Emmanuel Fleury writes: > Andrew Haley wrote: > > > > It seems that no-one has investigated what is really going on. All > > the gdb backtraces are without any debuginfo, and no-one has looked at > > where exactly the segfault happens. No-one has looked at the source > > code that faults, or the assembly code generated by the compiler. > > Yesterday, I asked the guy to compile with debug symbols and to redo the > trace. I'll keep you informed. I'll try to monitor him and get some more > information about it. > > But if the bug disappear when -O2 is turned off, I do suspect > (*eventually*) some problem in the ARM back-end... Am I right ?
It's possible, but probably not. Most bugs that appear at -O2 are coding errors, but let's wait before pushing to any judgment. Andrew.
