Re: switching font backends in .emacs.desktop causes crash

[Robert Pluim]

martin rudalics <address@hidden> writes:

>>      Thread 1 "emacs" received signal SIGSEGV, Segmentation fault.
>>      0x00007ffff446468d in XftCharIndex () from 
>> /usr/lib/x86_64-linux-gnu/libXft.so.2
>>      (gdb) bt
>>      #0  0x00007ffff446468d in XftCharIndex () at 
>> /usr/lib/x86_64-linux-gnu/libXft.so.2
>>      #1  0x0000555555758689 in xftfont_encode_char (font=<optimized out>, 
>> c=<optimized out>)
>>      at xftfont.c:532
>>      #2 0x00005555555d60a9 in get_char_glyph_code (char2b=<synthetic
>> pointer>, font=0x555557b49470, c=<optimized out>) at xdisp.c:25915
>>      #3  0x00005555555d60a9 in x_produce_glyphs (it=0x7fffffff9b00) at 
>> xdisp.c:28294
>>
>> Here we're using xftfont_encode_char, even though we've just closed
>> the XFT font backend and should be using the X one for the minibuffer?
>> Is the 'it' there caching XFT information?
>>
>> I have no idea how to fix this. I can reproduce at will if more gdb
>> info is needed (BTW, emacs-26 has the same issue).
>>
>>      Lisp Backtrace:
>>      "message" (0xffffc458)
>>      "desktop-read" (0xffffc7e0)
>>      0x563cec70 PVEC_COMPILED
>>      "run-hooks" (0xffffcae0)
>>      "command-line" (0xffffd388)
>>      "normal-top-level" (0xffffd700)
>
> This looks like Bug#23386: Segfault when messing with font-backend.

It is in fact exactly that. I tried the patch there, it works for
me. How come it never went in?

Robert