This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
egcs-971105 alphaev5?-dec-osf4.0 bootstrap failure
- To: egcs at cygnus dot com
- Subject: egcs-971105 alphaev5?-dec-osf4.0 bootstrap failure
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- Date: Tue, 11 Nov 1997 14:39:54 -0500 (EST)
- Cc: ghazi at caip dot rutgers dot edu
I got access to a new alpha system (alphaev5?-dec-osf4.0).
(I'll explain the "?" later.) When I build a clean egcs-971105
bootstrapping with cc for stage1, I get the following:
> stage1/xgcc -Bstage1/ -W -Wall -o f/fini f/fini.o f/proj-h.o
> ./f/fini `echo ./f/str-1t.fin | sed 's,^\./,,'` f/str-1t.j f/str-1t.h
> /bin/sh: 1143 Memory fault - core dumped
> make[2]: *** [f/str-1t.h] Error 139
Some info:
1. I used no special BOOT_CFLAGS when building egcs in this case.
2. gcc2 ss-971104 builds fine.
3. I can build egcs-971105 on the alphaev5? if I use LANGUAGES=c.
4. I can build all of egcs-971105 on a regular alpha-dec-osf4.0 with
no problem.
When I run c-torture, I get lots of lines which say this:
inst emulated pid=1855 <xgcc> va=0x11ffffc10 pc=0x1200098c8 inst=0x2829ffff
Another oddity is that config.guess from testgcc-971104 says
the system is:
> alphaev5-dec-osf4.0b
However egcs-971105 config.guess claims it is:
> alphaev56-dec-osf4.0
So we have OSF 4.0 vs 4.0b, and ev56 vs ev5. The configure
script definitely does different things depending on which of these is
specified. Which config.guess is getting it right?
Thanks,
--Kaveh
--
Kaveh R. Ghazi Project Manager
ghazi@caip.rutgers.edu ICon CMT Corp.