Обсуждение: 8.1.4 build failure on ICC 9.1

Hello,

I encountered this the other day and set up a build farm client for it.

http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=meerkat&dt=2006-05-25%2018:16:36

-- 
   === The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240   Providing the most comprehensive  PostgreSQL
solutionssince 1997             http://www.commandprompt.com/
 

"Joshua D. Drake" <jd@commandprompt.com> writes:
> I encountered this the other day and set up a build farm client for it.

> http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=meerkat&dt=2006-05-25%2018:16:36

That NaN problem has been discussed before, and I believe we concluded
it's a compiler bug.  See the archives for the switch to use to avoid
it.
        regards, tom lane

On Thu, 25 May 2006, Tom Lane wrote:

> "Joshua D. Drake" <jd@commandprompt.com> writes:
> > I encountered this the other day and set up a build farm client for it.
>
> > http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=meerkat&dt=2006-05-25%2018:16:36
>
> That NaN problem has been discussed before, and I believe we concluded
> it's a compiler bug.  See the archives for the switch to use to avoid
> it.

I think it was meant as a "feature" by intel, but one man's feature is
another man's bug ;)

The flag to use is -mp1

Also, I see that you are getting all of the tons of output also.  Those
drove me nuts.  I sent in a patch for configure to take some of those
-W flags out which is now applied to HEAD as well as the addition of the
-mp1 flag for the ICC compiler.

I was more interested in the failures on the HEAD build on that box.  I
have had no problems with pl/(perl|python) on my box, though it is using
9.0 vs 9.1, I don't expect that they would have broken things THAT
badly...