bug-gmp
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Compiler optimize bug(?) & test failures on HPUX 10.20


From: Kevin Ryde
Subject: Re: Compiler optimize bug(?) & test failures on HPUX 10.20
Date: 03 May 2001 07:40:57 +1000
User-agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/20.5

"Byers, Denton W." <address@hidden> writes:
> 
> I'm getting a compile error when compiling tstbit.c: 70 (in /mpz) with just
> the default settings assigned by configure.

Alas yes, it tickles a compiler bug in gcc 2.95.2.  Replacement below.

> I've seen lots of posts regarding insn errors in other software newsgroups
> but no general solutions other than compiler/source upgrades.  After
> searching for many hours, I read where someone was having problems compiling
> some other package with gcc's optimization flags.  On a hunch, I created a
> fresh source tree I got rid of the optimize flag altogether setting
> CFLAGS="-g" on configure's command line.

Yes turning down the optimizations often helps.

> Everything compiled cleanly but half the tests fail, some dumping
> core.
>
> hppa1.1-hp-hpux10.20

I think we've tested on such systems, so I'm not sure what would be
going wrong.  Please provide the whole output from "make check", or at
least the list of which tests come out as "PASS" and "FAIL".  Pick a
simple one if you like to try to trace the problem.

If you just want to get something working, a generic C build might
give more joy,

    ./configure --target=none-hp-hpux10.20 CFLAGS="-g -DNO_ASM"

or similar.


Attachment: tstbit.c
Description: Text Data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]