This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug ada/6669] Irix6.5 ada bootstrap fail in ada/targparm.adb


PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=6669



------- Additional Comments From ro at techfak dot uni-bielefeld dot de  2003-10-21 18:51 -------
Subject: Re:  Irix6.5 ada bootstrap fail in ada/targparm.adb

> You can close this bug if you do any of the following (in decreasing order of
> preference).  Otherwise, the original submitter is left with no way forward.
> 
> 1) Confirm that the bug is resolved in GNAT 3.15p for mips-sgi-irix6.2, make
> mips-sgi-irix6.2 binaries of that version of GNAT publicly available, and
> document both the version requirement and the location of the binaries.

I'm in the process of bootstrapping current mainline with GNAT 3.15p: the
bug is still present in that release.  This compiler seems to miscompile
the stage1 gcc, since I get ICEs when building libgcc at the end of stage1
with -O2.  Removing -O2 works around this for the moment.  I'll report
details when i submit the patch necessary to get ada to compile.

> 2) Same, but for GCC 3.1 or 3.2 -based GNAT instead of GNAT 3.15p.

Impossible since (at least up to the current Ada merge) GCC 3.x-based GNAT
doesn't work (i.e. complete make gnatlib_and_tools) on IRIX 6.

> 3) Remove -g from ALL_ADA_CFLAGS in stage1.  (Discouraged, as this will make
> debugging harder.)

I think there's another option that avoids all this: provide GCC 3.4-based
GNAT binaries for mips-sgi-irix6.[25].  I'll see how far I get with my
bootstrap attempt and may be able to provide them if this works out.

	Rainer


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]