This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Fix bootstrap failure caused by tree-ssa-loop.c
- From: Jan Hubicka <jh at suse dot cz>
- To: law at redhat dot com
- Cc: Jan Hubicka <jh at suse dot cz>, Andrew MacLeod <amacleod at redhat dot com>,Diego Novillo <dnovillo at redhat dot com>, Jan Hubicka <hubicka at ucw dot cz>,gcc-patches <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 15 Dec 2003 23:46:01 +0100
- Subject: Re: [tree-ssa] Fix bootstrap failure caused by tree-ssa-loop.c
- References: <20031211210135.GC27614@kam.mff.cuni.cz> <200312152242.hBFMgSH7018895@speedy.slc.redhat.com>
> In message <20031211210135.GC27614@kam.mff.cuni.cz>, Jan Hubicka writes:
> >That breaks not only by Zdenek's code. Any edge splitting or such would
> >break it. How it is supposed to work?
> Which is why PRE (for example) rebuilds the dominator tree.
>
> I don't think it's particularly interesting to describe how it is supposed
> to work since I don't think the current implementation is desirable long
> term since it's intimately ties to block indices. That's a long term
> recipe for disaster.
Yes, Zdenek already re-implemented it and it seems to me that he did
very good job about it.
I was quite surprised by this as for instance tail call probably got
this wrong while splitting the entry block to produce loopback edges.
I've completely missed this part of interface.
We really should think on making some kind of sanity checkers for such a
non-obvious parts of interaces when we don't have any documentation...
Honza
>
>
> jeff