On 11 November 2013 18:52, Martin Liška <marxin.li...@gmail.com> wrote:
>>> +2013-10-28 Martin Liska <marxin.li...@gmail.com> >>> + >>> + * gcc.dg/time-profiler-1.c: New test. >>> + * gcc.dg/time-profiler-2.c: Ditto. >>> + > > Yes, I do have commit right. I will bootstrap the patch, test Inkscape > instrumentation and commit it. > I have observed that the time-profiler-2.c has very unstable results, at least when executed under QEMU (ARM target). I am not sure what it is really trying to measure, but the problem is probably linked to the way QEMU handles threads. Any idea about to fix this? Thanks, Christophe.
