bug#63283: Emacs crash within fontset

bug#63283: Emacs crash within fontset

[Seppo Ronkainen via Bug reports for GNU Emacs, the Swiss army knife of text editors]

[-- Attachment #1.1: Type: text/plain, Size: 728 bytes --]

Honorable GNU developers !

For some reason the new Emacs 29.0.90 did not build at work the other day unless I provided configuration option --without-all (I plan to look into why later but that is not why I am writing this report now).

When I experimented with Emacs (29.0.90 --without-all) I noticed a crash defect that was easy to reproduce. It seemed to have with fonts to do. I have noticed the same problem both at work CentOS 7.9 and at home (currently on a raspberry pi 4) not that it matters but anyhow several machines / architectures. The problem seemed non-present when I built Emacs without configuration option --without-all

I attach details about my findings in an org file, hope it helps

Live long and prosper

[-- Attachment #1.2: Type: text/html, Size: 777 bytes --]

[-- Attachment #2: 20230504172048utc_emacs_crash_defect_report_to_gnu.org --]
[-- Type: application/octet-stream, Size: 2531 bytes --]



* 20230504172048utc Steps to reproduce

create a test file with some unicode character
#+begin_src emacs-lisp
  (with-temp-file (expand-file-name "~/examplefilewithunicodetext")
    (insert-char (char-from-name "LEFT WHITE CORNER BRACKET")))
#+end_src

~/examplefilewithunicodetext
#+begin_export
『
#+end_export


/home/user/.local/emacs-29.0.90_minimal/cons/src/config.h:131
#+begin_src c
/* Define to the options passed to configure. */
#define EMACS_CONFIG_OPTIONS "--prefix=/home/user/.local/emacs-29.0.90_minimal/inst --without-all"
#+end_src


run Emacs without configuration and after one second increase the font size a few times
#+begin_src sh
  PATH="/home/user/.local/emacs-29.0.90_minimal/inst/bin:${PATH}" gdb --ex run --args emacs -Q ~/examplefilewithunicodetext --eval "(run-at-time 1 '() (lambda() (dotimes (i 10) (text-scale-adjust 1))))"
#+end_src

expected behavior: emacs resizes the font

observed behavior: gdb reports emacs defect in fontset.c on line 555
src/fontset.c:549
#+begin_src c
/* 549 */ static Lisp_Object
/* 550 */ fontset_find_font (Lisp_Object fontset, int c, struct face /* *fa */ce,
/* 551 */                    int charset_id, bool fallback)
/* 552 */ {
/* 553 */   Lisp_Object vec, font_group;
/* 554 */   int i, charset_matched = 0, found_index;
/* 555 */   struct frame *f = (FRAMEP (FONTSET_FRAME (fontset))
/* 556 */                      ? XFRAME (FONTSET_FRAME (fontset))
/* 557 */                      : XFRAME (selected_frame));
#+end_src

#+begin_export
0x00000055557864c8 in fontset_find_font (fontset=fontset@entry=0x20, c=c@entry=12302, face=face@entry=0x5556519100, 
    charset_id=charset_id@entry=-1, fallback=fallback@entry=false) at fontset.c:555
555       struct frame *f = (FRAMEP (FONTSET_FRAME (fontset))
#+end_export





when I run a more recent Emacs 30 withoug specifying --without-all then no defect seem present
#+begin_export
--prefix=/home/user/.local/emacs-head-0d2fdf6e36d35e7ab64d8894e8d4c27b0cc06875/inst --with-gif=no --with-included-regex --with-jpeg=no --with-json=yes --with-mailutils=no --with-makeinfo=no --with-modules --with-rsvg=no --with-threads --with-tiff=no --with-xml2=yes --without-gpm --without-gsettings --without-imagemagick --without-javascript --without-sound
#+end_export

#+begin_src sh
  PATH="/home/user/.local/emacs-head-0d2fdf6e36d35e7ab64d8894e8d4c27b0cc06875/inst/bin:${PATH}" gdb --ex run --args emacs -Q ~/examplefilewithunicodetext --eval "(run-at-time 1 '() (lambda() (dotimes (i 10) (text-scale-adjust 1))))"
#+end_src

bug#63283: Emacs crash within fontset

[Ruijie Yu via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Seppo Ronkainen via "Bug reports for GNU Emacs, the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org> writes:

> Honorable GNU developers !
>
> For some reason the new Emacs 29.0.90 did not build at work the other day unless I provided configuration option --without-all (I plan to look
> into why later but that is not why I am writing this report now).
>
> When I experimented with Emacs (29.0.90 --without-all) I noticed a crash defect that was easy to reproduce. It seemed to have with fonts to do.
> I have noticed the same problem both at work CentOS 7.9 and at home (currently on a raspberry pi 4) not that it matters but anyhow several
> machines / architectures. The problem seemed non-present when I built Emacs without configuration option --without-all
>
> I attach details about my findings in an org file, hope it helps

Not quite sure if this is a bug per-se, since there are some of the
default configurations concerning fonts, which you have deliberately
disabled.

