public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* dwarf2: force_decl_die abort for CONST_DECL
@ 2004-04-09 17:33 Graham Stott
  2004-04-09 17:40 ` Andrew Pinski
  0 siblings, 1 reply; 3+ messages in thread
From: Graham Stott @ 2004-04-09 17:33 UTC (permalink / raw)
  To: gcc-bugs

All,

Compiling sid/component/gdb/gdb.cxx triggers an abort in dwarf2out:force_decl_die
becuase it's being handed a CONST_DECL.

  <const_decl 0xbd754a8c cpu_trap_handled
     type <enumeral_type 0xbd754948 cpu_trap_disposition asm_written unsigned SI
         size <integer_cst 0xbd86db7c constant 32>
         unit size <integer_cst 0xbd86db90 constant 4>
         align 32 symtab -1116402880 alias set -1 precision 2 min <integer_cst 0xbd7556f4 0> max <integer_cst 0xbd755708 3>
         values <tree_list 0xbd755618
             purpose <identifier_node 0xbd750a40 cpu_trap_unhandled
             bindings <0xbe39e6b8>
             local bindings <(nil)>>
             value <integer_cst 0xbd755730 constant 0>
             chain <tree_list 0xbd755640
                 purpose <identifier_node 0xbd750a80 cpu_trap_handled
                 bindings <0xbd72f53c>
                 local bindings <(nil)>>
                 value <integer_cst 0xbd755758 constant 1>
                 chain <tree_list 0xbd755668
                     purpose <identifier_node 0xbd750ac0 cpu_trap_reissue
                     bindings <0xbe39e820>
                     local bindings <(nil)>>
                     value <integer_cst 0xbd755780 constant 2>
                     chain <tree_list 0xbd755690
                         purpose <identifier_node 0xbd750b00 cpu_trap_skip
                         bindings <0xbe39e550>
                         local bindings <(nil)>>
                         value <integer_cst 0xbd7557a8 constant 3>>>>> context <namespace_decl 0xbe2fd6c0 sidutil>
         chain <type_decl 0xbd7549b4 cpu_trap_disposition>>
     readonly constant VOID file ../sid/component/gdb/../../include/sidcpuutil.h line 94
     align 1 context <namespace_decl 0xbe2fd6c0 sidutil> initial <integer_cst 0xbd755758 1> chain <const_decl 0xbd754a20 cpu_trap_unhandled>>

Anyone ideas on what needs fixing?

Graham


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: dwarf2: force_decl_die abort for CONST_DECL
  2004-04-09 17:33 dwarf2: force_decl_die abort for CONST_DECL Graham Stott
@ 2004-04-09 17:40 ` Andrew Pinski
  2004-04-09 18:53   ` Graham Stott
  0 siblings, 1 reply; 3+ messages in thread
From: Andrew Pinski @ 2004-04-09 17:40 UTC (permalink / raw)
  To: Graham Stott; +Cc: gcc-bugs, Andrew Pinski


On Apr 9, 2004, at 13:33, Graham Stott wrote:

> All,
>
> Compiling sid/component/gdb/gdb.cxx triggers an abort in 
> dwarf2out:force_decl_die
> becuase it's being handed a CONST_DECL.
(from an enumeration).
> Anyone ideas on what needs fixing?

No idea from me but this is already filed under PR 14718.

Thanks,
Andrew Pinski


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: dwarf2: force_decl_die abort for CONST_DECL
  2004-04-09 17:40 ` Andrew Pinski
@ 2004-04-09 18:53   ` Graham Stott
  0 siblings, 0 replies; 3+ messages in thread
From: Graham Stott @ 2004-04-09 18:53 UTC (permalink / raw)
  To: Andrew Pinski; +Cc: gcc-bugs

Andrew,

Thanks for rthe pointer to PR 14718

The problem is the new dwarf2out_imported_module_or_decl hook
it's passing a CONST_DECL directly down to force_decl_die.

I'm testing a patch which instead passes CONST_DECL's down
to force_type_die that does fix the aborts, but I've no idea
at present if the resulting dwarf2 info generated is correct.

Guess it's time to do some more digging/testing.

Time to build

Thanks
Graham


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2004-04-09 18:53 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-04-09 17:33 dwarf2: force_decl_die abort for CONST_DECL Graham Stott
2004-04-09 17:40 ` Andrew Pinski
2004-04-09 18:53   ` Graham Stott

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).