Re: cannot start Gnus with idna.el

Re: cannot start Gnus with idna.el

[Reiner Steib]

[ Sorry for the long delay. ]

On Sat, Mar 29 2008, NAKAJI Hiroyuki wrote:

> I noticed that recent CVS Emacs has a problem. For example, Gnus cannot
> start if idna.el is installed in the load-path directory.
>
> I found the problem under the combination of
>
> o FreeBSD/i386 8.0-CURRENT (Jan 29 2008)
> o Emacs 23.0.60 (CVS HEAD)
> o No Gnus 0.7 (CVS HEAD)
> o GNU Libidn 1.2
>
> I checked out the HEAD yesterday morning (JST), bootstrapped with gcc4
> and ran. And then, typed ( g n u s ) C-j in *scratch* buffer but nothing
> happened.
>
> Next, I evaled (setq debug-on-quit t) and saved the *Backtrace* which
> says the breakpoint is in idna.el:idna-to-ascii-response(). I attach the
> *Backtrace* file.
>
> The "while" in idna-to-ascii-response() seems infinite loop.
>
> (defun idna-to-ascii-response ()
>   (while (and (eq (process-status idna-to-ascii-process) 'run)
>               (null idna-to-ascii-response))
>     (accept-process-output idna-to-ascii-process 1))
>   idna-to-ascii-response)

Isn't this a bug in `idna.el' then?  Cc-ing Simon therefore.

> The idna-to-ascii-response is called from message.el and the failure
> itself occurs in
>
> (idna-to-ascii "räksmörgås")

Does it mean, that you can reproduce the problem when evaluating this
expression after starting "emacs -Q"?

> Of cource, after I removed idna.el from any directories of load-path,
> the error go away. (But there is another problem with Gnus which I'd
> like to report.)

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: cannot start Gnus with idna.el

[NAKAJI Hiroyuki]

Thank you for the comment.

After "make maintainer-clean" in emacs/lisp directory, I successfully
bootstrapped the latest emacs. And then I'm trying this problem again.

>>>>> In <v9y75z98of.fsf@marauder.physik.uni-ulm.de> 
>>>>>	Reiner Steib <reinersteib+gmane@imap.cc> wrote:

> > The "while" in idna-to-ascii-response() seems infinite loop.
> >
> > (defun idna-to-ascii-response ()
> >   (while (and (eq (process-status idna-to-ascii-process) 'run)
> >               (null idna-to-ascii-response))
> >     (accept-process-output idna-to-ascii-process 1))
> >   idna-to-ascii-response)

> Isn't this a bug in `idna.el' then?  Cc-ing Simon therefore.

I'm not sure.

> > The idna-to-ascii-response is called from message.el and the failure
> > itself occurs in
> >
> > (idna-to-ascii "räksmörgås")

> Does it mean, that you can reproduce the problem when evaluating this
> expression after starting "emacs -Q"?

No.

After starting "emacs -Q", loaded /some/where/idna.el, opened
gnus/lisp/message.el (no load of gnus) and evaluated this expression. It
was successfull.

(idna-to-ascii "räksmörgås")
"xn--rksmrgs-5wao1o"

Thanks.
-- 
NAKAJI Hiroyuki

bug#92: cannot start Gnus with idna.el

[Lars Magne Ingebrigtsen]

close 92
thanks

bug#92: cannot start Gnus with idna.el

[Lars Magne Ingebrigtsen]

Reiner Steib <reinersteib+gmane@imap.cc> writes:

>>   (idna-to-ascii "räksmörgås")
>>
>> If idna-to-ascii-response inf-loops, it's a bug in idna.el, the idn
>> program or Emacs' process handling on FreeBSD.
>> CC-ing Simon Josefsson, the author of idna.el and libidn.

This doesn't seem to be reproducible by anyone, so I'm closing the
report.

close 92

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

bug#92: cannot start Gnus with idna.el

[Glenn Morris]

Lars Magne Ingebrigtsen wrote:

> This doesn't seem to be reproducible by anyone, so I'm closing the
> report.
>
> close 92

It doesn't work that way. Either send a mail to 92-done (as I did in
this mail), explaining in the body why the bug is closed; or send

close 92
stop

at the start of the body in a message to control@debbugs.

cannot start Gnus with idna.el

[NAKAJI Hiroyuki]

[-- Attachment #1: Type: text/plain, Size: 980 bytes --]

Hi,

I noticed that recent CVS Emacs has a problem. For example, Gnus cannot
start if idna.el is installed in the load-path directory.

I found the problem under the combination of

o FreeBSD/i386 8.0-CURRENT (Jan 29 2008)
o Emacs 23.0.60 (CVS HEAD)
o No Gnus 0.7 (CVS HEAD)
o GNU Libidn 1.2

I checked out the HEAD yesterday morning (JST), bootstrapped with gcc4
and ran. And then, typed ( g n u s ) C-j in *scratch* buffer but nothing
happened.

Next, I evaled (setq debug-on-quit t) and saved the *Backtrace* which
says the breakpoint is in idna.el:idna-to-ascii-response(). I attach the
*Backtrace* file.

The "while" in idna-to-ascii-response() seems infinite loop.

(defun idna-to-ascii-response ()
  (while (and (eq (process-status idna-to-ascii-process) 'run)
              (null idna-to-ascii-response))
    (accept-process-output idna-to-ascii-process 1))
  idna-to-ascii-response)

The idna-to-ascii-response is called from message.el and the failure
itself occurs in


[-- Attachment #2: Type: text/plain, Size: 32 bytes --]

(idna-to-ascii "räksmörgås")


[-- Attachment #3: Type: text/plain, Size: 256 bytes --]

Of cource, after I removed idna.el from any directories of load-path,
the error go away. (But there is another problem with Gnus which I'd
like to report.)

Regards,

Note: this report is sent using another CVS Emacs several weeks old.
-- 
NAKAJI Hiroyuki

[-- Attachment #4: Backtrace.gnus.20080328 --]
[-- Type: application/octet-stream, Size: 2648 bytes --]