This is the value of `system-configuration-features' on my build:

--8<---------------cut here---------------start------------->8---
"ACL CAIRO DBUS FREETYPE GIF GLIB GMP GNUTLS GPM GSETTINGS HARFBUZZ JPEG
JSON LCMS2 LIBSYSTEMD LIBXML2 MODULES NATIVE_COMP NOTIFY INOTIFY PDUMPER
PGTK PNG RSVG SECCOMP SOUND SQLITE3 THREADS TIFF TOOLKIT_SCROLL_BARS
TREE_SITTER WEBP XIM GTK3 ZLIB"
--8<---------------cut here---------------end--------------->8---

Someone more qualified may be able to answer which feature(s) exactly
are needed for having fonts work correctly.

What is the value of this variable on a `--without-all' build?  And does
the issue occur on both GUI ("emacs -Q") session and TUI ("emacs -Q
-nw") session?

-- 
Best,


RY

[Please note that this mail might go to spam due to some
misconfiguration in my mail server -- still investigating.]

bug#63283: Emacs crash within fontset

[Eli Zaretskii]

> Date: Thu, 04 May 2023 19:04:48 +0000
> From:  Seppo Ronkainen via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
> 
> run Emacs without configuration and after one second increase the font size a few times
> #+begin_src sh
>   PATH="/home/user/.local/emacs-29.0.90_minimal/inst/bin:${PATH}" gdb --ex run --args emacs -Q ~/examplefilewithunicodetext --eval "(run-at-time 1 '() (lambda() (dotimes (i 10) (text-scale-adjust 1))))"
> #+end_src
> 
> expected behavior: emacs resizes the font
> 
> observed behavior: gdb reports emacs defect in fontset.c on line 555
> src/fontset.c:549
> #+begin_src c
> /* 549 */ static Lisp_Object
> /* 550 */ fontset_find_font (Lisp_Object fontset, int c, struct face /* *fa */ce,
> /* 551 */                    int charset_id, bool fallback)
> /* 552 */ {
> /* 553 */   Lisp_Object vec, font_group;
> /* 554 */   int i, charset_matched = 0, found_index;
> /* 555 */   struct frame *f = (FRAMEP (FONTSET_FRAME (fontset))
> /* 556 */                      ? XFRAME (FONTSET_FRAME (fontset))
> /* 557 */                      : XFRAME (selected_frame));
> #+end_src
> 
> #+begin_export
> 0x00000055557864c8 in fontset_find_font (fontset=fontset@entry=0x20, c=c@entry=12302, face=face@entry=0x5556519100, 
>     charset_id=charset_id@entry=-1, fallback=fallback@entry=false) at fontset.c:555
> 555       struct frame *f = (FRAMEP (FONTSET_FRAME (fontset))
> #+end_export

What do you mean by "gdb reports emacs defect"?  Please show the
complete GDB session, including everything that GDB shows on screen.

Also please send the data about your Emacs and system configuration,
as collected by "M-x report-emacs-bug RET".

Thanks.

bug#63283: Emacs crash within fontset

[Eli Zaretskii]

[Please use Reply All to reply, so that the bug address is on the CC
list, and this discussion is recorded by our bug tracker.]

> Date: Fri, 05 May 2023 21:19:18 +0000
> From: Seppo Ronkainen <sodr80@protonmail.com>
> 
> I was not aware about report-emacs-bug   thanks for the guidance
> 
> Following instructions from (share/emacs/29.0.90/etc/DEBUG:19) I rebuilt with options
> 
>   ./configure --enable-checking='yes,glyphs' --enable-check-lisp-object-type \
>     CFLAGS='-O0 -g3'
> 
> 
> 
> TL;DR
> Now it looks as if (of course) Emacs never crashed, but it seems as if insufficient font data caused an abort is that right? Hope this helps
> 
> fontset.c:970: Emacs fatal error: assertion failed: fontset_id_valid_p (face->fontset)
> Fatal error 6: Aborted

Thanks, but this is not enough info to find the root cause of the
problem.  Please run Emacs under GDB, like this:

  $ gdb /path/to/src/emacs
  ...
  (gdb) run

Then, when Emacs crashes or aborts, GDB will kick in and show its
"(gdb)" prompt.  Please type at the GDB prompt:

  (gdb) thread apply all bt

and post here everything that GDB prints in response.

Also, please show what this displays inside Emacs built --without-all:

  M-: system-configuration-features RET

bug#63283: Emacs crash within fontset

[Seppo Ronkainen via Bug reports for GNU Emacs, the Swiss army knife of text editors]

sure, here are captures of gdb and system-configuration-features from two build variants of 29.0.90 and one 26.3





/home/user/.local/emacs-29.0.90_minimal/inst/bin/emacs -Q ~/examplefilewithunicodetext

(describe-variable 'system-configuration-features)



system-configuration-features is a variable defined in ‘src/emacs.c’.

Its value is "GLIB GMP JSON PDUMPER SECCOMP X11 XIM GTK3"

String listing some of the main features this Emacs was compiled with.
An element of the form "FOO" generally means that HAVE_FOO was
defined during the build.

This is mainly intended for diagnostic purposes in bug reports.
Don’t rely on it for testing whether a feature you want to use is available.

  Probably introduced at or before Emacs version 25.1.

[back]















/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs -Q ~/examplefilewithunicodetext

(describe-variable 'system-configuration-features)



system-configuration-features is a variable defined in ‘C source code’.

