This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug optimization/13985] [3.4/3.5 regression] ICE in gcc.c-torture/compile/930621-1.c
- From: "ebotcazou at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 20 Apr 2004 09:24:30 -0000
- Subject: [Bug optimization/13985] [3.4/3.5 regression] ICE in gcc.c-torture/compile/930621-1.c
- References: <20040202213500.13985.cato@df.lth.se>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From ebotcazou at gcc dot gnu dot org 2004-04-20 09:24 -------
> I don't understand? The comment as I formulated it assumes very little
> knowledge about our implementation (basically just that loop->father is
> superloop of the loop). It only refers to the well-known definition of
> the natural loop.
That loop->father is a superloop of the loop is hardly problematic indeed :-)
But "all paths from the LOOP to LOOP->FATHER's latch" implicitly assumes that
you know the criterion based on which the loop are sorted and, given the lack of
hurry to approve the patch :-), I don't think everyone is as familiar as you
with this stuff.
> I was sort of thinking about extending the comments at the
> fix_{bb,loop}_placement in a similar way (i.e. to explain not only what
> they are doing, but also why it is so).
IMHO such a general comment would rather belong in the class of file-scope
comments (that cfgloopmanip.c lacks btw). As I said, the comment of
fix_bb_placement is nice because it matches the level of abstraction of the
function itself (which doesn't deal with latches).
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13985