bug#3659: FW: 23.0.95; crash on set-frame-font

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

Kenichi Handa <handa@m17n.org> writes:

> At the moment, no.  But, I'll investigate the code in a
> hurry.

Thanks.  Glenn has posted a backtrace in the bug tracker:

http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=3659

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <87eit9ugs9.fsf@stupidchicken.com>, Chong Yidong <cyd@stupidchicken.com> writes:

> Kenichi Handa <handa@m17n.org> writes:
> > At the moment, no.  But, I'll investigate the code in a
> > hurry.

> Thanks.  Glenn has posted a backtrace in the bug tracker:

> http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=3659

Glenn wrote:

> The May 1 and March 1 sources crash in the same way. The Jan 1 and Feb
> 1 sources don't build.

But Simon wrote:

> I narrowed it down to between 2009-05-21 (ok) and 2009-05-22 (coredump)
> via:

Doesn't it mean that the Glenn's crash of May 1 source is
different from Simon's crash of May 21 source?

Among the differeces of src/ChangeLog shown by Simon, these
seem harmless.

> 2009-05-21  Kenichi Handa  <handa@m17n.org>
>
> 	* fontset.c (Qlatin): Don't make it static.
>
> 2009-05-19  Kenichi Handa  <handa@m17n.org>
>
> 	* font.c (font_sort_entities): Rename from font_sort_entites.
> 	Callers changed.
>
> 2009-05-18  Kenichi Handa  <handa@m17n.org>
>
> 	* font.c (font_find_for_lface): Copy SPEC's FONT_TYPE too.

And the rest are changes to xfont.c for x font-backend.
But, Glenn's crash is in the code for xft font-backend here:

#0  0xfd3c8b0c in XftLockFace () from /usr/openwin/lib/libXft.so.2
#1  0x0035acdc in xftfont_open (f=0x6794d8, entity=-2133415224, pixel_size=14)
    at xftfont.c:290

This also suggests that their crashes are the different one.

Could you please try to run Emacs as this?

% emacs --xrm 'Emacs*FontBackend: x'

and

% emacs --xrm 'Emacs*FontBackend: xft'

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Kenichi Handa wrote:

> % emacs --xrm 'Emacs*FontBackend: x'

I cannot find a font that makes set-frame-font crash. However,
S-Mouse-1 'Increase buffer text size' about 5 times still causes a
crash (a separate issue?).

Program received signal SIGSEGV, Segmentation fault.
0xfcf31c60 in strcmp () from /lib/libc.so.1
(gdb) bt
#0  0xfcf31c60 in strcmp () from /lib/libc.so.1
#1  0xfd997600 in ComparePixmapDatas () from /usr/dt/lib/libXm.so.4
#2  0xfda8c8f0 in _XmRemoveHashEntry () from /usr/dt/lib/libXm.so.4
#3  0xfd9f4160 in XmDestroyPixmap () from /usr/dt/lib/libXm.so.4
#4  0xfd9cf6d4 in _XmArrowPixmapCacheDelete () from /usr/dt/lib/libXm.so.4
#5  0xfd9ce964 in Destroy () from /usr/dt/lib/libXm.so.4
#6  0xfd874390 in Phase2Destroy () from /usr/openwin/lib/libXt.so.4
#7  0xfd873a78 in Recursive () from /usr/openwin/lib/libXt.so.4
#8  0xfd873a78 in Recursive () from /usr/openwin/lib/libXt.so.4
#9  0xfd873ac8 in Recursive () from /usr/openwin/lib/libXt.so.4
#10 0xfd87408c in XtPhase2Destroy () from /usr/openwin/lib/libXt.so.4
#11 0xfd873c84 in _XtDoPhase2Destroy () from /usr/openwin/lib/libXt.so.4
#12 0xfd873a00 in XtDestroyWidget () from /usr/openwin/lib/libXt.so.4
#13 0x00368a30 in xm_destroy_instance (instance=0x965210) at lwlib-Xm.c:1747
#14 0x00362694 in destroy_one_instance (instance=0x965210)
    at lwlib/lwlib.c:1045
#15 0x00362788 in lw_destroy_all_widgets (id=65541)
    at lwlib/lwlib.c:1090
#16 0x000c9b68 in pop_down_menu (arg=-1600996040) at xmenu.c:1853
#17 0x0029c868 in unbind_to (count=10, value=542261264) at eval.c:3400
#18 0x000c9f58 in create_and_show_popup_menu (f=0x671c80, first_wv=0x994bc0, 
    x=500, y=422, for_click=1) at xmenu.c:1931
#19 0x000cac9c in xmenu_show (f=0x671c80, x=244, y=211, for_click=1, 
    keymaps=1, title=1615002964, error=0xffbf85e0) at xmenu.c:2152
#20 0x000c7678 in Fx_popup_menu (position=-1601031168, menu=-1596750128)
    at xmenu.c:505
#21 0x0029b330 in Ffuncall (nargs=3, args=0xffbf8720) at eval.c:3048
#22 0x0030adc4 in Fbyte_code (bytestr=1615002564, vector=-2143093804, 
    maxdepth=9) at bytecode.c:678
#23 0x0029bedc in funcall_lambda (fun=-2143093860, nargs=1, 
    arg_vector=0xffbf8a14) at eval.c:3232
#24 0x0029b5c0 in Ffuncall (nargs=2, args=0xffbf8a10) at eval.c:3091
#25 0x0029375c in Fcall_interactively (function=546805920, 
    record_flag=542261264, keys=-2142035960) at callint.c:868
#26 0x0029b374 in Ffuncall (nargs=4, args=0xffbf8ca8) at eval.c:3051
#27 0x0029aaf0 in call3 (fn=542446264, arg1=546805920, arg2=542261264, 
    arg3=542261264) at eval.c:2871
#28 0x001d31d8 in Fcommand_execute (cmd=546805920, record_flag=542261264, 
    keys=542261264, special=542261264) at keyboard.c:10449
#29 0x001bb474 in command_loop_1 () at keyboard.c:1901
#30 0x002973cc in internal_condition_case (bfun=0x1b8c78 <command_loop_1>, 
    handlers=542323144, hfun=0x1b8114 <cmd_error>) at eval.c:1512
#31 0x001b8768 in command_loop_2 () at keyboard.c:1359
#32 0x00296b5c in internal_catch (tag=542319272, 
    func=0x1b8740 <command_loop_2>, arg=542261264) at eval.c:1248
#33 0x001b86e8 in command_loop () at keyboard.c:1338
#34 0x001b7964 in recursive_edit_1 () at keyboard.c:953
#35 0x001b7c40 in Frecursive_edit () at keyboard.c:1015
#36 0x001b547c in main (argc=4, argv=0xffbf943c) at emacs.c:1852

Lisp Backtrace:
"x-popup-menu" (0xffbf8724)
"mouse-appearance-menu" (0xffbf8a14)
"call-interactively" (0xffbf8cac)




> % emacs --xrm 'Emacs*FontBackend: xft'

emacs -Q --xrm 'Emacs*FontBackend: xft' \
  --font '-Adobe-Courier-normal-normal-normal-*-14-*-*-*-m-90-iso10646-1'

(I have to specify a font else Emacs complains about my normal
misc-fixed not being defined.)

crashes on startup with:

