This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug rtl-optimization/15296] [3.4 only] Delayed branch scheduling causing invalid code on cris-*
- From: "hp at bitrange dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 26 May 2004 21:02:14 -0000
- Subject: [Bug rtl-optimization/15296] [3.4 only] Delayed branch scheduling causing invalid code on cris-*
- References: <20040505140818.15296.hp@gcc.gnu.org>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From hp at bitrange dot com 2004-05-26 21:02 -------
Subject: Re: [3.4 only] Delayed branch scheduling
causing invalid code on cris-*
On Wed, 26 May 2004, bangerth at dealii dot org wrote:
> Note
> that the target milestone describes an _intent_: "we'd like to have this
> bug fixed by that version". The known-to-work field describes a _fact_:
> "This bug has been tested against these versions/branches and has been
> verified to trigger/to not trigger the bug there".
Well, that's what I've been trying to say... The known-to
fields should state facts, but your current usage does not
follow the intent you state: when you write release identifiers
rather than branch identifiers, it can't be a *fact* until the
release is done.
Can you consider e.g. writing 3.3-branch, not 3.3.4 in the
known-to-work field? It doesn't have to be an exact branch
identifier as long as it's unique.
Perhaps also consider using the known-to-fail field for releases
only? If not, you'd have to iterate over all bugs to fix
known-to-fail fields for bugs that "disappear" for that field to
state a fact!
> I understand that the fact that we also use branch-names and not only
> release names in these fields can in a few occasions be confusing.
The problem seems to be that you use non-unique terms in those
fields: your refer to the 3.3 branch as 3.3.4, which will
hopefully be the name of a release as well.
Still, please document whatever usage of yours, supposedly in
management.html. Maybe I should open a bug report for that?
(No, *I* don't want to document it since I still consider this
usage confusing if not wrong, if nothing else than that it's
easy to mistakenly remove a "known to fail" field for a release
known to fail (QED). At least until you stop using the same
name for releases as branches.)
brgds, H-P
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15296