This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: template_type_parm and typename_type not supported by dump_type_prefix
- To: oldham at codesourcery dot com (Jeffrey Oldham)
- Subject: Re: template_type_parm and typename_type not supported by dump_type_prefix
- From: "John David Anglin" <dave at hiauly1 dot hia dot nrc dot ca>
- Date: Fri, 22 Jun 2001 19:57:50 -0400 (EDT)
- Cc: gcc-bugs at gcc dot gnu dot org, david dot billinghurst at riotinto dot com
> cc1plus Linux->mips-sgi-irix6.5 cross-compiler no error messages
> cc1plus mips-sgi-irix6.5 compiled using -g no error messages
> cc1plus mips-sgi-irix6.5 compiled using -O1 no error messages
> cc1plus mips-sgi-irix6.5 compiled using -O2 no error messages
> cc1plus mips-sgi-irix6.5 from bootstrap error messages
>
> I may have goofed up some of these test results.
I have been looking at it a bit and I suspect that dump_type_suffix
and dump_type_prefix have been miscompiled. The following tree leads
to sorry being called in dump_type_suffix:
(gdb) p debug_tree (t)
<template_type_parm 7adcb958 A readonly VOID
align 8 symtab 0 alias set 0 reference_to_this <reference_type 7adcb9c0>>
I am a bit puzzled as to why TEMPLATE_TYPE_PARM is not listed when I
do:
(gdb) ptype enum tree_code
This only seems to list the C tree codes.
Could this be part of the problem or just a gdb mystery? I would not be
surprised if the compiler optimization level affects the problem. My
build was "-O2 -g". I am currently doing a build with "-O3 -g -fPIC".
I think I need to look at the assembler code for dump_type_suffix to
see if the switch statement in it is ok.
Dave
--
J. David Anglin dave.anglin@nrc.ca
National Research Council of Canada (613) 990-0752 (FAX: 952-6605)