Closed jamessan closed 11 years ago
With these changes and building with GCC 4.8.1 I'm getting a SEGV. The only commit that logically could cause that would be f53f1ba. I'm not sure yet how to fix it, though.
Hmm, it seems to be an optimization problem. Compiling with -O0
or -O1
doesn't crash, but -O2
does. I'll see if I have time to narrow down which of the optimization flags that -O2
adds causes the problem.
As a side note, all of my testing is in an x86 VM since I use Multi-Arch on my amd64 systems.
I've been running this for the past few days and it seems to work well, other than needing -O1
on x86. I'll try to get some time in the next week or so to file GCC bugs for the ICE and optimization-related segfault.
Hi, the applicable parts were applied to jackyf/next. Thanks!
The other day, I hit an issue similar to #1 and wanted to debug it. Since there's no cupt-dbg package, I tried rebuilding and hit an ICE in GCC 4.6. I removed the g++-4.6 requirement[0] and tried building with a newer version. This produced a number of compilation problems, which I've fixed in this series of commits. I also added a cupt-dbg package so that I can more easily debug things in the future.
I have another patch series I've made for the next branch which I'll submit soon.
[0]: It'd be nice if the CMake GCC check were modified to check for a version of GCC >= 4.6 instead of requiring 4.6. I don't know enough about CMake to implement that, so I just removed the g++-4.6 check when I was doing my testing.