lele@idea.sublink.org (Lele Gaifax as idea's SysAdm) (02/04/91)
Hi, I'm here again...
I'm trying to set up gcc 1.39 on a Honeywell Bull XSuperteam X20, a
68020 box running unix sysv 2.2. Probably this name doesn't tell you
anything: it is made here in Italy...
I applied the coff patches, and I was able to build gas 1.36, and it
is running ok. With gcc instead I have a serious problem: everything
compiles ok, 'till stage1. I started to recompile gcc with itself, and
it seems to do his work ok. But the recompiled gcc doesn't produce the
right code...: at the end of the stage, it recompiles gnulib2.c with
the new gcc, and here "cpp got fatal signal 11".
I checked it all, and I think I have isolated the problem in
__builtin_alloca: if I put "#define USE_C_ALLOCA" everywhere, and
disable "#if (__GNUC__)... #define alloca __builtin_alloca" everything
goes well, and I built stage2 with that changes.
I tried to reset the situation, and recompile it with the executable in
stage2, but the same problem appeared.
A few minutes ago I tried using STACK_POINTER_OFFSET, a macro that is used
in expr.c in the body of __builtin_alloca: I set it to 0, without any news.
I looked in the other tm-* and xm-*, but I didn't find anything that can
enlight me...
Have you any hints?
Thanx a lot, lele.
BTW, thanks again to all those who answered to my previous posting!
--
<<< lele gaifas-the 1st sublinker in TN | lele@idea.sublink.org >>>
Se ti vien voglia di lavorare, siediti: vedrai che ti passa...