Its value is "GLIB GMP JSON PDUMPER SECCOMP X11 XIM GTK3"

String listing some of the main features this Emacs was compiled with.
An element of the form "FOO" generally means that HAVE_FOO was
defined during the build.

This is mainly intended for diagnostic purposes in bug reports.
Don’t rely on it for testing whether a feature you want to use is available.

  Probably introduced at or before Emacs version 25.1.




















/home/user/.local/emacs-29.0.90_minimal/cons/src/config.h:128
/* Summary of some of the main features enabled by configure. */
#define EMACS_CONFIG_FEATURES "GLIB GMP JSON PDUMPER SECCOMP X11 XIM GTK3"

/* Define to the options passed to configure. */
#define EMACS_CONFIG_OPTIONS "--prefix=/home/user/.local/emacs-29.0.90_minimal/inst --without-all"


> gdb --args /home/user/.local/emacs-29.0.90_minimal/inst/bin/emacs -Q ~/examplefilewithunicodetext
GNU gdb (Debian 10.1-1.7) 10.1.90.20210103-git
Copyright (C) 2021 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 "aarch64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /home/user/.local/emacs-29.0.90_minimal/inst/bin/emacs...
(gdb) run
Starting program: /home/user/.local/emacs-29.0.90_minimal/inst/bin/emacs -Q /home/user/examplefilewithunicodetext
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
[New Thread 0x7ff4a6fe80 (LWP 44415)]
[Detaching after vfork from child process 44416]
[Detaching after vfork from child process 44417]
[Detaching after vfork from child process 44418]

Thread 1 "emacs" received signal SIGSEGV, Segmentation fault.
0x00000055557864c8 in fontset_find_font (fontset=fontset@entry=0x30, c=c@entry=12302, face=face@entry=0x555664ee90, 
    charset_id=charset_id@entry=-1, fallback=fallback@entry=false) at fontset.c:555
555       struct frame *f = (FRAMEP (FONTSET_FRAME (fontset))
(gdb) thread apply all bt

Thread 2 (Thread 0x7ff4a6fe80 (LWP 44415) "gmain"):
#0  0x0000007ff6fa6ef4 in __GI___poll (fds=0x5555dd25c0, nfds=1, timeout=<optimized out>) at ../sysdeps/unix/sysv/linux/poll.c:41
#1  0x0000007ff74b7df8 in  () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#2  0x0000007ff74b7f24 in g_main_context_iteration () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#3  0x0000007ff74b7f7c in  () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#4  0x0000007ff74e18f4 in  () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#5  0x0000007ff71d4648 in start_thread (arg=0x7ff4a6f780) at pthread_create.c:477
#6  0x0000007ff6fb0c1c in thread_start () at ../sysdeps/unix/sysv/linux/aarch64/clone.S:78

Thread 1 (Thread 0x7ff5d6f700 (LWP 44408) "emacs"):
#0  0x00000055557864c8 in fontset_find_font (fontset=fontset@entry=0x30, c=c@entry=12302, face=face@entry=0x555664ee90, charset_id=charset_id@entry=-1, fallback=fallback@entry=false) at fontset.c:555
#1  0x0000005555787240 in fontset_font (fontset=fontset@entry=0x30, c=c@entry=12302, face=face@entry=0x555664ee90, id=-1) at fontset.c:766
#2  0x0000005555787938 in face_for_char (f=0x5556159e50, face=0x555664ee90, c=<optimized out>, pos=<optimized out>, object=<optimized out>) at fontset.c:995                                                                                                                                     
#3  0x00000055555c45d8 in FACE_FOR_CHAR (object=<optimized out>, pos=<optimized out>, character=<optimized out>, face=<optimized out>, f=<optimized out>) at dispextern.h:1911           
#4  get_next_display_element (it=it@entry=0x7fffff8978) at xdisp.c:8184
#5  0x00000055555c9988 in display_line (it=it@entry=0x7fffff8978, cursor_vpos=cursor_vpos@entry=0) at xdisp.c:24708
#6  0x00000055555cda60 in try_window (window=window@entry=0x555615a09d, pos=..., flags=flags@entry=1) at xdisp.c:20577
#7  0x00000055555e8cd0 in redisplay_window (window=<optimized out>, just_this_one_p=just_this_one_p@entry=false) at xdisp.c:19961
#8  0x00000055555eac14 in redisplay_window_0 (window=window@entry=0x555615a09d) at xdisp.c:17446
#9  0x0000005555713234 in internal_condition_case_1 (bfun=bfun@entry=0x55555eabe0 <redisplay_window_0>, arg=arg@entry=0x555615a09d, handlers=<optimized out>, hfun=hfun@entry=0x55555aa300 <redisplay_window_error>) at eval.c:1498
#10 0x00000055555a7348 in redisplay_windows (window=0x555615a09d) at xdisp.c:17416
#11 0x00000055555d6720 in redisplay_internal () at xdisp.c:16866
#12 0x00000055555d7a44 in redisplay () at xdisp.c:16049
#13 0x000000555569b6b0 in read_char (commandflag=commandflag@entry=1, map=map@entry=0x55560ddf03, prev_event=0x0, used_mouse_menu=used_mouse_menu@entry=0x7fffffe013, end_time=end_time@entry=0x0) at keyboard.c:2634
#14 0x000000555569d780 in read_key_sequence (keybuf=keybuf@entry=0x7fffffe140, prompt=prompt@entry=0x0, dont_downcase_last=dont_downcase_last@entry=false, can_return_switch_frame=can_return_switch_frame@entry=true, fix_current_buffer=fix_current_buffer@entry=true, prevent_redisplay=prevent_redisplay@entry=false) at keyboard.c:10081
#15 0x000000555569f19c in command_loop_1 () at lisp.h:1165
#16 0x00000055557131a0 in internal_condition_case (bfun=bfun@entry=0x555569f004 <command_loop_1>, handlers=handlers@entry=0x90, hfun=hfun@entry=0x5555693030 <cmd_error>) at eval.c:1474           
#17 0x000000555568c378 in command_loop_2 (handlers=handlers@entry=0x90) at keyboard.c:1131
#18 0x00000055557130f4 in internal_catch (tag=tag@entry=0xf390, func=func@entry=0x555568c344 <command_loop_2>, arg=arg@entry=0x90) at eval.c:1197
#19 0x000000555568c310 in command_loop () at lisp.h:1165
#20 0x0000005555692b44 in recursive_edit_1 () at keyboard.c:718
#21 0x0000005555692f54 in Frecursive_edit () at keyboard.c:801
#22 0x000000555558cecc in main (argc=3, argv=0x7fffffe638) at emacs.c:2529
(gdb) q
A debugging session is active.

        Inferior 1 [process 44408] will be killed.

Quit anyway? (y or n) y
> exit
exit

Process terminal finished














/home/user/.local/emacs-29.0.90_debug/cons/src/config.h:128
/* Summary of some of the main features enabled by configure. */
#define EMACS_CONFIG_FEATURES "GLIB GMP JSON PDUMPER SECCOMP X11 XIM GTK3"

/* Define to the options passed to configure. */
#define EMACS_CONFIG_OPTIONS "--prefix=/home/user/.local/emacs-29.0.90_debug/inst --without-all --enable-checking=yes,glyphs --enable-check-lisp-object-type 'CFLAGS=-O0 -g3'"

> gdb --args /home/user/.local/emacs-29.0.90_debug/inst/bin/emacs -Q ~/examplefilewithunicodetext
GNU gdb (Debian 10.1-1.7) 10.1.90.20210103-git
Copyright (C) 2021 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 "aarch64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /home/user/.local/emacs-29.0.90_debug/inst/bin/emacs...
(gdb) run
Starting program: /home/user/.local/emacs-29.0.90_debug/inst/bin/emacs -Q /home/user/examplefilewithunicodetext
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
[New Thread 0x7ff4a6fe80 (LWP 44369)]
[Detaching after vfork from child process 44370]
[Detaching after vfork from child process 44371]
[Detaching after vfork from child process 44372]

fontset.c:970: Emacs fatal error: assertion failed: fontset_id_valid_p (face->fontset)
Fatal error 6: Aborted
Backtrace:
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x246468)[0x5555796468]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x205cc0)[0x5555755cc0]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x2cba04)[0x555581ba04]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x3d33c4)[0x55559233c4]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x6d308)[0x55555bd308]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x8d054)[0x55555dd054]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xc1a10)[0x5555611a10]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xb3424)[0x5555603424]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xb059c)[0x555560059c]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xa7740)[0x55555f7740]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x30be18)[0x555585be18]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xa73b0)[0x55555f73b0]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xa5b74)[0x55555f5b74]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0xa314c)[0x55555f314c]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x213e58)[0x5555763e58]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x229ba4)[0x5555779ba4]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x20f678)[0x555575f678]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x30bd14)[0x555585bd14]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x20ecfc)[0x555575ecfc]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x30ae0c)[0x555585ae0c]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x20ec7c)[0x555575ec7c]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x20ddbc)[0x555575ddbc]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x20e028)[0x555575e028]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x208d90)[0x5555758d90]
/lib/aarch64-linux-gnu/libc.so.6(__libc_start_main+0xe8)[0x7ff6efee18]
/home/user/.local/emacs-29.0.90_debug/inst/bin/emacs(+0x3b368)[0x555558b368]

Thread 1 "emacs" received signal SIGABRT, Aborted.
raise (sig=6) at ../sysdeps/unix/sysv/linux/raise.c:50
50      ../sysdeps/unix/sysv/linux/raise.c: No such file or directory.
(gdb) thread apply all bt

Thread 2 (Thread 0x7ff4a6fe80 (LWP 44369) "gmain"):
#0  0x0000007ff6fa6ef4 in __GI___poll (fds=0x555611e270, nfds=1, timeout=<optimized out>) at ../sysdeps/unix/sysv/linux/poll.c:41
#1  0x0000007ff74b7df8 in  () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#2  0x0000007ff74b7f24 in g_main_context_iteration () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#3  0x0000007ff74b7f7c in  () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#4  0x0000007ff74e18f4 in  () at /lib/aarch64-linux-gnu/libglib-2.0.so.0
#5  0x0000007ff71d4648 in start_thread (arg=0x7ff4a6f780) at pthread_create.c:477
#6  0x0000007ff6fb0c1c in thread_start () at ../sysdeps/unix/sysv/linux/aarch64/clone.S:78

