We plan to switch to compiling with --std=c99 and targetting all new code
work to c99 best practices. First we need to make sure existing code
compiles with c99.
I came in on the tail end of this part of the conversation.
Yesterday afternoon I spent about half an hour doing a first pass
towards getting it compiling with -std=c99 (it didn't end up being a
trivial set of changes so I backed it out without even a local commit)
Things that become interesting with a strict -std=c99 compile are common
things like strdup() not being strictly part of c99, but rather POSIX or
BSD things (so you need a #define __USE_BSD or whatever somewhere
appropriate). strcasecmp and strncasecmp (both of which we use quite a
bit) need to be included from strings.h rather than string.h (the latter
usually automagically includes the former, but not in strict c99 mode).
Targetting -std=gnu99 was easier, from memory I only had to add
strings.h to the files using str[n]casecmp to get it building
On Tue, Apr 14, 2015, at 10:33 AM, Bron Gondwana wrote:
Present: Bron, Jeroen, Ken, Matt, Simon, Ellie
Anthony has resigned from the Cyrus Foundation due to personality
conflicts and cultural mismatch with the team. I have reverted the
commits that happened during the week that I was was away, and then
cherry-picked the ones that I was satisfied were stable or easily
fixable.
We still plan to go ahead with simplifying the configure script - but
we'll do it as small, testable changes and make sure they keep working
with our target plaforms and don't block everyone else's workflow.
We plan to switch to compiling with --std=c99 and targetting all new code
work to c99 best practices. First we need to make sure existing code
compiles with c99.
Ken: has a bunch of commits almost ready to go, but just needed to track
down one more bug first.
Matt: has been away for a month, moving and had a good holiday. Welcome
back. Is going to work on cleaning up lib/ - working with Jeroen to make
sure it all works with the configure changes.
Ken: libical - almost everything currently used by Cyrus in is libical
git. There's a bug on nonsense recrurrence startdate which causes it to
spin forever - needs to be fixed.
Ellie: new sync code on master. Will back it out from 2.5 for 2.5.1
release - probably tomorrow.
Bron: got it compiling with clang. Bugfixes for some issues at FastMail.
Master is on FastMail testing now, with Ellie's changes. Worked
together to make sync backwards compatible back to 2.4. Planning to ship
master code to FM production tomorrow if issues with dav database can be
tracked down.
We have a potential new programmer who applied to FastMail at just the
right time, and is interested in working on Cyrus. He's a programming
student in France. Bron and Pierre to Skype interview during the week.
Hoping to meet in person while Bron is in Europe in early May.
Cyrus talk at openSUSE/Kolab summit: need to prepare Abstract and Bio -
Hans is on my case!
Due to the daylight savings changes squashing Europe closer to Australia,
the UTC 2200 time doesn't work so well any more, so the next meeting will
be pushed later on Thursday: 2015-04-16T11:00:00Z.
Bron.
--
Bron Gondwana
br...@fastmail.fm
