This is the mail archive of the
mailing list for the GCC project.
Re: Successfull build gcc 3.0.4
- From: Janis Johnson <janis187 at us dot ibm dot com>
- To: Anders Carlsson <anders dot carlsson at mds dot mdh dot se>
- Cc: gcc at gcc dot gnu dot org
- Date: Wed, 6 Mar 2002 14:38:52 -0800
- Subject: Re: Successfull build gcc 3.0.4
- References: <200203050800.JAA00293@legolas.mdh.se>
On Tue, Mar 05, 2002 at 09:00:56AM +0100, Anders Carlsson wrote:
> I have successfully "make bootstrap-lean" on SPARC (SS10/RT626):
> $ config-guess
> $ gcc -v
> Reading specs from /opt/FSFgcc3/lib/gcc-lib/sparc-sun-solaris2.5.1/3.0.4/specs
> Configured with: ../configure --prefix=/opt/FSFgcc3
> Thread model: posix
> gcc version 3.0.4
Thanks! I've added this to the GCC 3.0 build status list at
I'm slightly uncomfortable doing that, though, because of your change to
the specs file that was warned against.
> However, something weird happened. Last fall I bootstrapped gcc 3.0.1 by
> using gcc 2.7.2, and it compiled through all stages without errors. Since
> I wanted to simplify compiling C++ programs, I added "-R /opt/FSFgcc3/lib" to
> the link command in the specs file of the installed version, despite
> GCC warnings about doing that.
> Ok. So now I tried to compile 3.0.4 with 3.0.1. It ran through the first two
> stages, and crashed out with "Error 1" somewhere in stage 3. I was too tired
> to debug what had happened. Could this be due to the configuration change
> above, something with mismatching libraries in the stage1/2 xgcc compiler?
I don't know; perhaps someone else will answer this question.
> Oh well. I cleaned out the objdir completely and bootstrapped 3.0.4 again,
> this time using the ancient 2.7.2 compiler. Voila. Everything ran as it
> /Anders Carlsson