Thread 1 (Thread 0x7ff5d6f700 (LWP 44362) "emacs"):
#0  raise (sig=6) at ../sysdeps/unix/sysv/linux/raise.c:50
#1  0x0000005555755cf0 in terminate_due_to_signal (sig=6, backtrace_limit=2147483647) at emacs.c:464
#2  0x000000555581ba04 in die (msg=0x55559f6900 "fontset_id_valid_p (face->fontset)", file=0x55559f6608 "fontset.c", line=970) at alloc.c:7707
#3  0x00000055559233c4 in face_for_char (f=0x55564e70f0, face=0x5556ac95e0, c=12302, pos=1, object=...) at fontset.c:970
#4  0x00000055555bd308 in FACE_FOR_CHAR (f=0x55564e70f0, face=0x5556ac95e0, character=12302, pos=1, object=...) at dispextern.h:1911
#5  0x00000055555dd054 in get_next_display_element (it=0x7fffff8818) at xdisp.c:8184
#6  0x0000005555611a10 in display_line (it=0x7fffff8818, cursor_vpos=0) at xdisp.c:24708
#7  0x0000005555603424 in try_window (window=..., pos=..., flags=1) at xdisp.c:20577
#8  0x000000555560059c in redisplay_window (window=..., just_this_one_p=false) at xdisp.c:19961
#9  0x00000055555f7740 in redisplay_window_0 (window=...) at xdisp.c:17446
#10 0x000000555585be18 in internal_condition_case_1 (bfun=0x55555f76fc <redisplay_window_0>, arg=..., handlers=..., hfun=0x55555f73e8 <redisplay_window_error>) at eval.c:1498                       
#11 0x00000055555f73b0 in redisplay_windows (window=...) at xdisp.c:17416
#12 0x00000055555f5b74 in redisplay_internal () at xdisp.c:16866
#13 0x00000055555f314c in redisplay () at xdisp.c:16049
#14 0x0000005555763e58 in read_char (commandflag=1, map=..., prev_event=..., used_mouse_menu=0x7fffffdf17, end_time=0x0) at keyboard.c:2634
#15 0x0000005555779ba4 in read_key_sequence (keybuf=0x7fffffe098, prompt=..., dont_downcase_last=false, can_return_switch_frame=true, fix_current_buffer=true, prevent_redisplay=false) at keyboard.c:10081
#16 0x000000555575f678 in command_loop_1 () at keyboard.c:1382
#17 0x000000555585bd14 in internal_condition_case (bfun=0x555575f1dc <command_loop_1>, handlers=..., hfun=0x555575e464 <cmd_error>) at eval.c:1474
#18 0x000000555575ecfc in command_loop_2 (handlers=...) at keyboard.c:1131
#19 0x000000555585ae0c in internal_catch (tag=..., func=0x555575ecd4 <command_loop_2>, arg=...) at eval.c:1197
#20 0x000000555575ec7c in command_loop () at keyboard.c:1109
#21 0x000000555575ddbc in recursive_edit_1 () at keyboard.c:718
#22 0x000000555575e028 in Frecursive_edit () at keyboard.c:801
#23 0x0000005555758d90 in main (argc=3, argv=0x7fffffe638) at emacs.c:2529
(gdb) q
A debugging session is active.

        Inferior 1 [process 44362] will be killed.

Quit anyway? (y or n) y
> exit
exit

Process terminal finished



















attempts to reproduce the abort within Emacs 26.3 built with the feature FREETYPE (was reported in the output):




/home/user/.local/emacs-26.3/inst/bin/emacs-26.3 -Q ~/examplefilewithunicodetext

(describe-variable 'system-configuration-features)


system-configuration-features is a variable defined in ‘C source code’.
Its value is
"XPM PNG SOUND DBUS GSETTINGS GLIB NOTIFY LIBSELINUX GNUTLS LIBXML2 FREETYPE XFT ZLIB TOOLKIT_SCROLL_BARS GTK3 X11 XDBE XIM MODULES THREADS"

Documentation:
String listing some of the main features this Emacs was compiled with.
An element of the form "FOO" generally means that HAVE_FOO was
defined during the build.

This is mainly intended for diagnostic purposes in bug reports.
Don’t rely on it for testing whether a feature you want to use is available.




/home/user/.local/emacs-26.3/cons/src/config.h:117
/* Define to the canonical Emacs configuration name. */
#define EMACS_CONFIGURATION "aarch64-unknown-linux-gnu"

/* Summary of some of the main features enabled by configure. */
#define EMACS_CONFIG_FEATURES "XPM PNG SOUND DBUS GSETTINGS GLIB NOTIFY LIBSELINUX GNUTLS LIBXML2 FREETYPE XFT ZLIB TOOLKIT_SCROLL_BARS GTK3 X11 XDBE XIM MODULES THREADS"

/* Define to the options passed to configure. */
#define EMACS_CONFIG_OPTIONS "--prefix=/home/user/.local/emacs-26.3/inst --with-gif=no --with-included-regex --with-jpeg=no --with-json=yes --with-mailutils=yes --with-makeinfo=no --with-modules --with-rsvg=no --with-threads --with-tiff=no --with-xml2=yes"




I rescale the buffer back and forth several times with


C-x C-0, C-x C-=, C-x C--, C-x C-+.