Program received signal SIGSEGV, Segmentation fault.
0xfd3c8b0c in XftLockFace () from /usr/openwin/lib/libXft.so.2
(gdb) bt
#0  0xfd3c8b0c in XftLockFace () from /usr/openwin/lib/libXft.so.2
#1  0x0035a728 in xftfont_open (f=0x66b810, entity=-2134315232, pixel_size=14)
    at xftfont.c:290
#2  0x002bf44c in font_open_entity (f=0x66b810, entity=-2134315232, 
    pixel_size=14) at font.c:2926
#3  0x002c1c8c in font_open_for_lface (f=0x66b810, entity=-2134315232, 
    attrs=0xffbf7408, spec=-2133418072) at font.c:3429
#4  0x002c1d50 in font_load_for_lface (f=0x66b810, attrs=0xffbf7408, 
    spec=-2133418072) at font.c:3454
#5  0x0015f494 in realize_x_face (cache=0xc902b8, attrs=0xffbf7408)
    at xfaces.c:5892
#6  0x0015eff8 in realize_face (cache=0xc902b8, attrs=0xffbf7408, 
    former_face_id=-1) at xfaces.c:5781
#7  0x0015bde0 in lookup_face (f=0x66b810, attr=0xffbf7408) at xfaces.c:4723
#8  0x00160fa0 in face_at_string_position (w=0x676910, string=1616937248, 
    pos=1, bufpos=0, region_beg=-1, region_end=-1, endptr=0xffbf7548, 
    base_face_id=MODE_LINE_FACE_ID, mouse_p=0) at xfaces.c:6532
#9  0x000ac10c in display_string (string=0x5269f0 "*scratch*", 
    lisp_string=542261264, face_string=1616937248, face_string_pos=1, start=0, 
    it=0xffbf7900, field_width=12, precision=-8, max_x=0, multibyte=0)
    at xdisp.c:18928
#10 0x000a8788 in display_mode_element (it=0xffbf7900, depth=4, field_width=0, 
    precision=-8, elt=1616937248, props=542261264, risky=0) at xdisp.c:17678
#11 0x000a8e40 in display_mode_element (it=0xffbf7900, depth=3, field_width=0, 
    precision=-8, elt=-1604295440, props=542261264, risky=0) at xdisp.c:17850
#12 0x000a8e40 in display_mode_element (it=0xffbf7900, depth=1, field_width=0, 
    precision=0, elt=-1604293320, props=542261264, risky=0) at xdisp.c:17850
#13 0x000a7784 in display_mode_line (w=0x676910, face_id=MODE_LINE_FACE_ID, 
    format=-1604293400) at xdisp.c:17362
#14 0x000a74a0 in display_mode_lines (w=0x676910) at xdisp.c:17307
#15 0x000a718c in redisplay_mode_lines (window=-2140706544, force=0)
    at xdisp.c:17265
#16 0x0008c1dc in echo_area_display (update_frame_p=1) at xdisp.c:9209
#17 0x00088ef0 in message3_nolog (m=1616680176, nbytes=65, multibyte=0)
    at xdisp.c:8083
#18 0x00088a8c in message3 (m=1616680176, nbytes=65, multibyte=0)
    at xdisp.c:8018
#19 0x0028b294 in Fmessage (nargs=2, args=0xffbf8244) at editfns.c:3389
#20 0x0029b13c in Ffuncall (nargs=3, args=0xffbf8240) at eval.c:3026
#21 0x0030adc4 in Fbyte_code (bytestr=1614711308, vector=-2143385060, 
    maxdepth=3) at bytecode.c:678
#22 0x0029bedc in funcall_lambda (fun=-2143385100, nargs=0, 
    arg_vector=0xffbf8504) at eval.c:3232
#23 0x0029b5c0 in Ffuncall (nargs=1, args=0xffbf8500) at eval.c:3091
#24 0x0030adc4 in Fbyte_code (bytestr=1614712136, vector=-2143384232, 
    maxdepth=10) at bytecode.c:678
#25 0x0029bedc in funcall_lambda (fun=-2143384280, nargs=1, 
    arg_vector=0xffbf87dc) at eval.c:3232
#26 0x0029b5c0 in Ffuncall (nargs=2, args=0xffbf87d8) at eval.c:3091
#27 0x0030adc4 in Fbyte_code (bytestr=1614698544, vector=-2143397824, 
    maxdepth=7) at bytecode.c:678
#28 0x0029bedc in funcall_lambda (fun=-2143397864, nargs=0, 
    arg_vector=0xffbf8aac) at eval.c:3232
#29 0x0029b5c0 in Ffuncall (nargs=1, args=0xffbf8aa8) at eval.c:3091
#30 0x0030adc4 in Fbyte_code (bytestr=1614695204, vector=-2143401164, 
    maxdepth=6) at bytecode.c:678
#31 0x0029bedc in funcall_lambda (fun=-2143401204, nargs=0, 
    arg_vector=0xffbf8cb8) at eval.c:3232
#32 0x0029b934 in apply_lambda (fun=-2143401204, args=542261264, eval_flag=1)
    at eval.c:3156
#33 0x00299ba0 in Feval (form=-1604600808) at eval.c:2418
#34 0x001b87bc in top_level_2 () at keyboard.c:1368
#35 0x002973cc in internal_condition_case (bfun=0x1b87a4 <top_level_2>, 
    handlers=542323144, hfun=0x1b8114 <cmd_error>) at eval.c:1512
#36 0x001b8818 in top_level_1 () at keyboard.c:1376
#37 0x00296b5c in internal_catch (tag=542319272, func=0x1b87cc <top_level_1>, 
    arg=542261264) at eval.c:1248
#38 0x001b86c0 in command_loop () at keyboard.c:1331
#39 0x001b7964 in recursive_edit_1 () at keyboard.c:953
#40 0x001b7c40 in Frecursive_edit () at keyboard.c:1015
#41 0x001b547c in main (argc=6, argv=0xffbf93ec) at emacs.c:1852

Lisp Backtrace:
"message" (0xffbf8244)
"display-startup-echo-area-message" (0xffbf8504)
"command-line-1" (0xffbf87dc)
"command-line" (0xffbf8aac)
"normal-top-level" (0xffbf8cb8)

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

I installed opensolaris 5.11 on x86 PC, but couldn't
reproduce these bugs.  Perhaps, they are specific to sparc
architecture.

In article <gmzlbwtb8n.fsf@fencepost.gnu.org>, Glenn Morris <rgm@gnu.org> writes:

> > % emacs --xrm 'Emacs*FontBackend: x'

> I cannot find a font that makes set-frame-font crash. However,
> S-Mouse-1 'Increase buffer text size' about 5 times still causes a
> crash (a separate issue?).

> Program received signal SIGSEGV, Segmentation fault.
> 0xfcf31c60 in strcmp () from /lib/libc.so.1
> (gdb) bt
> #0  0xfcf31c60 in strcmp () from /lib/libc.so.1
> #1  0xfd997600 in ComparePixmapDatas () from /usr/dt/lib/libXm.so.4
> #2  0xfda8c8f0 in _XmRemoveHashEntry () from /usr/dt/lib/libXm.so.4
> #3  0xfd9f4160 in XmDestroyPixmap () from /usr/dt/lib/libXm.so.4
> #4  0xfd9cf6d4 in _XmArrowPixmapCacheDelete () from /usr/dt/lib/libXm.so.4
> #5  0xfd9ce964 in Destroy () from /usr/dt/lib/libXm.so.4
> #6  0xfd874390 in Phase2Destroy () from /usr/openwin/lib/libXt.so.4
> #7  0xfd873a78 in Recursive () from /usr/openwin/lib/libXt.so.4
> #8  0xfd873a78 in Recursive () from /usr/openwin/lib/libXt.so.4
> #9  0xfd873ac8 in Recursive () from /usr/openwin/lib/libXt.so.4
> #10 0xfd87408c in XtPhase2Destroy () from /usr/openwin/lib/libXt.so.4
> #11 0xfd873c84 in _XtDoPhase2Destroy () from /usr/openwin/lib/libXt.so.4
> #12 0xfd873a00 in XtDestroyWidget () from /usr/openwin/lib/libXt.so.4
> #13 0x00368a30 in xm_destroy_instance (instance=0x965210) at lwlib-Xm.c:1747

I'm not 100% sure, but it seems that this bug is not related
to x font-backend.

[...]
> > % emacs --xrm 'Emacs*FontBackend: xft'>

> emacs -Q --xrm 'Emacs*FontBackend: xft' \
>   --font '-Adobe-Courier-normal-normal-normal-*-14-*-*-*-m-90-iso10646-1'

> (I have to specify a font else Emacs complains about my normal
> misc-fixed not being defined.)

> crashes on startup with:

> Program received signal SIGSEGV, Segmentation fault.
> 0xfd3c8b0c in XftLockFace () from /usr/openwin/lib/libXft.so.2
> (gdb) bt
> #0  0xfd3c8b0c in XftLockFace () from /usr/openwin/lib/libXft.so.2
> #1  0x0035a728 in xftfont_open (f=0x66b810, entity=-2134315232, pixel_size=14)
>     at xftfont.c:290
> #2  0x002bf44c in font_open_entity (f=0x66b810, entity=-2134315232, 
>     pixel_size=14) at font.c:2926

In a debugger, please check the values of filename
(LispString), index (LispInt), and xftfont (XftFont *) just
before calling XftLockFace at xftfont.c:290.

By the way, are there any fonts with that you can startup
Emacs with --xrm 'Emacs*FontBackend: xft'?

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Kenichi Handa wrote:

> In a debugger, please check the values of filename (LispString),
> index (LispInt), and xftfont (XftFont *) just before calling
> XftLockFace at xftfont.c:290.

A breakpoint in this area is triggered 4 times before the Emacs frame
appears, then once more after the frame appears. It crashes on the 5th
time.

The 5 values of filename are:
/usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
/usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
/usr/openwin/lib/X11/fonts/75dpi/courO14.pcf.Z
/usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
/usr/openwin/lib/X11/fonts/100dpi/courB10.pcf.Z

"index" is always 0.

"xftfont" is always 0 (perhaps I am printing it wrong? I use "pp".)

> By the way, are there any fonts with that you can startup
> Emacs with --xrm 'Emacs*FontBackend: xft'?

I didn't check them all, but these work:

-*-FreeMono-normal-normal-normal-*-*-*-*-*-m-0-iso10646-1
-monotype-Times New-Roman-normal-normal-normal-*-*-*-*-*-*-0-iso10646-1

These are both TrueType fonts. I did not find any of the fonts in the
75dpi directory that worked (but it was not an exhaustive search).

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

> I installed opensolaris 5.11 on x86 PC, but couldn't
> reproduce these bugs.  Perhaps, they are specific to sparc
> architecture.

