This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Update a stab string to reflect gcc3
- From: Daniel Berlin <dan at dberlin dot org>
- To: Jim Ingham <jingham at apple dot com>
- Cc: Daniel Jacobowitz <drow at mvista dot com>, Ira Ruben <ira at apple dot com>, Alexandre Oliva <aoliva at redhat dot com>, <gcc-patches at gcc dot gnu dot org>
- Date: Fri, 15 Mar 2002 21:50:05 -0500 (EST)
- Subject: Re: [PATCH] Update a stab string to reflect gcc3
On Fri, 15 Mar 2002, Jim Ingham wrote:
> Daniel,
>
> No real reason (for instance I was NOT planning on cluing a per objfile
> C++ ABI setting off of it right now.) But the whole gdb "set cp-abi"
> bit set me to thinking that there are in fact differences between gcc2 &
> gcc3 compiled code that gdb does need to know about, so just for grins I
> wanted it to be right as soon as possible. Then, if we ever care about
> some of the differences we will have the info we need to act on it.
>
> It is, after all, pretty gross that gdb has to grub through symbols
> looking for functions with an initial _Z to determine that we are using
> the gnu-v3 ABI. Wouldn't it be a lot cleaner if there was a
> gcc3_compiled stab that we could read?
No.
The ABI is not defined by GCC.
It's defined by the IA-64 C++ ABI document.
Thus, we use something not compiler specific to detect it, because it's
not compiler specific itself (of course, there are a few vendor
extensions, but this is in mangling).
Please try to keep that in mind when mucking about in the V3 abi stuff.
It's not just g++'s abi anymore, so you can't rely on only changing gcc to
get some desired behavior.
:)
--Dan