On 05/11/2015 02:41 PM, Mikhail Maltsev wrote:
On 09.05.2015 0:54, Jeff Law wrote:
Both patches are approved. Please install onto the trunk.
jeff
Sorry for delay. When I started to work on this task, I wrote that I'll
test the patches on couple of other platforms (not just x86). Probably I
should have done it earlier, because I missed a couple of important
details, which could break the build. Fortunately I did several tests
before merging into trunk, and I think I need an advice on testing (or
maybe some reworking).
It happens. This kind of problem is part of what Trevor's patches are
improving for us.
For many years, the preferred style of coding in GCC was to create
target macros, the conditionalize code based on those macros. That
results in a lot of code in GCC that is rarely actually compiled.
In general, is there a recommended set of targets that cover most
conditionally compiled code? Also, the GCC Wiki mentions some automated
test services and compile farm. Is it possible to use it to test a patch
on many targets?
There's a makefile fragment in contrib which will build a large number
of targets that you might find helpful. Of course without some baseline
to compare against, it's of less value.
Finally, I could try to break the patch into smaller pieces, though I
don't know if it's worth the efforts.
I doubt it's worth the effort at this point.
P.S. Bootstrapped/regtested on x86_64-unknown-linux-gnu {,-m32}
(C,C++,lto,objc,fortran,go), crosscompiled and regtested (C and C++
testsuites) on sh-elf, mips-elf, poweperpc-eabisim and arm-eabi simulators.
These seem like a reasonable set of targets, especially if you could add
one cc0 target (h8/300, v850, m68k come to mind as candidates). I also
doubt you need to do the full testing with simulators for this work.
I'd think that bootstrapping one target, then just build the cross tools
for the others would be fine.
jeff