I got access to a sparc, but was unable to reproduce the crash:

  SunOS m56-129-3.mit.edu 5.10 Generic_118833-36 sun4u sparc
  SUNW,Sun-Blade-1500

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

I forgot to mention the following configuration information about the
system that I tried, and failed, to reproduce this:

In GNU Emacs 23.0.95.2 (sparc-sun-solaris2.10, Motif Version 2.1.0)
 of 2009-06-26 on m56-129-3.mit.edu
Windowing system distributor `Sun Microsystems, Inc.', version 11.0.6620
configured using `configure  '--with-gif=no' '--with-x-toolkit=motif' 'CFLAGS=-g''

I noticed that Simon Marshall has the following:

In GNU Emacs 23.0.94.1 (sparc-sun-solaris2.8, Motif Version 2.1.0)
 of 2009-06-01 on risksun2
Windowing system distributor `Hummingbird Ltd.', version 11.0.100015
configured using `configure  '--x-includes=/usr/openwin/include:/usr/dt/ include:/usr/local/include:/usr/local/X11/include' '--x-libraries=/usr/openwin/lib:/usr/dt/lib:/usr/local/lib:/usr/local/X11/lib' '--with-x-toolkit=motif' '--with-jpeg=no' '--with-gif=no' '--with-tiff=no' 'CFLAGS=-g''

It's possible the Hummingbird X server is somehow responsible.  Simon,
could you try to compile and run using Sun's X?  Glenn, what X server
are you using, and what's your configuration?

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

Glenn Morris <rgm@gnu.org> writes:

> "xftfont" is always 0 (perhaps I am printing it wrong? I use "pp".)

Does the following patch prevent the crash?

*** trunk/src/xftfont.c.~1.24.~	2009-05-11 12:06:41.000000000 -0400
--- trunk/src/xftfont.c	2009-06-26 23:15:32.000000000 -0400
***************
*** 287,300 ****
    match = XftFontMatch (display, FRAME_X_SCREEN_NUMBER (f), pat, &result);
    FcPatternDestroy (pat);
    xftfont = XftFontOpenPattern (display, match);
-   ft_face = XftLockFace (xftfont);
-   UNBLOCK_INPUT;
- 
    if (! xftfont)
      {
        XftPatternDestroy (match);
        return Qnil;
      }
    /* We should not destroy PAT here because it is kept in XFTFONT and
       destroyed automatically when XFTFONT is closed.  */
    font_object = font_make_object (VECSIZE (struct xftfont_info), entity, size);
--- 287,301 ----
    match = XftFontMatch (display, FRAME_X_SCREEN_NUMBER (f), pat, &result);
    FcPatternDestroy (pat);
    xftfont = XftFontOpenPattern (display, match);
    if (! xftfont)
      {
+       UNBLOCK_INPUT;
        XftPatternDestroy (match);
        return Qnil;
      }
+   ft_face = XftLockFace (xftfont);
+   UNBLOCK_INPUT;
+ 
    /* We should not destroy PAT here because it is kept in XFTFONT and
       destroyed automatically when XFTFONT is closed.  */
    font_object = font_make_object (VECSIZE (struct xftfont_info), entity, size);

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Chong Yidong wrote:

> It's possible the Hummingbird X server is somehow responsible.  Simon,
> could you try to compile and run using Sun's X?  Glenn, what X server
> are you using, and what's your configuration?

Sorry for not providing this info. Sorry also if I have hijacked
Simon's bug report with something unrelated.

I can only display by ssh'ing into the Sun machine and forwarding to
my GNU/Linux machine.

In GNU Emacs 23.0.95.2 (sparc-sun-solaris2.10, Motif Version 2.1.0)
 of 2009-06-25 on foo
Windowing system distributor `The X.Org Foundation', version 11.0.10402000
configured using `configure  '--with-x-toolkit=motif' 'CFLAGS=-O0 -g3''

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Chong Yidong wrote:

> Does the following patch prevent the crash?

Yes; but the buffer-name in the mode-line displays as a series of
empty boxes. So do the fonts: bold, bold-italic, and others.

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

Glenn Morris <rgm@gnu.org> writes:

> Chong Yidong wrote:
>
>> Does the following patch prevent the crash?
>
> Yes; but the buffer-name in the mode-line displays as a series of
> empty boxes. So do the fonts: bold, bold-italic, and others.

So, on this system, XftFontMatch returns a match that XftFontOpenPattern
is unable to open.  That's strange.  Maybe you have a misconfigured X
server?

Handa-san, have you any ideas?

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <nkskhm9yyp.fsf@fencepost.gnu.org>, Glenn Morris <rgm@gnu.org> writes:

> Kenichi Handa wrote:
> > In a debugger, please check the values of filename (LispString),
> > index (LispInt), and xftfont (XftFont *) just before calling
> > XftLockFace at xftfont.c:290.

> A breakpoint in this area is triggered 4 times before the Emacs frame
> appears, then once more after the frame appears. It crashes on the 5th
> time.

> The 5 values of filename are:
> /usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
> /usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
> /usr/openwin/lib/X11/fonts/75dpi/courO14.pcf.Z
> /usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
> /usr/openwin/lib/X11/fonts/100dpi/courB10.pcf.Z

> "index" is always 0.

> "xftfont" is always 0 (perhaps I am printing it wrong? I use "pp".)

You should use the normal p(rint) command for xftfont
because it is not a Lisp object.  But, if "pp xftfont" shows
0, xftfont is actually NULL.

> > By the way, are there any fonts with that you can startup
> > Emacs with --xrm 'Emacs*FontBackend: xft'?

> I didn't check them all, but these work:

> -*-FreeMono-normal-normal-normal-*-*-*-*-*-m-0-iso10646-1
> -monotype-Times New-Roman-normal-normal-normal-*-*-*-*-*-*-0-iso10646-1

> These are both TrueType fonts. I did not find any of the fonts in the
> 75dpi directory that worked (but it was not an exhaustive search).

Hmmm, it seems that there's something wrong in you Xft (or
underlying fontconfig/freetype)

Please compile the attached program and run it (as below).
What is shown?

% gcc `pkg-config --cflags xft` `pkg-config --libs xft` -o xfttest xfttest.c 
% ./xfttest /usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z

---
Kenichi Handa
handa@m17n.org

--- xfttest.c ---
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <X11/Xlib.h>
#include <X11/Xft/Xft.h>

int
main (int argc, char **argv)
{
  Display *display;
  FcPattern *pat, *match;
  FcObjectSet *objset;
  FcFontSet *fontset;
  FcResult result;
  XftFont *font;
      
  display = XOpenDisplay (NULL);
  if (! display)
    {
      fprintf (stderr, "Can't open a display\n");
      exit (1);
    }
  if (argc < 2)
    {
      fprintf (stderr, "Usage: xfttest FONT-FILE-NAME\n");
      exit (1);
    }

  printf ("Listing a font by file name and index ... ");
  pat = FcPatternBuild (0, FC_FILE, FcTypeString,
			(FcChar8 *) argv[1],
			FC_INDEX, FcTypeInteger, 0, NULL);
  objset = FcObjectSetBuild (FC_FOUNDRY, FC_FAMILY, FC_WEIGHT, FC_SLANT,
			     FC_WIDTH, FC_PIXEL_SIZE, FC_SPACING, FC_SCALABLE,
			     FC_STYLE, FC_FILE, FC_INDEX, NULL);
  fontset = FcFontList (NULL, pat, objset);
  if (fontset && fontset->nfont > 0)
    {
      int i;

      printf ("found these fonts:\n");
      for (i = 0; i < fontset->nfont; i++)
	{
	  FcChar8 *file;

	  printf ("  %02d:", i + 1);
	  if (FcPatternGetString (fontset->fonts[i], FC_FILE, 0, &file)
	      == FcResultMatch)
	    printf ("%s\n", (char *) file);
	  else
	    printf ("no filename\n");
	}
      printf ("Matching the first one ... ");
      match = XftFontMatch (display, 0, fontset->fonts[0], &result);
      if (match)
	{
	  FcChar8 *file;
	  FcPatternGetString (match, FC_FILE, 0, &file);
	  printf (" matched\n  %s opening ... ", file);
	  font = XftFontOpenPattern (display, match);
	  if (font)
	    {
	      printf ("ok\n");
	      XftFontClose (display, font);
	    }
	  else
	    {
	      printf ("no\n");
	    }
	}
      else
	{
	  printf (" not matched\n");
	}
    }
  else
    {
      printf (" none\n");
    }

  exit (0);
}

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Kenichi Handa wrote:

> Hmmm, it seems that there's something wrong in you Xft (or
> underlying fontconfig/freetype)

That's entirely possible. I never normally use Solaris, I was only
trying to reproduce the initial report.

> Please compile the attached program and run it (as below).
> What is shown?
>
> % gcc `pkg-config --cflags xft` `pkg-config --libs xft` -o xfttest xfttest.c 

In case it's relevant, I have to manually add "-lX11" to get it to compile.
It then compiles with these warnings:

In file included from ./xfttest.c:4:
/usr/openwin/include/X11/Xlib.h:38: warning: ignoring #pragma ident 
In file included from /usr/openwin/include/X11/extensions/Xrender.h:33,
                 from /usr/openwin/include/X11/Xft/Xft.h:44,
                 from ./xfttest.c:5:
/usr/openwin/include/X11/Xutil.h:56: warning: ignoring #pragma ident 

> % ./xfttest /usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z

Listing a font by file name and index ... found these fonts:
  01:/usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
Matching the first one ...  matched
  /usr/openwin/lib/X11/fonts/Type1/cour.pfa opening ... ok

(the latter is a symlink to ./outline/Courier-Bold.pfa)

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <gi7hyvk083.fsf@fencepost.gnu.org>, Glenn Morris <rgm@gnu.org> writes:

> > % ./xfttest /usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z

> Listing a font by file name and index ... found these fonts:
>   01:/usr/openwin/lib/X11/fonts/75dpi/courR14.pcf.Z
> Matching the first one ...  matched
>   /usr/openwin/lib/X11/fonts/Type1/cour.pfa opening ... ok

> (the latter is a symlink to ./outline/Courier-Bold.pfa)

Ummm, it seems that there's no way to open a font by
specifying file and index with your Xft.  What is the value
of these macros in header files?

