← Back to index

incomplete stacktrace ?

albatros
#22979
Hello there.

I give opengl a callback for errors. I set a breakpoint in the callback to check in the stacktrace which of my calls to opengl generates an error.


Here is the stack trace from gdb : 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
Static tracepoint 2 at 0x555555562f5b: file main.c, line 402.
(gdb) bt
#0  MessageCallback (source=33350, type=33356, id=1282, severity=37190, length=92, message=0x7fffe6c70b40 "GL_INVALID_OPERATION error generated. The operation is invalid when the uniform is a matrix.",
    userParam=0x0) at main.c:402
#1  0x00007fffe59af378 in ?? () from /usr/lib/libnvidia-glcore.so.440.82
#2  0x00007fffe59af4c0 in ?? () from /usr/lib/libnvidia-glcore.so.440.82
#3  0x00007fffe59af782 in ?? () from /usr/lib/libnvidia-glcore.so.440.82
#4  0x00007fffe5b72a13 in ?? () from /usr/lib/libnvidia-glcore.so.440.82
#5  0x0000555555564f01 in init () at main.c:816
#6  0x0000555555563314 in main () at main.c:500
(gdb)

So my code on #0 (in the callback)
#1,2,3,4 stuffs in the nvidia driver code or somthing like thiks
#5, 6 back to my code

On codeclap I only get #0 and #1

continuing to step in the code enventually get me back to my code.
Edited by albatros on
spx
#22980
Indeed this was and at least partially still is a known problem (or limitation). This will improve a bit with 0.9.5 but codeclap may in some cases still have problems to find the frame base (in order to fetch the return address). I will continue to look into this but didn't give it priority for 0.9.5.

Thanks for the report, I will try to find a few common repro cases for further investigation.
Edited by spx on
← Back to index