This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch][graphite] Check cloog revision when configure.
On Fri, Nov 6, 2009 at 20:09, Tobias Grosser <grosser@fim.uni-passau.de> wrote:
> On Sat, 2009-11-07 at 09:40 +0800, Li Feng wrote:
>> On Sat, Nov 7, 2009 at 2:59 AM, Paolo Bonzini <bonzini@gnu.org> wrote:
>> > If this is the patch:
>> >
>> > - Â#if CLOOG_VERSION_MAJOR != 0 || CLOOG_VERSION_MINOR != 15
>> > + Â#if CLOOG_VERSION_MAJOR != 0 || CLOOG_VERSION_MINOR != 15 ||
>> > CLOOG_VERSION_REVISION < 5
>> > Â choke me
>> > Â #endif
>>
>> Em, this is the patch.
>>
>> - Â#if CLOOG_VERSION_MAJOR != 0 || CLOOG_VERSION_MINOR != 15
>> + Â#if CLOOG_VERSION_MAJOR != 0 || CLOOG_VERSION_MINOR != 15 ||
>> CLOOG_VERSION_REVISION < 4
>>
>> ÂI have noticed in Sebastian's last patch, the
>> CLOOG_VERSION_REVISION will not allow revision 4.
>>
>> @Sebastian, is there some special reason for this? I have tested
>> cloog 0.15.4, and it works (at least) with aermod in polyhedral benchmark.
>>
>
> .4 is enough to fix this problem.
>
> There exist newer CLooG versions that fix other problems:
>
> .5 - Compilation with -Wc++-compat PR40103
> .6 - Add --with-host-libstdcxx to allow usage of static compiled CLooG
> with static or dynamic libstdcxx
> .7 (not yet released) - Fixes some memory leaks
>
> I believe fixing .4 is required and probably .5 will also hit a lot of
> people so it would be helpful to warn them at the beginning.
>
> For .6 and .7 this does only affect a few people, who will find the
> solution on the mailinglists. Forcing all people to upgrade CLooG just
> to make life for a few a little easier does not sound useful to me.
>
> However I do not have a strong opinion about .6 and .7
Let's say that CLooG-PPL 0.15.7 or later revision is required for gcc4.5.
Sebastian