/usr/openwin/include/X11/Xft/Xft.h:

#define XFT_MAJOR	2
#define XFT_MINOR	1
#define XFT_REVISION	12

.../include/fontconfig/fontconfig.h:

#define FC_MAJOR	2
#define FC_MINOR	6
#define FC_REVISION	0

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Kenichi Handa wrote:

> Ummm, it seems that there's no way to open a font by specifying file
> and index with your Xft. What is the value of these macros in header
> files?
>
> /usr/openwin/include/X11/Xft/Xft.h:
>
> #define XFT_MAJOR	2
> #define XFT_MINOR	1
> #define XFT_REVISION	12

#define XFT_MAJOR       2
#define XFT_MINOR       1
#define XFT_REVISION    6

> .../include/fontconfig/fontconfig.h:
>
> #define FC_MAJOR	2
> #define FC_MINOR	6
> #define FC_REVISION	0

#define FC_MAJOR        2
#define FC_MINOR        2
#define FC_REVISION     3

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <pf3a9ianyp.fsf@fencepost.gnu.org>, Glenn Morris <rgm@gnu.org> writes:

> > /usr/openwin/include/X11/Xft/Xft.h:
> >
> > #define XFT_MAJOR	2
> > #define XFT_MINOR	1
> > #define XFT_REVISION	12

> #define XFT_MAJOR       2
> #define XFT_MINOR       1
> #define XFT_REVISION    6

> > .../include/fontconfig/fontconfig.h:
> >
> > #define FC_MAJOR	2
> > #define FC_MINOR	6
> > #define FC_REVISION	0

> #define FC_MAJOR        2
> #define FC_MINOR        2
> #define FC_REVISION     3

I don't know how old Xft 2.1.6 is, but, at least, fontconfig
2.2.3 is very old, it's released 5 years ago.

Is it possible to install the latest version of fontconfig,
and try with it?

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Glenn Morris]

Kenichi Handa wrote:

> I don't know how old Xft 2.1.6 is, but, at least, fontconfig
> 2.2.3 is very old, it's released 5 years ago.
>
> Is it possible to install the latest version of fontconfig,
> and try with it?

I tried, but I ran into dependency problems, and I could not get
everything to work right. If the conclusion is that the system I used
is odd or too old (though it passes all the configure tests), that's
fine by me.

Simon's problem might be different though. Sorry for the diversion!

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

Guys, I waited as I wasn't quite sure if we went thru a diversion on
this thread or not.

In any case, 23.0.96 still has the crash on set-frame-font introduced
between 2009-05-21 (ok) and 2009-05-22 (coredump). 

-----Original Message-----
From: Glenn Morris [mailto:rgm@gnu.org] 
Sent: 30 June 2009 23:27
To: Kenichi Handa
Cc: cyd@stupidchicken.com; monnier@iro.umontreal.ca; Marshall, Simon;
3659@emacsbugs.donarmstrong.com
Subject: Re: FW: 23.0.95; crash on set-frame-font

Kenichi Handa wrote:

> I don't know how old Xft 2.1.6 is, but, at least, fontconfig
> 2.2.3 is very old, it's released 5 years ago.
>
> Is it possible to install the latest version of fontconfig,
> and try with it?

I tried, but I ran into dependency problems, and I could not get
everything to work right. If the conclusion is that the system I used
is odd or too old (though it passes all the configure tests), that's
fine by me.

Simon's problem might be different though. Sorry for the diversion!


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

"Marshall, Simon" <Simon.Marshall@misys.com> writes:

> Guys, I waited as I wasn't quite sure if we went thru a diversion on
> this thread or not.
>
> In any case, 23.0.96 still has the crash on set-frame-font introduced
> between 2009-05-21 (ok) and 2009-05-22 (coredump).

Does it crash with the X server from Sun or X.org, rather than the
Hummingbird server?

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

> > Guys, I waited as I wasn't quite sure if we went thru a diversion on
> > this thread or not.
> >
> > In any case, 23.0.96 still has the crash on set-frame-font
introduced
> > between 2009-05-21 (ok) and 2009-05-22 (coredump).

> Does it crash with the X server from Sun or X.org, rather than the
> Hummingbird server?

Unfortunately I don't have access to the solaris box console, assuming
it still has one and if it does assuming it supports X.  It's somewhere
else in London - in an outsourced datacentre.  We only have access via
our PCs.


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

What would you like me to do to help progress this?  Or has it already
been fixed?

If the problem is that no one can reproduce the problem, I could try to
narrow down the 2009-05-21/22 change(s) that caused it...?

-----Original Message-----
From: Marshall, Simon 
Sent: 09 July 2009 15:38
To: 'Chong Yidong'
Cc: Kenichi Handa; monnier@iro.umontreal.ca; Glenn Morris;
3659@emacsbugs.donarmstrong.com
Subject: RE: FW: 23.0.95; crash on set-frame-font

> > Guys, I waited as I wasn't quite sure if we went thru a diversion on
> > this thread or not.
> >
> > In any case, 23.0.96 still has the crash on set-frame-font
introduced
> > between 2009-05-21 (ok) and 2009-05-22 (coredump).

> Does it crash with the X server from Sun or X.org, rather than the
> Hummingbird server?

Unfortunately I don't have access to the solaris box console, assuming
it still has one and if it does assuming it supports X.  It's somewhere
else in London - in an outsourced datacentre.  We only have access via
our PCs.


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

"Marshall, Simon" <Simon.Marshall@misys.com> writes:

> What would you like me to do to help progress this?  Or has it already
> been fixed?
>
> If the problem is that no one can reproduce the problem, I could try to
> narrow down the 2009-05-21/22 change(s) that caused it...?

No one else seems to have a Solaris box with a Hummingbird X server
available.  If you can, please try to narrow down the change.  Thanks.

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

Guys, I can confirm it is this change that started the problem:

2009-05-21  Kenichi Handa  <handa@m17n.org>

	* fontset.c (Qlatin): Don't make it static.

	* xfont.c (xfont_chars_supported, xfont_supported_scripts):
	New functions.
	(xfont_scripts_cache, xfont_scratch_props): New variables.
	(Qlatin, Vscalable_fonts_allowed): Extern it.
	(xfont_list_pattern): Argument changed.  Callers changed.
	Check Vscalable_fonts_allowed.  Check the support of a script.
	(xfont_list): Don't reject a font spec with :script property.
	(xfont_has_char): Fix setting of encoding.
	(syms_of_xfont): Staticpro and initialize xfont_scripts_cache
and
	xfont_scratch_props.

These were made at revisions 1.172 of src/fontset.c and 1.29 of
src/xfont.c.

If you need more info, what would you like me to do?

The core dump is:

risksun2 123> gdb src/emacs
GNU gdb 6.7.1
Copyright (C) 2007 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show
copying"
and "show warranty" for details.
This GDB was configured as "sparc-sun-solaris2.8"...
(gdb) run -Q
Starting program:
/homedev/marshals/ftp/emacs-23.1-pretests/emacs-2009-05-21+2/src/emacs
-Q
warning: Temporarily disabling breakpoints for unloaded shared library
"/usr/lib/ld.so.1"

