From: johnhm@vaxc.cc.monash.edu.au Subject: Re: GCC 2.11b problems Date: 20 May 1992 04:40:06 GMT
In article <2417@ariel.its.unimelb.EDU.AU>, rab@ariel.ucs.unimelb.edu.au (Richard Alan Brown) writes:
> I have been running the MCC interim version of linux with no apparent
> problems (although I hadn't been doing a lot with it). Yesterday I
> installed the 2.11blib, 2.11bmisc, binutils and 0.96include archives
> from banjo and followed the installation instructions. Everything went
> without a hitch. I ran the "inst2.x" and then the "fixfiles" scripts. I
> booted linux with the 0.96 Bootimage.
>
> Everything was looking good until I tried to compile the linux 0.96.
> (Yes, I had deleted the includes and installed 0.96include.tar into
> /usr, and removed the "-nostincludes" from all the Makefiles).
> I get that old:
>
> GENERAL PROTECTION: 0000
> EIP ...
> EFLAGS: ...
> ...
> gcc: Internal Compiler Error: program cc1 got fatal signal 11
>
> ... which I seem to recall people used to get when they used incompatible
> versions of cpp/cc1/as/ .. etc ... But I have reinstalled all of these !!!
>
> I also noted that repeating the make resulted in a similar error from a
> different point in the compile! Sometimes cpp crashed, sometimes cc1, ...
> If I change to the directory where the error occured and try to make the
> offending module by hand, everything works ... ?!
>
>
> rab
> ------------------------------------------------------------------------
> Richard Brown | E-mail: rab@tauon.ph.unimelb.EDU.AU
> School of Physics | Phone : +61 3 344 5081
> University of Melbourne | Fax : +61 3 347 4783
> Parkville Victoria AUSTRALIA 3052 | Telex : AA35185
The above just about sums up the problem and temporary solution.
From what I've read in gnu.gcc.* this is a due to a Known bug in GCC 2.x and might
be fixed in GCC 2.2. Having Xwindows when running a make is useful because you can cd to
the directory that has the file the compiler failed on and paste the failing gcc
make command onto the command line.
Does someone else have more details about the "future" of the signal 11 bug?
--John H. Morris E-mail: com258a@vaxc.cc.monash.edu.au Monash University Computer Centre Phone: +61 3 5654763 Wellington Road, FAX: +61 3 5654746 Clayton 3168, Australia.