This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: libtool litter
- To: rth at redhat dot com
- Subject: Re: libtool litter
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Mon, 14 May 2001 20:34:04 -0700
- Cc: dave at hiauly1 dot hia dot nrc dot ca, law at redhat dot com, oliva at lsd dot ic dot unicamp dot br, gcc-bugs at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
- Organization: CodeSourcery, LLC
- References: <20010514125855.A868@redhat.com>
>>>>> "Richard" == Richard Henderson <rth@redhat.com> writes:
Richard> On Mon, May 14, 2001 at 03:55:15PM -0400, John David
Richard> Anglin wrote:
>> The first thing that goes into c++ assembler files compiled
>> with exceptions is an import declaration for `terminate'. Is
>> that ok?
Richard> I'm thinking that things should be adjusted such that the
Richard> c++ front end presents a function callback to get at that
Richard> data rather than creating it so early. Cause we don't
Richard> even know if terminate will actually be needed at this
Richard> point.
Yes. It's a little cheesy that protect_cleanup_actions is a variable,
rather than a function. It's weird to ask the front-end to create an
expression tree for a call when it's not in the scope of any function.
So, I think we should make protect_cleanup_actions a function-pointer.
NULL if we don't need it, defined by front-ends as necessary.
Then, terminate_node should be protected by a function in the C++
front-end that creates it lazily.
Dave, Jeff -- would one of you send me the target triplet? I gather
all I have to do is run cc1plus on any file to see the bug? And
describe the bug a little so that I can know I'm seeing it.
Thanks,
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com