Program received signal SIGSEGV, Segmentation fault.
0x0009890c in encode_char (charset=0xffbed244, c=192) at charset.c:1804
1804          deunifier = CHARSET_DEUNIFIER (charset);
(gdb) where
#0  0x0009890c in encode_char (charset=0xffbed244, c=192) at
charset.c:1804
#1  0x001bf1b0 in xfont_chars_supported (chars=8343645, xfont=0x866680, 
    encoding=<value optimized out>, repertory=<value optimized out>) at
xfont.c:275
#2  0x001c0a84 in xfont_list_pattern (display=0x8c8000, 
    pattern=0xffbed228 "-*-*-*-*-*-*-*-*-*-*-*-*-*-*", registry=3737601,
script=3737601)
    at xfont.c:365
#3  0x001c0bc4 in xfont_list (frame=<value optimized out>, spec=<value
optimized out>)
    at xfont.c:566
#4  0x0017d47c in font_list_entities (frame=9140228, spec=10350340) at
font.c:2823
#5  0x0017d638 in Flist_fonts (font_spec=10350340, frame=9140228,
num=3737601, prefer=10350340)
    at font.c:4256
#6  0x000ce620 in Fx_list_fonts (pattern=2641595, face=<value optimized
out>, frame=9140228, 
    maximum=3737601, width=<value optimized out>) at xfaces.c:1905
#7  0x00167fac in Feval (form=<value optimized out>) at eval.c:2390
#8  0x00167eb8 in Feval (form=<value optimized out>) at eval.c:2361
#9  0x0016ab80 in FletX (args=2641485) at eval.c:1028
#10 0x001680dc in Feval (form=<value optimized out>) at eval.c:2323
#11 0x00164760 in Fcall_interactively (function=8733249,
record_flag=3737649, keys=3771140)
    at callint.c:364
#12 0x00168b68 in Ffuncall (nargs=4, args=<value optimized out>) at
eval.c:3051
#13 0x00168d1c in call3 (fn=3940449, arg1=8733249, arg2=3737649,
arg3=3737601) at eval.c:2871
#14 0x000fffe4 in Fexecute_extended_command (prefixarg=3737601) at
keyboard.c:10560
#15 0x00168b94 in Ffuncall (nargs=2, args=<value optimized out>) at
eval.c:3045
#16 0x001658a4 in Fcall_interactively (function=3797569,
record_flag=3737601, keys=3611648)
    at callint.c:868
#17 0x00168b68 in Ffuncall (nargs=4, args=<value optimized out>) at
eval.c:3051
#18 0x00168d1c in call3 (fn=3940449, arg1=3797569, arg2=3737601,
arg3=3737601) at eval.c:2871
#19 0x0010cc78 in command_loop_1 () at keyboard.c:1901
#20 0x001672f8 in internal_condition_case (bfun=0x10c858
<command_loop_1>, handlers=3804305, 
    hfun=0x106170 <cmd_error>) at eval.c:1512
#21 0x001055fc in command_loop_2 () at keyboard.c:1359
#22 0x001673ec in internal_catch (tag=3797353, func=0x1055d0
<command_loop_2>, arg=3737601)
    at eval.c:1248
#23 0x00105f80 in command_loop () at keyboard.c:1338
#24 0x00106300 in recursive_edit_1 () at keyboard.c:953
#25 0x00106460 in Frecursive_edit () at keyboard.c:1015
#26 0x000fb818 in main (argc=2, argv=0xffbee39c) at emacs.c:1852
(gdb) 


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <63C2A154B1708946B60726AFDBA00AC006680307@ukmailemea01.misys.global.ad>, "Marshall, Simon" <Simon.Marshall@misys.com> writes:

> Guys, I can confirm it is this change that started the problem:
> 2009-05-21  Kenichi Handa  <handa@m17n.org>

> 	* fontset.c (Qlatin): Don't make it static.

> 	* xfont.c (xfont_chars_supported, xfont_supported_scripts):
> 	New functions.
> 	(xfont_scripts_cache, xfont_scratch_props): New variables.
> 	(Qlatin, Vscalable_fonts_allowed): Extern it.
> 	(xfont_list_pattern): Argument changed.  Callers changed.
> 	Check Vscalable_fonts_allowed.  Check the support of a script.
> 	(xfont_list): Don't reject a font spec with :script property.
> 	(xfont_has_char): Fix setting of encoding.
> 	(syms_of_xfont): Staticpro and initialize xfont_scripts_cache
> and
> 	xfont_scratch_props.

> These were made at revisions 1.172 of src/fontset.c and 1.29 of
> src/xfont.c.

Thank you for tracking down the problem to here.

> If you need more info, what would you like me to do?

> The core dump is:

[...]
> Program received signal SIGSEGV, Segmentation fault.
> 0x0009890c in encode_char (charset=0xffbed244, c=192) at charset.c:1804
> 1804          deunifier = CHARSET_DEUNIFIER (charset);
> (gdb) where
> #0  0x0009890c in encode_char (charset=0xffbed244, c=192) at
> charset.c:1804
> #1  0x001bf1b0 in xfont_chars_supported (chars=8343645, xfont=0x866680, 
>     encoding=<value optimized out>, repertory=<value optimized out>) at
> xfont.c:275

Please show me the output of this gdb command.

(gdb) p *charset
(gdb) xcharset charset->id

Note that you must run gdb in the directory where the
executable Emacs was made.

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <63C2A154B1708946B60726AFDBA00AC00668060C@ukmailemea01.misys.global.ad>, "Marshall, Simon" <Simon.Marshall@misys.com> writes:

> (gdb) p *charset
> $1 = {
>   id = 1, 
>   hash_index = 3678208, 
>   dimension = 1073758208, 
>   code_space = {1294009101, 8830976, 1, 40, -4271696, 1, -4271232, 0, 0,
> 1483848, 805306368, 
>     15794175, 3737601, 3737601, 1, 4}, 

Ummm, the content of this structure is completely broken, or
"charset" points incorrect memory.

> (gdb) xcharset charset->id
> Attempt to extract a component of a value that is not a structure
> pointer.

> Was the last cmd wrong?  Let me know what else I can do...

It's a bug of old .gdbinit, and is now fixed both in the
trunk and EMACS_23_1_RC branch.  It seems that you are not
using the latest emacs code, but as we already found the
culprit change, we can debug using the latest code.  Please
recompile the latest code of the trunk or branch with
  % make CFLAGS=-g clean all

When you get segfault with that code, please show me the
backtrace again, and the results of these commands.

(gdb) up 2   ;; go up to the function xfont_list_pattern
(gdb) pp registry
(gdb) pp script

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

With a CFLAGS=-g version of yesterday's trunk, the crash happens in a
different place (it aborts semi-cleanly).  This is true of the
2009-05-20 version as well:

risksun2 134> gdb ./emacs
GNU gdb 6.7.1
Copyright (C) 2007 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later
<http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show
copying"
and "show warranty" for details.
This GDB was configured as "sparc-sun-solaris2.8"...
SIGINT is used by the debugger.
Are you sure you want to change it? (y or n) [answered Y; input not from
terminal]
DISPLAY = dtp020071.misys.global.ad:0.0
TERM = xterm
Breakpoint 1 at 0x1a331c: file emacs.c, line 433.
Breakpoint 2 at 0x1d6a9c: file sysdep.c, line 1132.
(gdb) r -Q
Starting program:
/homedev/marshals/ftp/emacs-23.1-pretests/emacs-2009-07-23-g/src/emacs
-Q
warning: Temporarily disabling breakpoints for unloaded shared library
"/usr/lib/ld.so.1"

Breakpoint 1, abort () at emacs.c:433
433       kill (getpid (), SIGABRT);
(gdb) where
#0  abort () at emacs.c:433
#1  0x00280904 in internal_condition_case (bfun=0x1a913c
<command_loop_1>, handlers=5385361, 
    hfun=0x1a85c4 <cmd_error>) at eval.c:1487
#2  0x001a8c24 in command_loop_2 () at keyboard.c:1359
#3  0x002801e0 in internal_catch (tag=5378409, func=0x1a8bf8
<command_loop_2>, arg=5314561)
    at eval.c:1248
#4  0x001a8b9c in command_loop () at keyboard.c:1338
#5  0x001a7e54 in recursive_edit_1 () at keyboard.c:953
#6  0x001a8114 in Frecursive_edit () at keyboard.c:1015
#7  0x001a5948 in main (argc=2, argv=0xffbee394) at emacs.c:1854
(gdb) up
#1  0x00280904 in internal_condition_case (bfun=0x1a913c
<command_loop_1>, handlers=5385361, 
    hfun=0x1a85c4 <cmd_error>) at eval.c:1487
1487        abort ();
(gdb) list 
1482
1483      /* Since Fsignal will close off all calls to x_catch_errors,
1484         we will get the wrong results if some are not closed now.
*/
1485    #if HAVE_X_WINDOWS
1486      if (x_catching_errors ())
1487        abort ();
1488    #endif
1489
1490      c.tag = Qnil;
1491      c.val = Qnil;
(gdb) 

With a CFLAGS="-g -O2" version I do not get a crash anymore.

So: -g aborts in 2009-05-22 and 2009-07-23 versions
But: -g -O2 crashes in 2009-05-22 and appears ok in 2009-07-23 versions

What to do?  

-----Original Message-----
From: Kenichi Handa [mailto:handa@m17n.org] 
Sent: 24 July 2009 02:08
To: Marshall, Simon
Cc: cyd@stupidchicken.com; monnier@iro.umontreal.ca; rgm@gnu.org;
3659@emacsbugs.donarmstrong.com
Subject: Re: FW: 23.0.95; crash on set-frame-font

In article
<63C2A154B1708946B60726AFDBA00AC00668060C@ukmailemea01.misys.global.ad>,
"Marshall, Simon" <Simon.Marshall@misys.com> writes:

> (gdb) p *charset
> $1 = {
>   id = 1, 
>   hash_index = 3678208, 
>   dimension = 1073758208, 
>   code_space = {1294009101, 8830976, 1, 40, -4271696, 1, -4271232, 0,
0,
> 1483848, 805306368, 
>     15794175, 3737601, 3737601, 1, 4}, 

Ummm, the content of this structure is completely broken, or
"charset" points incorrect memory.

> (gdb) xcharset charset->id
> Attempt to extract a component of a value that is not a structure
> pointer.

> Was the last cmd wrong?  Let me know what else I can do...

It's a bug of old .gdbinit, and is now fixed both in the
trunk and EMACS_23_1_RC branch.  It seems that you are not
using the latest emacs code, but as we already found the
culprit change, we can debug using the latest code.  Please
recompile the latest code of the trunk or branch with
  % make CFLAGS=-g clean all

When you get segfault with that code, please show me the
backtrace again, and the results of these commands.

