This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Red-black trees was Re: [PATCH] : PR libstdc++ 3349
- To: Phil Edwards <pedwards at disaster dot jaj dot com>
- Subject: Re: Red-black trees was Re: [PATCH] : PR libstdc++ 3349
- From: Benjamin Kosnik <bkoz at redhat dot com>
- Date: Thu, 23 Aug 2001 17:08:59 -0700 (PDT)
- cc: Sylvain Pion <Sylvain dot Pion at sophia dot inria dot fr>, gcc-patches at gcc dot gnu dot org, libstdc++ at gcc dot gnu dot org
> I really really wish I could read the thread surrounding the original change.
> This PR has been bothering me.
Yeah this is a bummer.
The history of the "John Potter in 1999" change is
completely lost. Perhaps somebody will re-find the bug that caused this?
I think this is a sane way to proceed.
> Unless there are objections, tomorrow morning I'll apply the patch from
>
> http://gcc.gnu.org/ml/gcc-patches/2001-08/msg00993.html
>
> along with the testcase, and a more descriptive ChangeLog. It's a regression
> from 2.95 and 3.0, so it goes on the branch as well.
Sounds good. As long as a testsuite entry is checked in as well, we won't
run into the problem of wondering what this fix was in 2 years when we
can't read the archives of this list on the web
:(
> Also I'll be writing up some notes on "hinting" insertion, pointing out
> the conditions under which the hint will help, versus those where we have
> to fall back on straight insertion. In the PR's testcase, it wouldn't
> have helped, but we didn't fall back like we should have.
Great, thanks.
> The same thing should be tested on multimaps; the insert_equal() function
> might suffer from the same bug. I haven't checked yet.
Huh. Yeah, it would be nice to get all this stuff at once.
thanks