This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: -ftree-check for review
- From: "Sebastian Pop" <sebastian dot pop at inria dot fr>
- To: tromey at redhat dot com
- Cc: "Nic Volanschi" <nic dot volanschi at free dot fr>, gcc-patches at gcc dot gnu dot org
- Date: Sun, 6 May 2007 17:11:19 -0400
- Subject: Re: -ftree-check for review
- References: <200704281544.05708.nic.volanschi@free.fr> <m31whwwjuz.fsf@fleche.redhat.com>
Hi,
Sorry for the time it took me to get to this. In general, I agree with Tom
that we need more work on this patch for minimizing the number of bug reports.
On 5/4/07, Tom Tromey <tromey@redhat.com> wrote:
This approach turns the dumper from something
that is a convenience to GCC developers to something that is required
for proper operation of other parts of GCC. So, that's surprising.
Agreed, if we consider that the gimple pretty printers are user visible
from now on, we have to consider the question of the stability of this
representation and the stability of the way we print the information:
in the case of the "0B", if we fix the pretty printers to print "NULL"
instead, then the condates looking for "0B" will just stop working,
and we get other bug reports asking back for the feature :( so we'd
better get the right thing from the start.
Please realize that I can't approve or reject your patch. I'm trying
to understand what this is about, and presenting my view about
potential problems, since I see this as more than a mere patch but
really a major new user-visible addition -- and consequently a
potential major new maintenance burden. In any case the reviewer,
whoever that may be, is who you must please -- not me. HTH.
Many thanks for the directions, ideas, and encouragements,
Sebastian