(gdb) up 2   ;; go up to the function xfont_list_pattern
(gdb) pp registry
(gdb) pp script

---
Kenichi Handa
handa@m17n.org


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <63C2A154B1708946B60726AFDBA00AC00676D571@ukmailemea01.misys.global.ad>, "Marshall, Simon" <Simon.Marshall@misys.com> writes:

> With a CFLAGS=-g version of yesterday's trunk, the crash happens in a
> different place (it aborts semi-cleanly).  This is true of the
> 2009-05-20 version as well:

Ummm, weird...

> Breakpoint 1, abort () at emacs.c:433
> 433       kill (getpid (), SIGABRT);
> (gdb) where
> #0  abort () at emacs.c:433
> #1  0x00280904 in internal_condition_case (bfun=0x1a913c
> <command_loop_1>, handlers=5385361, 
>     hfun=0x1a85c4 <cmd_error>) at eval.c:1487

The code around there is this:

#if HAVE_X_WINDOWS
  if (x_catching_errors ())
    abort ();
#endif

So, it seems that some X error happened.  Please check the
contents of x_error_message to see what kind of error
happened.

And also please start Emacs, do ESC : (x-synchronize t),
then do M-x set-frame-font.

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

> > Breakpoint 1, abort () at emacs.c:433
> > 433       kill (getpid (), SIGABRT);
> > (gdb) where
> > #0  abort () at emacs.c:433
> > #1  0x00280904 in internal_condition_case (bfun=0x1a913c
> > <command_loop_1>, handlers=5385361, 
> >     hfun=0x1a85c4 <cmd_error>) at eval.c:1487
> 
> The code around there is this:
> 
> #if HAVE_X_WINDOWS
>   if (x_catching_errors ())
>     abort ();
> #endif
> 
> So, it seems that some X error happened.  Please check the
> contents of x_error_message to see what kind of error
> happened.
> 
> And also please start Emacs, do ESC : (x-synchronize t),
> then do M-x set-frame-font.

Ok, I get:

(gdb) p x_error_message 
$1 = (struct x_error_message_stack *) 0xa61100
(gdb) p *x_error_message 
$2 = {
  string =
"\000\233\333\000\000O\305\220\000\004\000\004\000'\000\026\000\000\030\
001\000Q\0301\000Q\030\001\000\000\000\000\000\000\0301\000\275\264\000\
000\000\000\037\000+\000\004\000)\000\026\000\000\201\201\000Q\0301\000Q
\030\001\000\000\000\000\000\000\0301\000\300b\000\000\000\000\037\000T\
000\004\000F\000\026\000\0002\271\000Q\0301\000Q\030\001\000\000\000\000
\000\000\0301\000\300\252\000\000\000\000\037\000\232\000\004\000?\000\0
26\000\000\272Y\000Q\0301\000Q\030\001\000\000\000\000\000\000\0301\000\
302.\000\000\000\000\037\000\331\000\004\0004\000\026\000\0002\271\000Q\
0301\000Q\030\001\000\000\000\000\000\000\0301\000\000\000\000\000\234\2
23\031\000Q\0301\000Q\030\001\000\234\222\271", 
  dpy = 0xa0e000, 
  prev = 0x0
}
(gdb) 

Which doesn't look overly helpful.  

I see it has a leading \000.  Does it mean there were more
x_catch_errors() calls than x_uncatch_errors()?  I tried to put a
breakpoint on x_error_catcher(), but it was never triggered before the
call to abort().

What next to try?  Thanks, Simon.


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <63C2A154B1708946B60726AFDBA00AC00676DA9E@ukmailemea01.misys.global.ad>, "Marshall, Simon" <Simon.Marshall@misys.com> writes:

> Ok, I get:

> (gdb) p x_error_message 
> $1 = (struct x_error_message_stack *) 0xa61100
> (gdb) p *x_error_message 
> $2 = {
>   string =
> "\000\233\333\000\000O\305\220\000\004\000\004\000'\000\026\000\000\030\
> 001\000Q\0301\000Q\030\001\000\000\000\000\000\000\0301\000\275\264\000\
> 000\000\000\037\000+\000\004\000)\000\026\000\000\201\201\000Q\0301\000Q
> \030\001\000\000\000\000\000\000\0301\000\300b\000\000\000\000\037\000T\
> 000\004\000F\000\026\000\0002\271\000Q\0301\000Q\030\001\000\000\000\000
> \000\000\0301\000\300\252\000\000\000\000\037\000\232\000\004\000?\000\0
> 26\000\000\272Y\000Q\0301\000Q\030\001\000\000\000\000\000\000\0301\000\
> 302.\000\000\000\000\037\000\331\000\004\0004\000\026\000\0002\271\000Q\
> 0301\000Q\030\001\000\000\000\000\000\000\0301\000\000\000\000\000\234\2
> 23\031\000Q\0301\000Q\030\001\000\234\222\271", 
>   dpy = 0xa0e000, 
>   prev = 0x0
> }
> (gdb) 

> Which doesn't look overly helpful.  

> I see it has a leading \000.  Does it mean there were more
> x_catch_errors() calls than x_uncatch_errors()?  I tried to put a
> breakpoint on x_error_catcher(), but it was never triggered before the
> call to abort().

> What next to try?

Then, try this:

(1) start emacs under gdb
(2) when the emacs window appears, type C-z (if you are
(using M-x gdb, type C-c C-z) to suspend emacs.
(3) set breakpoint in Fsignal
(4) continue
(5) do M-x set-frame-font
(6) when Emacs stops in Fsignal, do
  (gdb) pp error_symbol
  (gdb) pp data
  (gdb) bt full

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

"Marshall, Simon" <Simon.Marshall@misys.com> writes:

> #3  0x00259ab0 in wrong_type_argument (predicate=5385985, value=5314561)
> at data.c:115
> #4  0x000df1c8 in load_charset_map_from_vector (charset=0x369f30,
> vec=5314561, control_flag=2)
>     at charset.c:608

In that gdb session, please do

f 4
p val
xtype

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

> > #3  0x00259ab0 in wrong_type_argument (predicate=5385985,
value=5314561)
> > at data.c:115
> > #4  0x000df1c8 in load_charset_map_from_vector (charset=0x369f30,
> > vec=5314561, control_flag=2)
> >     at charset.c:608
> 
> In that gdb session, please do
> 
> f 4
> p val
> xtype

I had to restart the session.  However, at that point:

(gdb) f 4
#4  0x000df1c8 in load_charset_map_from_vector (charset=0x369f30,
vec=5314561, control_flag=2)
    at charset.c:608
608               CHECK_NATNUM (val);
(gdb) p val
$13 = 5314561
(gdb) xtype
Lisp_Symbol
(gdb) pp val
nil
(gdb) p vec
$14 = 5314561
(gdb) xtype
Lisp_Symbol
(gdb) pp vec
nil
(gdb) p len
$15 = 810619904
(gdb) p i
$16 = 0
(gdb)


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

What happens if you do the following?

(This is not a final fix, but it might be good enough for 23.1).

*** emacs/src/charset.c.~1.179.2.1.~	2009-07-07 09:40:19.000000000 -0400
--- emacs/src/charset.c	2009-07-26 15:33:32.000000000 -0400
***************
*** 654,660 ****
      map = CHARSET_UNIFY_MAP (charset);
    if (STRINGP (map))
      load_charset_map_from_file (charset, map, control_flag);
!   else
      load_charset_map_from_vector (charset, map, control_flag);
  }
  
--- 654,660 ----
      map = CHARSET_UNIFY_MAP (charset);
    if (STRINGP (map))
      load_charset_map_from_file (charset, map, control_flag);
!   else if (VECTORP (map))
      load_charset_map_from_vector (charset, map, control_flag);
  }
  

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

In article <63C2A154B1708946B60726AFDBA00AC00676DBAF@ukmailemea01.misys.global.ad>, "Marshall, Simon" <Simon.Marshall@misys.com> writes:

> Breakpoint 3, Fsignal (error_symbol=5385409, data=6603597) at
> eval.c:1641
> 1641      register struct handler *allhandlers = handlerlist;
> (gdb) where
> #0  Fsignal (error_symbol=5385409, data=6603597) at eval.c:1641
> #1  0x002813e8 in xsignal (error_symbol=5385409, data=6603597) at
> eval.c:1751
> #2  0x00281480 in xsignal2 (error_symbol=5385409, arg1=5385985,
> arg2=5314561) at eval.c:1775
> #3  0x00259ab0 in wrong_type_argument (predicate=5385985, value=5314561)
> at data.c:115
> #4  0x000df1c8 in load_charset_map_from_vector (charset=0x369f30,
> vec=5314561, control_flag=2)
>     at charset.c:608
> #5  0x000df57c in load_charset (charset=0x369f30, control_flag=2) at
> charset.c:658
> #6  0x000e61b8 in encode_char (charset=0x369f30, c=192) at
> charset.c:1807

Ok, so the Fsignal call is because "*charset" is completely
broken (according to your previous mail).

> #7  0x003337c4 in xfont_chars_supported (chars=9656829, xfont=0xbf4200,
> encoding=0x369f30, 
>     repertory=0x0) at xfont.c:238
> #8  0x0033458c in xfont_supported_scripts (display=0xa0e000,
> fontname=0xc83dfb "--------0------", 
>     props=5454596, encoding=0x369f30) at xfont.c:328

Your previous backtrace didn't show that
xfont_supported_scripts is called.  It may be because the
previous backtrace was made by the emacs compiled with -O2.

