Gleb Smirnoff <[email protected]> writes: > On Tue, Sep 17, 2013 at 10:04:52AM -0400, Dan Espen wrote: > D> > Just got a very different panic. It may be unrelated > D> > to the free/malloc problems, but since fvwm was rock stable > D> > for the last 10 years for me, I'm prone to think that all > D> > my recent crashes are caused by the same single problem. > D> > > D> > (gdb) bt > D> > #0 0x00000000004daab8 in FlocaleDrawString (dpy=0x804831000, > flf=0x80482f980, > D> > fws=0x80482aec0, flags=0) at Flocale.c:1983 > D> ... > D> > Core saved, I can provide more info. > D> > D> Starting to look like a memory overlay somewhere. > D> In which case valgrind or one of the other memory checkers would be > D> more useful. > D> > D> Do "list". > D> Print fws, *fws, comb_chars, i. > > (gdb) list > 1978 while(comb_chars[i].c.byte1 != 0 && > comb_chars[i].c.byte2 != 0) > 1979 { > 1980 /* draw composing character on top of > corresponding > 1981 "real" character */ > 1982 FlocaleWinString tmp_fws = *fws; > 1983 int offset = > pixel_pos[comb_chars[i].position]; > 1984 char *buf2; > 1985 int out_len; > 1986 curr_len = > FlocaleChar2bOneCharToUtf8(comb_chars[i].c, > 1987 buf); > Current language: auto; currently minimal > (gdb) p fws > $1 = (FlocaleWinString *) 0x80482aec0 > (gdb) p *fws > $2 = {str = 0x804a0f230 "Shells", e_str = 0x804a37850 "Shells", str2b = 0x0, > gc = 0x80482f840, colorset = 0x804a13410, win = 18874669, x = 3, y = 33, > len = 0, clip_region = 0x0, flags = {text_rotation = 0, has_colorset = 1, > has_clip_region = 0}} > (gdb) p comb_chars > $3 = (superimpose_char_t *) 0x804a37850 > (gdb) p i > $4 = 0 > (gdb)
p pixel_pos p *comb_chars p char_len -- Dan Espen
