#3824: [PATCH] Unitialised variable
----------------------+----------------------
Reporter: flatcap | Owner: mutt-dev
Type: defect | Status: new
Priority: major | Milestone:
Component: build | Version: 1.5.24
Resolution: | Keywords: patch
----------------------+----------------------
Comment (by kevin8t8):
Tamo, thank you very much. You have my deep gratitude for taking the
time to look into this ticket, #3825, and the other fixes you've
helped with recently. (You too, Brendan and Vincent).
Rich, I'd like to take the opportunity to rant back a bit.
Your torrent of tickets is not helping me with the release. We're now
ONE week away, and you started this crusade just two weeks before
release.
As a new contributor, I need to look more closely at your patches, but
unfortunately I've found that they have been sloppy. See ticket
#3820, or #3814, or #3825. This means I can't trust your patches,
which in turn means I need to spend a lot of time looking at them
before I can commit them at this point.
Furthermore, your response to checking the return value instead of
discarding it in #3820, and other patches I've seen doing the same
thing (e.g #3818) makes me feel like the most important thing to you
is making your "warning tools" happy.
And we have the same problem here. Yes, *theoretically*, initializing
sp is good, but there IS NO actual bug here. The callers of
mutt_parse_rfc822_line() -> mutt_parse_references() are all checking
and passing a non-null p.
I don't need 20 tickets of hastily done fixes that may break more
things than they fix. I don't need all possible error path resource
handling freed up. I don't need a patch with 31 "build warning fixes"
removing and changing all sorts of lines just because clang said it's
okay. I just don't have the time right now to wade through all these
patches, triple-checking correctness and looking for real issues.
What I do need right now is patches that are carefully thought about,
and that are for concrete problems. This close to release, every
change needs to be done considerately.
-Kevin
--
Ticket URL: <https://dev.mutt.org/trac/ticket/3824#comment:4>
Mutt <http://www.mutt.org/>
The Mutt mail user agent
