Re: glibc fclose prob? was: Problem egcs-2.93.02

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

At 18:09 25.01.99 , Jeffrey A Law wrote:
>
>  In message <4.1.19990125170304.042875f0@mail.lauterbach.com>you write:
>  > No, the problem is that you need newer binutils, which are not yet
>  > released. hjl is working on backporting these patches to the upcoming
>  > binutils-2.9.1.0.20. Current mainline as of now won't build anyways due to
>  > other problems (probably due to CONSTANT_P_RTX handling again? rth?):
>Don't be sure.  Those are the same symptoms I've had with ppc-linux, even
>with the latest binutils snapshots.

Hmm, you know it worked for me last week with cvs-980120. I'll try again
when the problem below is solved.

>
>  > _eprintf
>  > _bb
>  > ../../../egcsm/gcc/varasm.c:3265: Internal compiler error in function
>  > decode_rtx_const
>  > Please submit a full bug report to `egcs-bugs@cygnus.com'.
>  > See <URL:http://egcs.cygnus.com/faq.html#bugreport> for details.
>  > make[4]: *** [libgcc2.a] Error 1
>  > make[4]: Leaving directory `/home/fsirl/obj/cvsm/gcc'
>  > make[3]: *** [stmp-multilib-sub] Error 2
>  > make[3]: Leaving directory `/home/fsirl/obj/cvsm/gcc'
>  > make[2]: *** [stmp-multilib] Error 1
>  > make[2]: Leaving directory `/home/fsirl/obj/cvsm/gcc'
>  > make[1]: *** [bootstrap-lean] Error 2
>  > make[1]: Leaving directory `/home/fsirl/obj/cvsm/gcc'
>  > make: *** [bootstrap-lean] Error 2
>  > 
>  > Last week cvs-980120 of egcs-mainline built fine with the gas-981127
>  > binutils snapshot, as I reported here on the list.
>Can you look at what "x"  looks like in decode_rtx_const?
>
>p debug_rtx(x)
>
>If it's a CONSTANT_P_RTX then it's something rth needs to look at.

Seems really to be a C_P_R problem:
(gdb) r
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: /home/fsirl/obj/cvsm/gcc/./cc1 libgcc2.i -quiet -dumpbase
libgcc2.c -mstrict-align -g -g1 -O2 -W -Wall -version -fPIC -o libgcc2.s
Cannot insert breakpoint 1:
Temporarily disabling shared library breakpoints:
1 2 
GNU C version egcs-2.93.03 19990125 (gcc2 ss-980609 experimental)
(ppc-redhat-linux) compiled by GNU C version egcs-2.91.60 19981201
(egcs-1.1.1 release).

Breakpoint 6, decode_rtx_const (mode=SImode, x=0x1c8ada0, value=0x7fffeb98)
at ../../../egcsm/gcc/varasm.c:3265
3265          abort ();
(gdb) p debug_rtx(x)
(constant_p_rtx:SI (reg/v:SI 83))
$1 = void
(gdb) 

BTW, this reminds me of a small nit in the Makefiles, the c++ parts of
libgcc, eg. tinfo.cc are built with -g compared with -g1 of the C parts. Is
this the way it's meant to be?

Franz.