(text-scale-adjust INC)


to both limits which where indicated in the mode line

(Fundamental -15)
(Fundamental +29)





> gdb --args /home/user/.local/emacs-26.3/inst/bin/emacs-26.3 -Q ~/examplefilewithunicodetext
GNU gdb (Debian 10.1-1.7) 10.1.90.20210103-git
Copyright (C) 2021 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 "aarch64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /home/user/.local/emacs-26.3/inst/bin/emacs-26.3...
(gdb) run
Starting program: /home/user/.local/emacs-26.3/inst/bin/emacs-26.3 -Q /home/user/examplefilewithunicodetext
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
[New Thread 0x7ff3354c40 (LWP 45101)]
[Detaching after vfork from child process 45102]
[Detaching after vfork from child process 45103]
[Detaching after vfork from child process 45104]
[New Thread 0x7ff2890c40 (LWP 45105)]
[New Thread 0x7ff1f00c40 (LWP 45106)]
[Thread 0x7ff1f00c40 (LWP 45106) exited]
[Thread 0x7ff3354c40 (LWP 45101) exited]
[Thread 0x7ff3681040 (LWP 45094) exited]
[Inferior 1 (process 45094) exited normally]
(gdb) q
> exit
exit

Process terminal finished









> uname -a
Linux raspberrypi 6.1.21-v8+ #1642 SMP PREEMPT Mon Apr  3 17:24:16 BST 2023 aarch64 GNU/Linux

% sha512sum emacs-29.0.90.tar.gz
5cbd47142ccd845b8da33f5c2cf62088b5076182a91e4cda7de6c5eef980743221705da8e6fa825bc0f6e7dcc0625b78382bdd12e319e0ce76ec3d4690557696  emacs-29.0.90.tar.gz










Live long and prosper

Sent with Proton Mail secure email.

------- Original Message -------
On Saturday, May 6th, 2023 at 10:19 AM, Eli Zaretskii <eliz@gnu.org> wrote:


> [Please use Reply All to reply, so that the bug address is on the CC
> list, and this discussion is recorded by our bug tracker.]
> 
> > Date: Fri, 05 May 2023 21:19:18 +0000
> > From: Seppo Ronkainen sodr80@protonmail.com
> > 
> > I was not aware about report-emacs-bug thanks for the guidance
> > 
> > Following instructions from (share/emacs/29.0.90/etc/DEBUG:19) I rebuilt with options
> > 
> > ./configure --enable-checking='yes,glyphs' --enable-check-lisp-object-type \
> > CFLAGS='-O0 -g3'
> > 
> > TL;DR
> > Now it looks as if (of course) Emacs never crashed, but it seems as if insufficient font data caused an abort is that right? Hope this helps
> > 
> > fontset.c:970: Emacs fatal error: assertion failed: fontset_id_valid_p (face->fontset)
> > Fatal error 6: Aborted
> 
> 
> Thanks, but this is not enough info to find the root cause of the
> problem. Please run Emacs under GDB, like this:
> 
> $ gdb /path/to/src/emacs
> ...
> (gdb) run
> 
> Then, when Emacs crashes or aborts, GDB will kick in and show its
> "(gdb)" prompt. Please type at the GDB prompt:
> 
> (gdb) thread apply all bt
> 
> and post here everything that GDB prints in response.
> 
> Also, please show what this displays inside Emacs built --without-all:
> 
> M-: system-configuration-features RET

bug#63283: Emacs crash within fontset

[Eli Zaretskii]

> Date: Sat, 06 May 2023 19:12:53 +0000
> From: Seppo Ronkainen <sodr80@protonmail.com>
> Cc: 63283@debbugs.gnu.org
> 
> sure, here are captures of gdb and system-configuration-features from two build variants of 29.0.90 and one 26.3

Thanks.  Please repeat this session again, and this time, when Emacs
aborts and GDB shows its prompt, do this:

  (gdb) thread 1
  (gdb) pp fontset

If GDB says that it doesn't know the command "pp", type

  (gdb) source /path/to/emacs/src/.gdbinit
  (gdb) pp fontset

and post everything GDB prints after that.

And one more data point: in the same session do this:

  (gdb) frame 2
  (gdb) print font_driver_list
  (gdb) print f->font_driver_list

and show the values GDB prints.

bug#63283: Emacs crash within fontset

[Seppo Ronkainen via Bug reports for GNU Emacs, the Swiss army knife of text editors]

more data



% gdb --args /home/user/.local/emacs-29.0.90_debug/inst/bin/emacs -Q ~/examplefilewithunicodetext
GNU gdb (Debian 10.1-1.7) 10.1.90.20210103-git
Copyright (C) 2021 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 "aarch64-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<https://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
    <http://www.gnu.org/software/gdb/documentation/>.