Anyway, the above fontname has a strange format.  For such a
fontname, xfont_supported_scripts should not be called from
xfont_list_pattern.

> #9  0x00334e30 in xfont_list_pattern (display=0xa0e000, 
>     pattern=0xffbeca84 "-*-*-*-*-*-*-*-*-*-*-*-*-*-*", registry=5314561,
> script=5314561)
>     at xfont.c:466

Please set the breakpoint at xfont.c:414 with conditional:
  buf[1]=='-' && buf[2]=='-' && buf[3]=='-' && buf[4]=='-'
(I think that is enough for matching with "--------0------")

When emacs stops at that breakpoint, execute the code one
line by one (by "n" command) until the line 466 (i.e. the
call of xfont_supported_scripts) while checking why none of
"if" conditions that leads to "continue" aren't satisfied.

---
Kenichi Handa
handa@m17n.org

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

> What happens if you do the following?
> 
> (This is not a final fix, but it might be good enough for 23.1).
> 
> *** emacs/src/charset.c.~1.179.2.1.~	2009-07-07 09:40:19.000000000
-0400
> --- emacs/src/charset.c	2009-07-26 15:33:32.000000000 -0400
> ***************
> *** 654,660 ****
>       map = CHARSET_UNIFY_MAP (charset);
>     if (STRINGP (map))
>       load_charset_map_from_file (charset, map, control_flag);
> !   else
>       load_charset_map_from_vector (charset, map, control_flag);
>   }
>   
> --- 654,660 ----
>       map = CHARSET_UNIFY_MAP (charset);
>     if (STRINGP (map))
>       load_charset_map_from_file (charset, map, control_flag);
> !   else if (VECTORP (map))
>       load_charset_map_from_vector (charset, map, control_flag);
>   }
  
Unfortunately, it still dies on M-x set-frame-font with a CFLAGS=-g
build, though it no longer seems to go into Fsignal with a CFLAGS=-g -O2
build.

The CFLAGS=-g build is similar to the way it was with 2009-05-21 builds:

Program received signal SIGSEGV, Segmentation fault.
0x000e6380 in encode_char (charset=0x369f68, c=192) at charset.c:1819
1819              code_index = GET_TEMP_CHARSET_WORK_ENCODER (c);
(gdb) where
#0  0x000e6380 in encode_char (charset=0x369f68, c=192) at
charset.c:1819
#1  0x003337f8 in xfont_chars_supported (chars=9656829, xfont=0xc55b80, 
    encoding=0x369f68, repertory=0x0) at xfont.c:238
#2  0x003345c0 in xfont_supported_scripts (display=0xa0e000, 
    fontname=0xc83dfb "--------0------", props=5454596,
encoding=0x369f68)
    at xfont.c:328
#3  0x00334e64 in xfont_list_pattern (display=0xa0e000, 
    pattern=0xffbeca84 "-*-*-*-*-*-*-*-*-*-*-*-*-*-*", registry=5314561,

    script=5314561) at xfont.c:466
#4  0x00335240 in xfont_list (frame=7075332, spec=5454852) at
xfont.c:512
#5  0x002a6fc8 in font_list_entities (frame=7075332, spec=12894212)
    at font.c:2882
#6  0x002ad38c in Flist_fonts (font_spec=12894212, frame=7075332,
num=5314561, 
    prefer=12894212) at font.c:4314
#7  0x00146e8c in Fx_list_fonts (pattern=4221827, face=5314561,
frame=7075332, 
    maximum=5314561, width=5314561) at xfaces.c:1904
#8  0x002830f4 in Feval (form=4221813) at eval.c:2390
#9  0x00282f24 in Feval (form=4221773) at eval.c:2361
#10 0x0027f738 in FletX (args=4221717) at eval.c:1028
#11 0x00282d7c in Feval (form=4221709) at eval.c:2323
#12 0x0027a668 in Fcall_interactively (function=10049921,
record_flag=5314609, 
    keys=5348100) at callint.c:364
#13 0x002849c0 in Ffuncall (nargs=4, args=0xffbed578) at eval.c:3051
#14 0x00284160 in call3 (fn=5521529, arg1=10049921, arg2=5314609,
arg3=5314561)
    at eval.c:2871
#15 0x001c2b00 in Fcommand_execute (cmd=10049921, record_flag=5314609, 
    keys=5314561, special=5314561) at keyboard.c:10524
#16 0x001c3074 in Fexecute_extended_command (prefixarg=5314561)
    at keyboard.c:10637
#17 0x00284930 in Ffuncall (nargs=2, args=0xffbed888) at eval.c:3045
#18 0x0027ce5c in Fcall_interactively (function=5378625,
record_flag=5314561, 
    keys=5348100) at callint.c:868
#19 0x002849c0 in Ffuncall (nargs=4, args=0xffbedb40) at eval.c:3051
#20 0x00284160 in call3 (fn=5521529, arg1=5378625, arg2=5314561,
arg3=5314561)
    at eval.c:2871
#21 0x001c2b00 in Fcommand_execute (cmd=5378625, record_flag=5314561, 
    keys=5314561, special=5314561) at keyboard.c:10524
#22 0x001ab868 in command_loop_1 () at keyboard.c:1903
#23 0x00280aa4 in internal_condition_case (bfun=0x1a9170
<command_loop_1>, 
    handlers=5385361, hfun=0x1a85f8 <cmd_error>) at eval.c:1512
#24 0x001a8c58 in command_loop_2 () at keyboard.c:1359
#25 0x00280214 in internal_catch (tag=5378409, func=0x1a8c2c
<command_loop_2>, 
    arg=5314561) at eval.c:1248
#26 0x001a8bd0 in command_loop () at keyboard.c:1338
#27 0x001a7e88 in recursive_edit_1 () at keyboard.c:953
#28 0x001a8148 in Frecursive_edit () at keyboard.c:1015
#29 0x001a597c in main (argc=2, argv=0xffbee394) at emacs.c:1854

Lisp Backtrace:
"x-list-fonts" (0xffbecf14)
"completing-read" (0xffbed07c)
"let*" (0xffbed234)
"call-interactively" (0xffbed57c)
"execute-extended-command" (0xffbed88c)
"call-interactively" (0xffbedb44)
(gdb) 

I'll try Kenichi Handa's suggestion soon.


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

In frame 9:

#9  0x00334e30 in xfont_list_pattern (display=0xa0e000, pattern=0xffbeca84 "-*-*-*-*-*-*-*-*-*-*-*-*-*-*", registry=5314561,
script=5314561)
    at xfont.c:466

could you do

p names[0]

and show the result?

bug#3659: FW: 23.0.95; crash on set-frame-font

[Chong Yidong]

Also, could you check if this has any effect?  Thanks.

*** emacs/src/xfont.c.~1.30.~	2009-05-21 12:36:30.000000000 -0400
--- emacs/src/xfont.c	2009-07-27 13:01:10.000000000 -0400
***************
*** 411,417 ****
  	    continue;
  	  entity = font_make_entity ();
  	  xfont_decode_coding_xlfd (indices[i], -1, buf);
! 	  font_parse_xlfd (buf, entity);
  	  ASET (entity, FONT_TYPE_INDEX, Qx);
  	  /* Avoid auto-scaled fonts.  */
  	  if (XINT (AREF (entity, FONT_DPI_INDEX)) != 0
--- 411,418 ----
  	    continue;
  	  entity = font_make_entity ();
  	  xfont_decode_coding_xlfd (indices[i], -1, buf);
! 	  if (font_parse_xlfd (buf, entity) < 0)
! 	    continue;
  	  ASET (entity, FONT_TYPE_INDEX, Qx);
  	  /* Avoid auto-scaled fonts.  */
  	  if (XINT (AREF (entity, FONT_DPI_INDEX)) != 0

bug#3659: FW: 23.0.95; crash on set-frame-font

[Kenichi Handa]

Hi, Simon.  Your investigation helps a lot, thank you.  I've
just installed a fix to the branch and the trunk.  Could you
please try the latest code?

---
Kenichi Handa
handa@m17n.org

In article <63C2A154B1708946B60726AFDBA00AC0067E980A@ukmailemea01.misys.global.ad>, "Marshall, Simon" <Simon.Marshall@misys.com> writes:

> > > #7  0x003337c4 in xfont_chars_supported (chars=9656829,
> xfont=0xbf4200,
> > > encoding=0x369f30, 
> > >     repertory=0x0) at xfont.c:238
> > > #8  0x0033458c in xfont_supported_scripts (display=0xa0e000,
> > > fontname=0xc83dfb "--------0------", 
> > >     props=5454596, encoding=0x369f30) at xfont.c:328
> > 
> > Your previous backtrace didn't show that
> > xfont_supported_scripts is called.  It may be because the
> > previous backtrace was made by the emacs compiled with -O2.
> > 
> > Anyway, the above fontname has a strange format.  For such a
> > fontname, xfont_supported_scripts should not be called from
> > xfont_list_pattern.
> > 
> > > #9  0x00334e30 in xfont_list_pattern (display=0xa0e000, 
> > >     pattern=0xffbeca84 "-*-*-*-*-*-*-*-*-*-*-*-*-*-*",
> registry=5314561,
> > > script=5314561)
> > >     at xfont.c:466
> > 
> > Please set the breakpoint at xfont.c:414 with conditional:
> >   buf[1]=='-' && buf[2]=='-' && buf[3]=='-' && buf[4]=='-'
> > (I think that is enough for matching with "--------0------")
> > 
> > When emacs stops at that breakpoint, execute the code one
> > line by one (by "n" command) until the line 466 (i.e. the
> > call of xfont_supported_scripts) while checking why none of
> > "if" conditions that leads to "continue" aren't satisfied.

