gcc Heisenbug
August 6th, 2008 by 
Multimedia Mike
In my last post, when I endeavored to enter a gcc bug regarding the internal compiler error when compiling FFmpeg on PPC, I could not reproduce the problem when I separated the compiler from ccache. I finally got, umm, lucky when the FFmpeg tree got into a state when the compiler threw the error without ccache’s involvement.
So I followed the steps outlined for a proper, useful bug reporting page, as listed on gcc’s bug reporting page. I added the -save-temps option to the command line…
…and the compiler no longer crashes.
The term here is Heisenbug.
Anyway, I have finally entered a bug in order to start tracking this issue.
Posted in FATE Server | 
6 Comments »
August 6th, 2008 at 11:48 pm
You may want to run a memory test if you haven’t done so yet.
August 6th, 2008 at 11:57 pm
To do so would be to challenge the infallibility of Apple hardware. Heresy.
August 7th, 2008 at 2:06 am
A memory error would be unlikely to trigger the bug at the same
place every time.
August 7th, 2008 at 7:47 am
Have you tried running gcc under valgrind?
August 7th, 2008 at 8:11 am
No, though I did try running it under gdb, which is something that is not allowed, apparently.
September 12th, 2008 at 7:04 am
Running gcc under gdb “not allowed”? No, that’s absolutely fine:
http://gcc.gnu.org/wiki/DebuggingGCC