Bug in pict3d leads to crash in DrRacket

Questions & Answers
1

There has been a bug with pict3d for some time now.

Already posted the problem a couple of times on the racket discord. Hopefully someone here can fix it

2

Thanks for the bump. Perhaps unsurprisingly, I'm unable to reproduce this bug on macOS. Are there other linux desktop users that can test this?

1 Like
3

Please, could anyone else with linux ( or another os besides mac os ) try to replicate the bug?

4

I see the bug as well. The current-pict3d-legacy? parameter does not change the results.

5

This appears to be a pretty fundamental issue in racket/draw. The following program crashes for me (on 8.17):

#lang racket
(require racket/gui/base)

(define canvas (new canvas% [parent (new frame% [label ""])] [style '(gl)]))
(send (send canvas get-dc) get-gl-context)
6

For me, too:

philip@bastet:/tmp$ cat sam.rkt 
#lang racket
(require racket/gui/base)

(define canvas (new canvas% [parent (new frame% [label ""])] [style '(gl)]))
(send (send canvas get-dc) get-gl-context)
philip@bastet:/tmp$ racket sam.rkt 

(process:410960): Gtk-WARNING **: 16:25:02.222: Locale not supported by C library.
        Using the fallback 'C' locale.
Gtk-Message: 16:25:02.312: Failed to load module "colorreload-gtk-module"
Gtk-Message: 16:25:02.312: Failed to load module "window-decorations-gtk-module"
invalid memory reference.  Some debugging context lost
  context...:
   /gnu/store/z0qlv55n78aajs2f6r5jcq1rfcwrdlrb-racket-vm-cs-8.17/opt/racket-vm/collects/racket/private/more-scheme.rkt:266:2: call-with-exception-handler
   /gnu/store/z0qlv55n78aajs2f6r5jcq1rfcwrdlrb-racket-vm-cs-8.17/opt/racket-vm/collects/racket/private/promise.rkt:104:0: force/generic
   /gnu/store/1c5x53nqspxpsqqipvzzf99wpas532qb-racket-8.17/lib/racket/pkgs/gui-lib/mred/private/wx/gtk/gl-context.rkt:352:0: make-gtk-drawable-gl-context
   /gnu/store/1c5x53nqspxpsqqipvzzf99wpas532qb-racket-8.17/lib/racket/pkgs/gui-lib/mred/private/wx/gtk/gl-context.rkt:465:0: make-gtk-widget-gl-context
   /gnu/store/1c5x53nqspxpsqqipvzzf99wpas532qb-racket-8.17/lib/racket/pkgs/gui-lib/mred/private/wx/gtk/dc.rkt:190:4: get-gl-context method in dc%
   .../private/arrow-higher-order.rkt:379:33
   body of "/tmp/sam.rkt"

(The Gtk locale and module-loading messages are probably due to my local setup, but the rest shouldn't be.)

7

This code is not crashing for me on Linux. I'm running Mint 20, which is fairly old, and Racket from git(only 2 commits behind).

8

Are you suggesting pict3d is bugged because of an error with racket/draw?

9

Yes, I do not think the bug is in pict3d.

1 Like
10

I tried this on another Linux machine and was still not able to replicate it. This was a Linux Mint 22 machine.

It would be helpful to know more about which systems are crashing:
kernel version
X or Wayland
graphics driver
etc.

Also, reproducing the crash in gdb might reveal more about where the problem lies. If not familiar with gdb, try something like:

# gdb racket
(gdb) run crash.rkt
<crash occurs>
(gdb) bt full
<gdb stack trace>
1 Like
11

| LiberalArtist
August 31 |

  • | - |

samth:

This appears to be a pretty fundamental issue in racket/draw. The following program crashes for me (on 8.17):

For me, too:

philip@bastet:/tmp$ cat sam.rkt 
#lang racket
(require racket/gui/base)

(define canvas (new canvas% [parent (new frame% [label ""])] [style '(gl)]))
(send (send canvas get-dc) get-gl-context)

FWIW, this program doesn't crash for me on macOS 12.7.6 (!) running Racket 8.17 [cs].