> Ok, font_parse_xlfd() returns -1 from font.c:1107, though the return
> value is ignored.  After xfont.c:414 we have:

> (gdb) list
> 410		  if (i > 0 && xstrcasecmp (indices[i - 1], indices[i])
> == 0)
> 411		    continue;
> 412		  entity = font_make_entity ();
> 413		  xfont_decode_coding_xlfd (indices[i], -1, buf);
> 414		  font_parse_xlfd (buf, entity);
> 415		  ASET (entity, FONT_TYPE_INDEX, Qx);
> 416		  /* Avoid auto-scaled fonts.  */
> 417		  if (XINT (AREF (entity, FONT_DPI_INDEX)) != 0
> 418		      && XINT (AREF (entity, FONT_AVGWIDTH_INDEX)) == 0)
> 419		    continue;
> (gdb) pp entity
> #<font-entity nil   nil nil nil nil nil nil nil nil nil nil>

> And as expected after the 415 we have:

> (gdb) pp entity
> #<font-entity x   nil nil nil nil nil nil nil nil nil nil>

> The condition at 423 is true:

> (gdb) list
> 418		      && XINT (AREF (entity, FONT_AVGWIDTH_INDEX)) == 0)
> 419		    continue;
> 420		  /* Avoid not-allowed scalable fonts.  */
> 421		  if (NILP (Vscalable_fonts_allowed))
> 422		    {
> 423		      if (XINT (AREF (entity, FONT_SIZE_INDEX)) == 0)
> 424			continue;
> 425		    }
> 426		  else if (CONSP (Vscalable_fonts_allowed))
> 427		    {

> So we make the continue at 424 which jumps us to 443:

> (gdb) list
> 438		      if (! CONSP (tail))
> 439			continue;
> 440		    }
> 441	
> 442		  /* Update encoding and repertory if necessary.  */
> 443		  if (! EQ (registry, AREF (entity,
> FONT_REGISTRY_INDEX)))
> 444		    {
> 445		      registry = AREF (entity, FONT_REGISTRY_INDEX);
> 446		      if (font_registry_charsets (registry, &encoding,
> &repertory) < 0)
> 447			encoding = NULL;

> (gdb) pp registry
> nil
> (gdb) pp entity
> #<font-entity x   nil nil nil nil nil nil nil nil nil nil>

> The condition at 443 is false which jumps us to 449:

> (gdb) list
> 444		    {
> 445		      registry = AREF (entity, FONT_REGISTRY_INDEX);
> 446		      if (font_registry_charsets (registry, &encoding,
> &repertory) < 0)
> 447			encoding = NULL;
> 448		    }
> 449		  if (! encoding)
> 450		    /* Unknown REGISTRY, not supported.  */
> 451		    continue;
> 452		  if (repertory)
> 453		    {

> (gdb) p encoding
> $3 = (struct charset *) 0x369f78
> (gdb) p *encoding
> $4 = {
>   id = 707602944, 
>   hash_index = 0, 
>   dimension = 707592192, 
>   code_space = {0, 757727232, 0, 754974720, 0, -1074790400, 0,
> 1076101120, 0, 0, 0, 704643072, 0, 628305194, 0, 628304429}, 
>   code_space_mask = 0x2a000000 <Address 0x2a000000 out of bounds>, 
>   code_linear_p = 627322154, 
>   iso_chars_96 = 0, 
>   iso_final = 707601764, 
>   iso_revision = 0, 
>   emacs_mule_id = 627322149, 
>   ascii_compatible_p = 1677721600, 
>   supplementary_p = 1828716544, 
>   compact_codes_p = 0, 
>   method = 1660944384, 
>   min_code = 0, 
>   max_code = 1677721600, 
>   char_index_offset = 0, 
>   min_char = 1879048192, 
>   max_char = 0, 
>   invalid_code = 627310592, 
>   fast_map =
> "\000\000\000\000-%s-%s-%s-%s-%s-%s-%s-%s-%s-%s-%s\000\000\000\000\000\0
> 00\000@$\000\000\000\000\000\000roman\000\000\000charcell\000\000\000\00
> 0\000\000\000\000mono\000\000\000\000proportional\000\000\000\000pixelsi
> ze\000\000\000\000\000\000\000Ultra-Light\000\000\000\000\000Light\000\0
> 00\000Semi-Bold\000\000\000\000\000\000\000Bold\000\000\000\000Italic\00
> 0\000Oblique\000weight\000\000sl", 
>   code_offset = 1946157056, 
>   unified_p = 2003395700
> }
> (gdb) p repertory
> $5 = (struct charset *) 0x0

> So both of those conditions are false, so we reach 459:

> (gdb) list
> 454		      if (NILP (script)
> 455			  || xfont_chars_supported (chars, NULL,
> encoding, repertory))
> 456			list = Fcons (entity, list);
> 457		      continue;
> 458		    }
> 459		  if (memcmp (props, &(AREF (entity,
> FONT_FOUNDRY_INDEX)),
> 460			      sizeof (Lisp_Object) * 7)
> 461		      || ! EQ (AREF (entity, FONT_SPACING_INDEX),
> props[7]))
> 462		    {
> 463		      memcpy (props, &(AREF (entity,
> FONT_FOUNDRY_INDEX)),
> (gdb) pp xfont_scratch_props
> [nil nil nil nil nil nil nil nil]
> (gdb) pp props[0]
> nil
> (gdb) pp props[1]
> nil
> (gdb) pp props[2]
> nil
> (gdb) pp props[3]
> nil
> (gdb) pp props[4]
> nil
> (gdb) pp props[5]
> nil
> (gdb) pp props[6]
> nil
> (gdb) pp props[7]
> nil
> (gdb) pp entity
> #<font-entity x   nil nil nil nil nil nil nil nil nil nil>

> The condition at 459 is true so we go to 463.  at the call on 466:

> (gdb) p indices[i]
> $7 = 0xc7adfb "--------0------"
> (gdb) pp xfont_scratch_props
> [  nil nil nil nil nil nil]
> (gdb) p *encoding
> $8 = {
>   id = 707602944, 
>   hash_index = 0, 
>   dimension = 707592192, 
>   code_space = {0, 757727232, 0, 754974720, 0, -1074790400, 0,
> 1076101120, 0, 0, 0, 704643072, 0, 628305194, 0, 628304429}, 
>   code_space_mask = 0x2a000000 <Address 0x2a000000 out of bounds>, 
>   code_linear_p = 627322154, 
>   iso_chars_96 = 0, 
>   iso_final = 707601764, 
>   iso_revision = 0, 
>   emacs_mule_id = 627322149, 
>   ascii_compatible_p = 1677721600, 
>   supplementary_p = 1828716544, 
>   compact_codes_p = 0, 
>   method = 1660944384, 
>   min_code = 0, 
>   max_code = 1677721600, 
>   char_index_offset = 0, 
>   min_char = 1879048192, 
>   max_char = 0, 
>   invalid_code = 627310592, 
>   fast_map =
> "\000\000\000\000-%s-%s-%s-%s-%s-%s-%s-%s-%s-%s-%s\000\000\000\000\000\0
> 00\000@$\000\000\000\000\000\000roman\000\000\000charcell\000\000\000\00
> 0\000\000\000\000mono\000\000\000\000proportional\000\000\000\000pixelsi
> ze\000\000\000\000\000\000\000Ultra-Light\000\000\000\000\000Light\000\0
> 00\000Semi-Bold\000\000\000\000\000\000\000Bold\000\000\000\000Italic\00
> 0\000Oblique\000weight\000\000sl", 
>   code_offset = 1946157056, 
>   unified_p = 2003395700
> }

> Hope this helps!  Let me know what else you would like me to do.

> Simon.


> "Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contrac
 t between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing. 

bug#3659: FW: 23.0.95; crash on set-frame-font

[Marshall, Simon]

> Hi, Simon.  Your investigation helps a lot, thank you.  I've
> just installed a fix to the branch and the trunk.  Could you
> please try the latest code?

I just did a cvs update of the trunk (with no other changes in my
working copy) and it seems fine now.  I'll continue to use this version
until the next pretest.

Many thanks for fixing this!  Simon.


"Misys" is the trade name for Misys plc (registered in England and Wales). Registration Number: 01360027. Registered office: One Kingdom Street, London W2 6BL, United Kingdom. For a list of Misys group operating companies please go to http://www.misys.com/corp/About_Us/misys_operating_companies.html. This email and any attachments have been scanned for known viruses using multiple scanners. This email message is intended for the named recipient only. It may be privileged and/or confidential. If you are not the named recipient of this email please notify us immediately and do not copy it or use it for any purpose, nor disclose its contents to any other person. This email does not constitute the commencement of legal relations between you and Misys plc. Please refer to the executed contract between you and the relevant member of the Misys group for the identity of the contracting party with which you are dealing.