For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from /home/user/.local/emacs-29.0.90_debug/inst/bin/emacs...
(gdb) source /home/user/.local/emacs-29.0.90_debug/cons/src/.gdbinit
Warning: /home/user/../lwlib: No such file or directory.
SIGINT is used by the debugger.
Are you sure you want to change it? (y or n) [answered Y; input not from terminal]
DISPLAY = :0
TERM = xterm
Breakpoint 1 at 0x205be8: file emacs.c, line 427.
Breakpoint 2 at 0x1c4388: file xterm.c, line 26126.
(gdb) run
Starting program: /home/user/.local/emacs-29.0.90_debug/inst/bin/emacs -Q /home/user/examplefilewithunicodetext
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
[New Thread 0x7ff4a8ee80 (LWP 52196)]

(process:52189): Gtk-WARNING **: 21:29:18.069: Locale not supported by C library.
        Using the fallback 'C' locale.
[Detaching after vfork from child process 52197]
[Detaching after vfork from child process 52198]
[Detaching after vfork from child process 52199]

fontset.c:970: Emacs fatal error: assertion failed: fontset_id_valid_p (face->fontset)

Thread 1 "emacs" hit Breakpoint 1, terminate_due_to_signal (sig=6, backtrace_limit=2147483647) at emacs.c:427
427       signal (sig, SIG_DFL);
(gdb) thread 1
[Switching to thread 1 (Thread 0x7ff5d6f700 (LWP 52189))]
#0  terminate_due_to_signal (sig=6, backtrace_limit=2147483647) at emacs.c:427
427       signal (sig, SIG_DFL);
(gdb) bt 5
#0  terminate_due_to_signal (sig=6, backtrace_limit=2147483647) at emacs.c:427
#1  0x000000555581ba04 in die (msg=0x55559f6900 "fontset_id_valid_p (face->fontset)", file=0x55559f6608 "fontset.c", line=970) at alloc.c:7707
#2  0x00000055559233c4 in face_for_char (f=0x55564c8470, face=0x5556292b90, c=12302, pos=1, object=XIL(0)) at fontset.c:970
#3  0x00000055555bd308 in FACE_FOR_CHAR (f=0x55564c8470, face=0x5556292b90, character=12302, pos=1, object=XIL(0)) at dispextern.h:1911
#4  0x00000055555dd054 in get_next_display_element (it=0x7fffff8a88) at xdisp.c:8184
(More stack frames follow...)

Lisp Backtrace:
"redisplay_internal (C function)" (0x0)
(gdb) frame 2
#2  0x00000055559233c4 in face_for_char (f=0x55564c8470, face=0x5556292b90, c=12302, pos=1, object=XIL(0)) at fontset.c:970
970       eassert (fontset_id_valid_p (face->fontset));
(gdb) info locals
fontset = XIL(0)
rfont_def = XIL(0x555605b2b8)
charset = XIL(0x300effff7380)
face_id = 85
id = 1432078920
(gdb) print font_driver_list
$1 = (struct font_driver_list *) 0x55560d77e0
(gdb) print f->font_driver_list
$2 = (struct font_driver_list *) 0x5556274ae0
(gdb) frame 3
#3  0x00000055555bd308 in FACE_FOR_CHAR (f=0x55564c8470, face=0x5556292b90, character=12302, pos=1, object=XIL(0)) at dispextern.h:1911
(gdb) p face->font
$5 = (struct font *) 0x0
(gdb) p face->fontset
$6 = -1
(gdb) q



if fontset carried -1 .. which in turn



/home/user/.local/emacs-29.0.90_debug/cons/src/fontset.c:970
  eassert (fontset_id_valid_p (face->fontset));
  fontset = FONTSET_FROM_ID (face->fontset);
  eassert (!BASE_FONTSET_P (fontset));




/home/user/.local/emacs-29.0.90_debug/cons/src/fontset.c:176
static bool
fontset_id_valid_p (int id)
{
  return (id >= 0 && id < ASIZE (Vfontset_table) - 1);
}



relevant commit ?
3c0dda2663e38635163f0fd6c19748c6eba1c3c8
Avoid asserting for fontset validity before it is used







Live long and prosper

Sent with Proton Mail secure email.

------- Original Message -------
On Saturday, May 6th, 2023 at 7:32 PM, Eli Zaretskii <eliz@gnu.org> wrote:


> > Date: Sat, 06 May 2023 19:12:53 +0000
> > From: Seppo Ronkainen sodr80@protonmail.com
> > Cc: 63283@debbugs.gnu.org
> > 
> > sure, here are captures of gdb and system-configuration-features from two build variants of 29.0.90 and one 26.3
> 
> 
> Thanks. Please repeat this session again, and this time, when Emacs
> aborts and GDB shows its prompt, do this:
> 
> (gdb) thread 1
> (gdb) pp fontset
> 
> If GDB says that it doesn't know the command "pp", type
> 
> (gdb) source /path/to/emacs/src/.gdbinit
> (gdb) pp fontset
> 
> and post everything GDB prints after that.
> 
> And one more data point: in the same session do this:
> 
> (gdb) frame 2
> (gdb) print font_driver_list
> (gdb) print f->font_driver_list
> 
> 
> and show the values GDB prints.

bug#63283: Emacs crash within fontset

[Eli Zaretskii]

