dwarf patches fail bootstrapping coff platforms]

[Robert Lipe]

This problem was introduced with Mark's most recent change in cp/.  This
is why gcc/ChangeLog didn't have any smoking guns.  These also were off
the radar of my build/test cycle from yesterday.

We've traded one bootstrap problem for another. :-)

If I do
   cvs update  -r gcc-2_95-branch -D990616 method.c call.c init.c decl2.c
kick off
   make CC="stage2/xgcc -Bstage2/" cc1plus
then run the attached script, the program builds.
   cvs update  -r gcc-2_95-branch -D990617 method.c call.c init.c decl2.c
and do the same, the assembler complains.

This seems to point to something in in Mark's 6/16 change.   Maybe it
was the change or maybe his change exposed something else.

I have no doubt that he'll figure out it. :-)

Thanx,
RJL


Robert Lipe wrote:
> This is really strange.   Yesterday morning I did an update of dwarfout.c
> only and performed a full bootstrap on OpenServer.   It succeeded.
> 
> Arturo just point out that today a bootstrap fails.   I can confirm that
> it does.   I don't see any smoking gnus in ChangeLog.
> 
> $ cat /tmp/a 
> ./xgcc -B./  -g --save-temps \
> -B/usr/local/i686-pc-sco3.2v5.0.5/bin/ -c -O2 -fno-implicit-templates \
> -mcoff -I/play/negcs/i686-pc-sco3.2v5.0.5/coff/libio \
>   -I/play/egcs/libio -nostdinc++ \
> /play/egcs/libio/fstream.cc
> 
> (robertl) rjlhome:/play/negcs/gcc
> $ sh /tmp/a
> Assembler: fstream.cc
>         aline 598(cline 3): Unbalanced Symbol Table Entries-Too Many Scope Beginnings
> 
> 
> The assembly in question looks like this:
> 
> .L1734:                                                                         .L1721:
>         call __terminate
>         .align 4
> .L1730:
>         leal -168(%ebp),%esp                                                            popl %ebx
>         popl %esi                                                                       popl %edi
>         movl %ebp,%esp
>         popl %ebp
>         ret
> HERE->  .def    .ef;    .val    .;      .scl    101;    .line   3;      .endef          .def    __11fstreambasei;       .val    .;      .scl    -1;     .endef
>         .align 4                                                                        .def    __11fstreambaseii;      .val    __11fstreambaseii;      .scl    2;      .tag    fstreambase;    .size   112;    .type   0150;   .endef
>         .globl __11fstreambaseii
> __11fstreambaseii:
>         .def    .bf;    .val    .;      .scl    101;    .line   58;     .endef          .def    this;   .val    8;      .scl    9;      .tag    fstreambase;    .size   112;    .type   030;    .endef                             
> 
> 
> ----- End forwarded message -----