Re: Bootstrap failure with enable-checking in _Unwind_GetGR (Zack's function.c changes broken?)

[Zack Weinberg <zack at codesourcery dot com>]

On Sun, Sep 02, 2001 at 01:30:04PM +0200, Andreas Jaeger wrote:
> Zack Weinberg <zack@codesourcery.com> writes:
> 
> > Here's a patch.  Would you mind testing?  If I try to do an
> > always-collect build it'll take three days.  It has run to the end of
> > stage1, and I'm doing the C testsuite, which should only take the rest
> > of today.
> 
> This seems to fix the issue I reported but now I get the following
> failure:
...
> /builds/gcc/checking/i686-pc-linux-gnu/libstdc++-v3/include/bits/std_ostream.h:194: Internal
>    error #19991203.
> /builds/gcc/checking/i686-pc-linux-gnu/libstdc++-v3/include/bits/std_ostream.h:194: Internal
>    compiler error in expand_call_inline, at cp/optimize.c:881
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
> make[4]: *** [misc-inst.lo] Error 1

I'm getting the same failure in a build with the default checking
settings.  This is the ICE location:

  /* Close the block for the parameters.  */
  scope_stmt = build_stmt (SCOPE_STMT, DECL_INITIAL (fn));
  SCOPE_NO_CLEANUPS_P (scope_stmt) = 1;
  my_friendly_assert (DECL_INITIAL (fn)
		      && TREE_CODE (DECL_INITIAL (fn)) == BLOCK,
		      19991203);
  remap_block (scope_stmt, NULL_TREE, id);
  STMT_EXPR_STMT (expr)
    = chainon (STMT_EXPR_STMT (expr), scope_stmt);

And sure enough, DECL_INITIAL (fn) is bogus.

(gdb) call debug_tree (fn)
 <function_decl 0x409a92f4 __comp_dtor
    type <method_type 0x409a53cc>
    asm_written used public static in_system_header decl_5 QI
    file .../libstdc++-v3/include/bits/std_ostream.h line 194
    context <record_type 0x40991ca8 sentry>
    abstract_origin <function_decl 0x409a5438 sentry>
    arguments <parm_decl 0x409a93cc __in_chrg>
    result <result_decl 0x40c2fca8>
    initial <error_mark 0x40141450>	*****
    template-info 0x406476f0
    (mem:QI (symbol_ref/i:SI ("_ZNSo6sentryD1Ev")) 0)
    chain <function_decl 0x409a5654 operator b>>

This does not seem like it would be related to my changes; on the
other hand, I don't see anything else in the recent change logs that
would have triggered it.

Is anyone seeing this failure *without* the patch I sent Andreas?

zw