> Date: Sat, 06 May 2023 22:36:35 +0000
> From: Seppo Ronkainen <sodr80@protonmail.com>
> Cc: 63283@debbugs.gnu.org
> 
> (gdb) frame 2
> #2  0x00000055559233c4 in face_for_char (f=0x55564c8470, face=0x5556292b90, c=12302, pos=1, object=XIL(0)) at fontset.c:970
> 970       eassert (fontset_id_valid_p (face->fontset));
> (gdb) info locals
> fontset = XIL(0)
> rfont_def = XIL(0x555605b2b8)
> charset = XIL(0x300effff7380)
> face_id = 85
> id = 1432078920
> (gdb) print font_driver_list
> $1 = (struct font_driver_list *) 0x55560d77e0
> (gdb) print f->font_driver_list
> $2 = (struct font_driver_list *) 0x5556274ae0
> (gdb) frame 3
> #3  0x00000055555bd308 in FACE_FOR_CHAR (f=0x55564c8470, face=0x5556292b90, character=12302, pos=1, object=XIL(0)) at dispextern.h:1911
> (gdb) p face->font
> $5 = (struct font *) 0x0
> (gdb) p face->fontset
> $6 = -1
> (gdb) q

Thanks.  If you apply the patch below, does it avoid the abort?

> relevant commit ?
> 3c0dda2663e38635163f0fd6c19748c6eba1c3c8
> Avoid asserting for fontset validity before it is used

No, I don't think so.  We'd just abort earlier in that function.

diff --git a/src/fontset.c b/src/fontset.c
index f196dee..e004cd9 100644
--- a/src/fontset.c
+++ b/src/fontset.c
@@ -967,6 +967,9 @@ face_for_char (struct frame *f, struct face *face, int c,
 #endif
     }
 
+  if (face->fontset < 0 && !face->font)
+    return face->id;
+
   eassert (fontset_id_valid_p (face->fontset));
   fontset = FONTSET_FROM_ID (face->fontset);
   eassert (!BASE_FONTSET_P (fontset));

bug#63283: Emacs crash within fontset

[Seppo Ronkainen via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Your suggested patch successfully avoided the abort, great !

Where the abort would previously occur now instead the buffer renders all characters with four digit code point rectangular representations which I would consider preferable to an abort the entire emacs server session with thirty or so windows. Tested with visiting a large file emacs/src/xdisp.c and Emacs swapps back to render regular characters when text is scaled back. I would vote for your suggested fix, well done thank you !




Live long and prosper

Sent with Proton Mail secure email.

------- Original Message -------
On Sunday, May 7th, 2023 at 6:48 AM, Eli Zaretskii <eliz@gnu.org> wrote:


> > Date: Sat, 06 May 2023 22:36:35 +0000
> > From: Seppo Ronkainen sodr80@protonmail.com
> > Cc: 63283@debbugs.gnu.org
> > 
> > (gdb) frame 2
> > #2 0x00000055559233c4 in face_for_char (f=0x55564c8470, face=0x5556292b90, c=12302, pos=1, object=XIL(0)) at fontset.c:970
> > 970 eassert (fontset_id_valid_p (face->fontset));
> > (gdb) info locals
> > fontset = XIL(0)
> > rfont_def = XIL(0x555605b2b8)
> > charset = XIL(0x300effff7380)
> > face_id = 85
> > id = 1432078920
> > (gdb) print font_driver_list
> > $1 = (struct font_driver_list *) 0x55560d77e0
> > (gdb) print f->font_driver_list
> > $2 = (struct font_driver_list *) 0x5556274ae0
> > (gdb) frame 3
> > #3 0x00000055555bd308 in FACE_FOR_CHAR (f=0x55564c8470, face=0x5556292b90, character=12302, pos=1, object=XIL(0)) at dispextern.h:1911
> > (gdb) p face->font
> > $5 = (struct font *) 0x0
> > (gdb) p face->fontset
> > $6 = -1
> > (gdb) q
> 
> 
> Thanks. If you apply the patch below, does it avoid the abort?
> 
> > relevant commit ?
> > 3c0dda2663e38635163f0fd6c19748c6eba1c3c8
> > Avoid asserting for fontset validity before it is used
> 
> 
> No, I don't think so. We'd just abort earlier in that function.
> 
> diff --git a/src/fontset.c b/src/fontset.c
> index f196dee..e004cd9 100644
> --- a/src/fontset.c
> +++ b/src/fontset.c
> @@ -967,6 +967,9 @@ face_for_char (struct frame *f, struct face *face, int c,
> #endif
> }
> 
> + if (face->fontset < 0 && !face->font)
> 
> + return face->id;
> 
> +
> eassert (fontset_id_valid_p (face->fontset));
> 
> fontset = FONTSET_FROM_ID (face->fontset);
> 
> eassert (!BASE_FONTSET_P (fontset));

bug#63283: Emacs crash within fontset

[Eli Zaretskii]

> Date: Mon, 08 May 2023 21:22:21 +0000
> From: Seppo Ronkainen <sodr80@protonmail.com>
> Cc: 63283@debbugs.gnu.org
> 
> Your suggested patch successfully avoided the abort, great !
> 
> Where the abort would previously occur now instead the buffer renders all characters with four digit code point rectangular representations which I would consider preferable to an abort the entire emacs server session with thirty or so windows. Tested with visiting a large file emacs/src/xdisp.c and Emacs swapps back to render regular characters when text is scaled back. I would vote for your suggested fix, well done thank you !

Displaying the characters as hex codes is the expected behavior in
this case, since Emacs is unable to find a font that suits its needs.

So I've now installed the change on the emacs-29 branch, and I'm
closing the bug.

Thanks.