[mpich-discuss] Configure error

[Dave Goodell]

After a cron job runs tonight, you can grab a tarball based on 1.2.1p1 that shouldn't have this problem from here:

http://www.mcs.anl.gov/research/projects/mpich2/downloads/tarballs/nightly/1.2.1/

You'll want to pick a revision >=r7073 to get the fix.

-Dave

On Aug 18, 2010, at 1:40 PM CDT, Dave Goodell wrote:

> I thought this sounded familiar.  It's a bug we fixed a while back on the trunk, but never made it to the 1.2.x series: http://lists.mcs.anl.gov/pipermail/mpich-discuss/2010-February/006480.html
> 
> Basically, the problem is that we don't correctly handle spaces in many of the "precious" autoconf variables, especially CPPFLAGS.  It was fixed on the trunk by r6220-6226.
> 
> 1.3b1 shouldn't suffer from this problem, if you are willing to use it.  I'll see if there's a way to back-port any of this to the 1.2.1 branch.  The workaround I mentioned earlier might also still work for you, but you might hit a different set of related problems later on instead.
> 
> -Dave
> 
> On Aug 18, 2010, at 12:54 PM CDT, Dave Goodell wrote:
> 
>> Hmm... it looks like we have a bug in the configure stage somewhere, but I haven't found it yet.  For some reason your CPPFLAGS argument is clobbering some internal additions to the CPPFLAGS, which confuses the nemesis sub-configure.
>> 
>> You might be able to work around this by adding the following entries to your CPPFLAGS argument for now:
>> 
>> -I/home/ssul/work/distros2/mpich2/mpich2-1.2.1p1/src/openpa/src
>> -I${PWD}/src/openpa/src
>> 
>> -Dave
>> 
>> On Aug 18, 2010, at 11:28 AM CDT, Seung-Jin Sul wrote:
>> 
>>> On 08/18/2010 12:16 PM, Dave Goodell wrote:
>>>> Can you also send us the following two files from the build directory?
>>>> 
>>>> config.log
>>>> src/mpid/ch3/channels/nemesis/config.log
>>>> 
>>>> Nemesis thinks that you don't have a working OpenPA configuration, but the OPA configure output seems OK to me.  In particular, I'll need that second file to see why Nemesis doesn't like OPA's config.
>>>> 
>>>> Otherwise, without seeing the logs, it seems like there's a chance it could be related to providing an "--exec-prefix" argument, since I don't think that we test that capability.
>>>> 
>>>> -Dave
>>>> 
>>>> On Aug 18, 2010, at 10:24 AM CDT, Seung-Jin Sul wrote:
>>>> 
>>>>> Hi,
>>>>> While I configured MPICH2, I found an error.
>>>>> Could you please look into the c.txt I've attached?
>>>>> Thanks.
>>>>> 
>>>>> 
>>>>> Seung.
>>>>> <c.txt>_______________________________________________
>>>>> mpich-discuss mailing list
>>>>> mpich-discuss at mcs.anl.gov
>>>>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>>>> _______________________________________________
>>>> mpich-discuss mailing list
>>>> mpich-discuss at mcs.anl.gov
>>>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>>> OK. I've attached them.
>>> And I found the same error without "--exec-prefix ".
>>> 
>>> Seung.
>>> <config.log><config.log>_______________________________________________
>>> mpich-discuss mailing list
>>> mpich-discuss at mcs.anl.gov
>>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>> 
>> _______________________________________________
>> mpich-discuss mailing list
>> mpich-discuss at mcs.anl.gov
>> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> 
> _______________________________________________
> mpich-discuss mailing list
> mpich-discuss at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss