scide segmentation fault on launch

[ahihi]

Environment

• SuperCollider version: 3.13.0
• Operating system: Debian 12 (bookworm) i386
• Other details (Qt version, audio driver, etc.): Qt 5.15.8

Steps to reproduce

launch scide.

Expected vs. actual behavior

it segfaults immediately, expected it not to.

this occurs with SC installed from apt, as well as with a manually built 3.13.0 (-DSC_QT=ON -DCMAKE_BUILD_TYPE=Debug -DNATIVE=ON -DCMAKE_INSTALL_PREFIX=~/usr -DSUPERNOVA=OFF). command line sclang seems to work fine.

gdb doesnt tell me much:

GNU gdb (Debian 13.1-3) 13.1
...
Reading symbols from editors/sc-ide/scide...
(gdb) run
Starting program: /home/ahihi/Documents/supercollider/build-debug/editors/sc-ide/scide
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/i386-linux-gnu/libthread_db.so.1".
[New Thread 0xa9477b40 (LWP 15626)]
[New Thread 0x9d7ffb40 (LWP 15627)]
[New Thread 0x9ca74b40 (LWP 15628)]
[Detaching after fork from child process 15629]
[Detaching after fork from child process 15630]
[Detaching after fork from child process 15631]
[New Thread 0x9c273b40 (LWP 15634)]
[New Thread 0x9ba72b40 (LWP 15635)]
[New Thread 0x9b271b40 (LWP 15636)]
[New Thread 0x9aa70b40 (LWP 15637)]
[New Thread 0x9a26fb40 (LWP 15638)]
[New Thread 0x99a6eb40 (LWP 15639)]
[New Thread 0x9926db40 (LWP 15640)]
[New Thread 0x98a6cb40 (LWP 15641)]
[New Thread 0x9826bb40 (LWP 15642)]
[New Thread 0x97a6ab40 (LWP 15643)]
[New Thread 0x96a68b40 (LWP 15645)]
[New Thread 0x97269b40 (LWP 15644)]
[New Thread 0x96267b40 (LWP 15646)]
[Detaching after fork from child process 15647]
[New Thread 0x95a66b40 (LWP 15648)]
[New Thread 0x950ffb40 (LWP 15649)]
[New Thread 0x948feb40 (LWP 15650)]
[New Thread 0x940fdb40 (LWP 15651)]

Thread 1 "scide" received signal SIGSEGV, Segmentation fault.
__GI_memcpy () at ../sysdeps/i386/i686/memcpy.S:70
70	../sysdeps/i386/i686/memcpy.S: No such file or directory.
(gdb) bt
#0  __GI_memcpy () at ../sysdeps/i386/i686/memcpy.S:70
#1  0x938ec800 in ?? ()
Backtrace stopped: previous frame inner to this frame (corrupt stack?)

[ahihi]

valgrind gives me a little more info:

==20781== Invalid write of size 4
==20781==    at 0x4848348: __GI_memcpy (vg_replace_strmem.c:1121)
==20781==    by 0x141635F0: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x14676D10: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x14BF4684: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x14A60F14: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x141BE366: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x141BEDEE: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x1418DB76: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x1418FE49: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x5195E84: ??? (in /usr/lib/i386-linux-gnu/sse2/libQt5Gui.so.5.15.8)
==20781==    by 0x5196215: ??? (in /usr/lib/i386-linux-gnu/sse2/libQt5Gui.so.5.15.8)
==20781==    by 0x548A662: QPlatformBackingStore::composeAndFlush(QWindow*, QRegion const&, QPoint const&, QPlatformTextureList*, bool) (in /usr/lib/i386-linux-gnu/sse2/libQt5Gui.so.5.15.8)
==20781==  Address 0x2b628000 is not stack'd, malloc'd or (recently) free'd
==20781==
==20781==
==20781== Process terminating with default action of signal 11 (SIGSEGV)
==20781==  Access not within mapped region at address 0x2B628000
==20781==    at 0x4848348: __GI_memcpy (vg_replace_strmem.c:1121)
==20781==    by 0x141635F0: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x14676D10: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x14BF4684: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x14A60F14: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x141BE366: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x141BEDEE: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x1418DB76: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x1418FE49: ??? (in /usr/lib/i386-linux-gnu/dri/i915_dri.so)
==20781==    by 0x5195E84: ??? (in /usr/lib/i386-linux-gnu/sse2/libQt5Gui.so.5.15.8)
==20781==    by 0x5196215: ??? (in /usr/lib/i386-linux-gnu/sse2/libQt5Gui.so.5.15.8)
==20781==    by 0x548A662: QPlatformBackingStore::composeAndFlush(QWindow*, QRegion const&, QPoint const&, QPlatformTextureList*, bool) (in /usr/lib/i386-linux-gnu/sse2/libQt5Gui.so.5.15.8)

looks like it may be a graphics driver issue.

[JordanHendersonMusic]

Is that a 32bit system?

[ahihi]

yes.

[JordanHendersonMusic]

So 32bit support isn't guaranteed on linux (according to the main readme), only windows (I don't know why). However there are additional steps you can take to compile for raspberry pi, did you try those? This would explain why the command line version works and the ide does not.

[ahihi]

it didnt occur to me as this is an Asus EeePC 900 (Intel Atom) rather than a Pi, but indeed compiling without QtWebEngine seems to fix it! thank you :)