This is the mail archive of the
gcc-prs@gcc.gnu.org
mailing list for the GCC project.
Re: bootstrap/3314: --enable-threads does not seem to be working for HP-UX 11.00
- From: ljrittle at gcc dot gnu dot org
- To: gcc-bugs at gcc dot gnu dot org, gcc-prs at gcc dot gnu dot org, harri dot pasanen at trema dot com, jdonner at schedsys dot com, ljrittle at gcc dot gnu dot org, nobody at gcc dot gnu dot org
- Date: 18 Feb 2002 23:01:07 -0000
- Subject: Re: bootstrap/3314: --enable-threads does not seem to be working for HP-UX 11.00
- Reply-to: ljrittle at gcc dot gnu dot org, gcc-bugs at gcc dot gnu dot org, gcc-prs at gcc dot gnu dot org, harri dot pasanen at trema dot com, jdonner at schedsys dot com, ljrittle at gcc dot gnu dot org, nobody at gcc dot gnu dot org, gcc-gnats at gcc dot gnu dot org
Synopsis: --enable-threads does not seem to be working for HP-UX 11.00
Responsible-Changed-From-To: ljrittle->unassigned
Responsible-Changed-By: ljrittle
Responsible-Changed-When: Mon Feb 18 15:01:05 2002
Responsible-Changed-Why:
Just marking that I am no longer working on this problem
(sorry to have kept it assigned to myself for so long
with no final progress).
In general, a complete fix requires moving all control from
gcc/config.gcc to gcc/configure.in . It would be better
to provide all legal values (and a hint for the default
for the target) for --enable-threads in gcc/config.gcc
and centralize the seclection of the thread model after
config.gcc is parsed. You will note that the control
logic is currently spread to the following locations:
- before config.gcc is parsed (in configure)
- per-host in config.gcc with much duplication and
many annoying subtle differences
- after config.gcc is parsed (back in configure)
When the user specifies a value for --enable-threads that
is not legal for the target or specifies --enable-threads
without a value and the target default is single; the
"early configure" should abort with a clear message.
A patch was posted that handles the second issue without
moving all the control logic as should (IMHO) be done.
It moved most control logic related to thread model
selection within configure to after config.gcc was
parsed without touching config.gcc. This less than
ideal improvement was never fully accepted or rejected.
http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=3314