12

I'm getting this on Wayland. Here's the backtrace:

Thread 1 "gracketcs" received signal SIGSEGV, Segmentation fault.
Downloading source file /build/libx11-aL6a2q/libx11-1.8.7/build/src/../../src/QuExt.c
0x00007ffff54e9205 in XQueryExtension (dpy=dpy@entry=0x555555787a90, name=name@entry=0x7ffff092317f "GLX",
    major_opcode=major_opcode@entry=0x5555561e9ad4, first_event=first_event@entry=0x7fffffffdb34,
    first_error=first_error@entry=0x5555561e9ad8) at ../../src/QuExt.c:48
warning: 48     ../../src/QuExt.c: No such file or directory
(gdb) where
#0  0x00007ffff54e9205 in XQueryExtension
    (dpy=dpy@entry=0x555555787a90, name=name@entry=0x7ffff092317f "GLX", major_opcode=major_opcode@entry=0x5555561e9ad4, first_event=first_event@entry=0x7fffffffdb34, first_error=first_error@entry=0x5555561e9ad8) at ../../src/QuExt.c:48
#1  0x00007ffff091f5dc in InitDisplayInfoEntry (dpy=0x555555787a90) at ../src/GLX/libglxmapping.c:639
#2  __glXLookupDisplay (dpy=dpy@entry=0x555555787a90) at ../src/GLX/libglxmapping.c:755
#3  0x00007ffff09206dd in glXQueryVersion (dpy=0x555555787a90, major=0x54adabf8, minor=0x54adac48) at ../src/GLX/libglx.c:1166
#4  0x0000000051c5d1cf in ??? ()
#5  0x0000000000000026 in ??? ()
#6  0x0000555555574c3a in S_call_help (tc_in=tc_in@entry=0x555555635440 <S_G>, singlep=singlep@entry=1, lock_ts=lock_ts@entry=0)
    at racket/src/ChezScheme/c/schlib.c:249
#7  0x0000555555574ecd in S_call (argcnt=2, cp=0x43ab0afd, tc=0x555555635440 <S_G>) at racket/src/ChezScheme/c/schlib.c:212
#8  Scall2 (cp=0x43ab0afd, x1=x1@entry=0x43ac175d, x2=x2@entry=0x43a64869) at racket/src/ChezScheme/c/schlib.c:166
#9  0x000055555556749c in racket_boot (ba=ba@entry=0x7fffffffdd80) at racket/src/cs/c/boot.c:256
#10 0x0000555555565718 in bytes_main
    (argc=<optimized out>, argv=<optimized out>, wm_is_gracket_or_x11_arg_count=0, gracket_guid_or_x11_args=0x55555561e127 "0x0")
    at racket/src/cs/c/main.c:287
#11 0x00007ffff7c2a1ca in __libc_start_call_main
    (main=main@entry=0x555555564e60 <main>, argc=argc@entry=1, argv=argv@entry=0x7fffffffdfc8)
    at ../sysdeps/nptl/libc_start_call_main.h:58
#12 0x00007ffff7c2a28b in __libc_start_main_impl
    (main=0x555555564e60 <main>, argc=1, argv=0x7fffffffdfc8, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffffffdfb8) at ../csu/libc-start.c:360
#13 0x0000555555564fe5 in _start ()

And the system is running an Intel graphics card with the xe driver, kernel 6.11.0 on Ubuntu.

1 Like
13

This post mentioning Wayland remember me that on my Linux (Ubuntu) i no more use Wayland and use Xorg simply because Racket does not works with Wayland. I switched one year ago, so i can not remember what happened, ah si, the package manager GUI was freezing and Racket was not usable for this reason (but i could have installed package from CLI , annoying solution)...I do not know if i'm off topic, perheaps the problem is not coming from Racket but with Wayland or the Racket code needs to be adapted for Wayland, i do not know...

update: note i can not confirm that with Wayland, as i use fvwm crystal with an uptime of 21 days...i can not check it now :sweat_smile:

14

Running with GDK_BACKEND=x11 fixes the crash. I recommend trying that as a workaround.

2 Likes
15

I think i provided all that information on the github issue

Could you elaborate a little more on how to use gdb or the like ? i have no experience using such tools

Can confirm running pict3d on x server doesnt crash drracket. The problem at least for me happens with wayland

16

currently using X11 and no problem with DrRacket to run the test from github issue

mattei@acer:~$ echo $XDG_SESSION_TYPE
x11
mattei@acer:~$ echo $WAYLAND_DISPLAY

mattei@acer:~$ loginctl show-session 3 -p Type
Type=x11

2025-09-07-000412_1920x1080_scrot
2025-09-07-000412_1920x1080_scrot1920×1080 296 KB

but can not test it under Wayland as it requires closing all my session

1 Like