bug#14264: 24.3.50; SEGV in char_table_ref's use of SUB_CHAR_TABLE_P

[Stefan Monnier]

>> > Sorry, still no cigar.
>> Any other info I can usefully provide?
> Well, it would be useful to find out how come Emacs comes with such a
> bogus character code in the middle of what is supposed to be pure
> ASCII text.  Can you set a breakpoint in scan_for_column, just before
> it calls composition_compute_stop_pos, conditioned on scan = 1645,
> then step inside composition_compute_stop_pos, and see how come it
> fetches such a character code?  Is buffer text being corrupted by
> something?

Any news on this?

I couldn't reproduce the bug right now, but I also couldn't find the
relevant article in my news server.  Can someone still reproduce
this bug?  If so, can he/she provide an updated recipe (maybe using a
public NNTP server, so it doesn't depend on which server one uses.
I guess the gmane server would be a good choice)?


        Stefan