[Bug bootstrap/66448] [6 Regression] Bootstrap fails on darwin after r224161

["iains at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66448

--- Comment #15 from Iain Sandoe <iains at gcc dot gnu.org> ---
(In reply to Aldy Hernandez from comment #14)
> Created attachment 35734 [details]
> Patch to ICE on other duplicate DW_AT_ attributes

> I can't reproduce any of these with a cross compiler.  I'm building a cross
> with:
> 
> --enable-languages=c,c++ --disable-bootstrap --disable-multilib
> --target=i686-darwin

The triple for the platform I'm testing on is :  x86_64-apple-darwin12

I doubt that i686/x86_64 is significant since the errors show on both multilib
variants.

However, I'd avoid the default version - and suggest at least
xxx-apple-darwin10.

> ...but I still can't get cc1plus to trigger any ICE with the attached patch.

me neither - nevertheless...

> > warning: invalid DWARF generated by the compiler: DIE 0x00000176 has multiple
> > AT_artificial attributes in
> > '/var/folders/tj/17r7407j14d324dzf67cnvxm000114/T//ccNJ6qak.o'.
> 
> Similarly here.
> 
> The attached patch adds ICEs for duplicate
> DW_AT_{artificial,decl_line,decl_file} tags.  Since I'm flying blind here,
> would you mind finding for which files we trigger ICEs and attach the
> preprocessed files so I can reproduce with a cc1/cc1plus?  I assume just one
> of each type will do.

OK. so â the ICEs don't fire - but, nevertheless, the output seems to be
incorrect.

I'll attach the intermediate files in a follow-on post.

for the case of:
./gcc/xgcc -Bgcc
/GCC/gcc-trunk/gcc/testsuite/gcc.c-torture/execute/20020412-1.c  -m64   -O3 -g 
-o t -save-temps

dwarfdump --verify 20020412-1.o
----------------------------------------------------------------------
 File: 20020412-1.o (x86_64)
----------------------------------------------------------------------
Verifying Compile Unit Header chain... ok
Verifying .debug_info... 
error: Range is not in parent
  BLCK {0x000001da}: [0x0000000000000190 - 0x0000000000000204) ERROR
0x0000020a: contains multiple  AT_decl_file attributes.
ERROR 0x0000020a: contains multiple  AT_decl_line attributes.
ERROR 0x0000020a: contains multiple  AT_decl_file attributes.
ERROR 0x0000020a: contains multiple  AT_decl_line attributes.
ERROR 0x0000028d: contains multiple  AT_decl_file attributes.
ERROR 0x0000028d: contains multiple  AT_decl_line attributes.

======

using binutils objdump (which groks mach-o)
â although I realise you don't have a cross-assembler (although I'm working on
that too ;) )

Contents of the .debug_info section:

  Compilation Unit @ offset 0x0:
   Length:        0x39d (32-bit)
   Version:       2
   Abbrev Offset: 0x0

<snip>

    <209>   DW_AT_const_value : 5
 <2><20a>: Abbrev Number: 11 (DW_TAG_structure_type)
    <20b>   DW_AT_decl_file   : 1
    <20c>   DW_AT_decl_line   : 43
    <20d>   DW_AT_decl_file   : 1
    <20e>   DW_AT_decl_line   : 43
    <20f>   DW_AT_decl_file   : 1
    <210>   DW_AT_decl_line   : 43
    <211>   DW_AT_sibling     : <0x222>
 <3><215>: Abbrev Number: 7 (DW_TAG_member)

so it appears that the tools agree about the binary .. 

=====