This is the mail archive of the
gdb-patches@sources.redhat.com
mailing list for the GDB project.
Re: [RFA] More C++ v3 tweaks
- To: Michael Elizabeth Chastain <mec at shout dot net>
- Subject: Re: [RFA] More C++ v3 tweaks
- From: Michael Snyder <msnyder at cygnus dot com>
- Date: Thu, 08 Nov 2001 10:58:15 -0800
- CC: gdb-patches at sources dot redhat dot com, jimb at redhat dot com
- Organization: Red Hat
- References: <200111080914.DAA14911@duracef.shout.net>
Michael Elizabeth Chastain wrote:
>
> This patch is approved for both HEAD and gdb_5-1-2001-07-29-branch.
>
> I changesproofread this in an unusual wachangesy: I performed the same substitutions
> of s/void/(void|)/ and s/const &/const ?&/ on my own copy, then
> compared against Michael Snyder's version.
>
> I ran before-and-after tests with:
>
> gcc: 2.95.3 3.0.2 HEAD gcc-3_0-branch
> gdb: HEAD gdb_5_1-2001-07-29-branch
>
> I'm a little bit concerned that this test now PASSes on a lot
> of test output like this:
>
> # gcc-3.0.2-gdb-HEAD/gdb.log
> (gdb) ptype VA
> type = class VA {
> public:
> int va;
>
> VA & operator=(VA const&);
> VA(VA const&);
> VA(VA const&);
> VA();
> VA();
> }
> (gdb) PASS: gdb.c++/virtfunc.exp: ptype VA (obsolescent gcc or gdb)
>
> The duplicate constructor names are confusing.
> There is an open PR for this (PR gdb/27).
>
> Not only does it PASS, but it says "obsolescent gcc or gdb"!
>
> However, there is another test script that covers this (classes.exp),
> so it is okay to let virtfunc.exp PASS on this.
I think the issue of duplicate constructors is separate from
the issue of whether we should accept VA() as well as VA(void).
But I do agree that I wish we (someone) would get around to
addressing all these issues.
Michael