP &optional GNUS-COMMAND-METHOD LAST NO-ENCODE)
| 
| Not documented.
`----

Obviously not. Maybe my draft groups are broken? Can I regenerate their
active files or overview database without causing harm?

Thanks,
-- 
Andreas Fuchs, <asf@acm.org>, asf@jabber.at, antifuchs
Was I helpful?  Let others know:
http://svcs.affero.net/rm.php?r=antifuchs

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]

Re: "End of file during parsing" in Gcc handling

[Simon Josefsson]

Andreas Fuchs <asf@void.at> writes:

> Today, Simon Josefsson <jas@extundo.com> wrote:
>>> I just tried to send a mail in recent CVS gnus, and a semi-recent CVS
>>> emacs (2002-10-06), to discover that when Gcc handling takes place, I
>>> get an error that is not traceable. toggle-debug-on-error did not
>>> reveal a backtrace. Stepping with edebug shows some _very_ weird
>>> behaviour:
>> 
>> Perhaps (setq debug-on-signal t) gives a backtrace?
>
> Nope, and neither does -on-quit. I fear emacs is out of its mind. Or I
> am.
>
> I tried this with an older emacs (debian's) with current CVS gnus, and
> there the problem is present, too (but only when I load my gnus config
> )-:).
>
> The message doesn't seem to come from any of the gnus sources. WTF?

The error is from the emacs lisp reader, so most likely it is a
missing parentheses error or similar somewhere.  Maybe the GCC
contains unbalanced quotes or parentheses?

Re: "End of file during parsing" in Gcc handling

[Andreas Fuchs]

[-- Attachment #1: Type: text/plain, Size: 823 bytes --]

Today, Simon Josefsson <jas@extundo.com> wrote:
>> The message doesn't seem to come from any of the gnus sources. WTF?
> 
> The error is from the emacs lisp reader, so most likely it is a
> missing parentheses error or similar somewhere.  Maybe the GCC
> contains unbalanced quotes or parentheses?

Gcc: nnfolder+archive:mail.2002-10

is my Gcc header. I don't notice anything unusual about
it. Hmm. I just thought up a possible cause: had a power failure some
days ago (gnus was running), and when the machine came up again, I had
several ancient, long ago killed and unactivatable nnarchive
groups. The active file was messed up, as well. Could that be the
cause?


-- 
Andreas Fuchs, <asf@acm.org>, asf@jabber.at, antifuchs
Was I helpful?  Let others know:
http://svcs.affero.net/rm.php?r=antifuchs

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]

Re: "End of file during parsing" in Gcc handling

[Kai Großjohann]

Andreas Fuchs <asf@void.at> writes:

> is my Gcc header. I don't notice anything unusual about
> it. Hmm. I just thought up a possible cause: had a power failure some
> days ago (gnus was running), and when the machine came up again, I had
> several ancient, long ago killed and unactivatable nnarchive
> groups. The active file was messed up, as well. Could that be the
> cause?

How about hitting `g' over the server in the server buffer?

kai
-- 
~/.signature is: umop ap!sdn    (Frank Nobis)

Re: "End of file during parsing" in Gcc handling

[Andreas Fuchs]

[-- Attachment #1: Type: text/plain, Size: 778 bytes --]

Today, Kai GroÃ.johann <Kai.Grossjohann@cs.uni-dortmund.de> wrote:
>> is my Gcc header. I don't notice anything unusual about it. Hmm. I
>> just thought up a possible cause: had a power failure some days ago
>> (gnus was running), and when the machine came up again, I had several
>> ancient, long ago killed and unactivatable nnarchive groups. The
>> active file was messed up, as well. Could that be the cause?
> How about hitting `g' over the server in the server buffer?

Hm, did that. Seems like that was not the cause - I still get the error
message. )-:

Arghl, I'm out of ideas. Does anybody else have these problems?

-- 
Andreas Fuchs, <asf@acm.org>, asf@jabber.at, antifuchs
Was I helpful?  Let others know:
http://svcs.affero.net/rm.php?r=antifuchs

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]

Re: "End of file during parsing" in Gcc handling

[Andreas Fuchs]

[-- Attachment #1: Type: text/plain, Size: 518 bytes --]

Today, Kai Großjohann <Kai.Grossjohann@cs.uni-dortmund.de> wrote:
> How about hitting `g' over the server in the server buffer?

Ah, I found it. Couldn't enter my latest archive group, either, and
looked at the .nov file to it. Naturally, it was full of NULL bytes,
which I deleted.

I love filesystem corruption.

Thanks for all your help,
a happy (again) Gnuser (:

-- 
Andreas Fuchs, <asf@acm.org>, asf@jabber.at, antifuchs
Was I helpful?  Let others know:
http://svcs.affero.net/rm.php?r=antifuchs

[-- Attachment #2: Type: application/pgp-signature, Size: 188 bytes --]