Re: Crash in gc_sweep (SIGH!)

[Kim F. Storm]

Eli Zaretskii <address@hidden> writes:

>> From: "Kim F. Storm" <address@hidden>
>> Date: Mon,  9 Apr 2007 01:32:01 +0200 (CEST)
>> 
>> Following my last bug report on W32 display error, I noticed that
>> the image didn't get through.  So I tried various stuff to decode 
>> the base64 stuff in various ways like this:
>> 
>> - copy it to a new file buffer named x.gif, mark it, do 
>> base64-decode-region, 
>>   and C-c C-c to view the image  (no luck!)
>
> This works for me, if I invoke image-toggle-display by hand (you may
> need to load image-mode before that).  It is bound to C-c C-c only in
> buffers that visit image files.
>
>> - create new file buffer y.gif, set file coding system to binary, copy image
>> data into it, base64 decode it and C-c C-c  (no luck)
>
> Works for me if I do the first part (``create new file buffer y.gif'')
> with "M-x set-visited-file-name".  How did you do that part in your
> case?
>
> It also works to decode the base64 data externally (with `recode'),
> then visit the resulting GIF directly.

What I meant with "no luck" was that it displayed a square box 
instead of the image itself.  

After restarting Emacs, it seems that I can also display the image
using the method that failed before.  Maybe the gif image library
messed something up when I invoked it on bad data, which caused it to
fail even worse on subsequent attemts.

I'm not sure how to see the library version.

It is the one from redhat 9.0 libgif.so.4.1.0,
It has this in the data section:

GIF_LIBRARY     IBMPC  Version 4.0,     Eric S. Raymond,        Jan 24 2003,   
23:33:26


I observed at one time that the base64-decode-region decoded the
data into something different than on other occasions, but maybe
I pasted the wrong base64 data ... it's a little difficult to
say as all base64 looks the same to me :-)


-- 
Kim F. Storm <address@hidden> http://www.cua.dk