Dev meeting minutes 2012-01-09

Dev meeting minutes 2012-01-09

Minutes of the meetings between Spring developers are archived here.
Post Reply
User avatar
Spring Developer
Posts: 2299
Joined: 28 Jun 2007, 07:30

Dev meeting minutes 2012-01-09

Post by jK »

Date: 2012-01-09
Present: abma, jK, Tobi, zerver


_The meeting_____________________________________________________
>> jK joined!
>> zerver joined!
>> abma joined!
>> Nightwatch joined!
>> Tobi joined!
<< Nightwatch left!
>> Nightwatch joined!
abma: hey!
Tobi: hi
zerver: hi
Tobi: did you already do meeting?
zerver: no
abma: shall we start? doesn't look like hoijui / kloot will join
zerver: yeah

can luasocket be added to the main-repo as branch ?
abma: anything against adding it as branch to the main repo?
jK: why not
abma: i think it can be used as it is... still has many todo, but i think it can be done that way...
jK: you can always add feature branches
abma: why not.. uhm, because its design is failure or so :)
abma: ok, then... next (other questions i had better fit to WVTTK)
zerver: if the design is fail, then delete branch :P

build slave configuration
abma: [RoX]Tobi: that was your point?
Tobi: ya
Tobi: at first I thought it would be useful to build oldest + newest boost version in some slave
Tobi: to ensure our minimum requirment (i.e., that in cmake) actually works
Tobi: and that it still works with latest boost
Tobi: so now koshi built all boost versions from 1.37 to 1.48
abma: newest one imo would be really helpfull, as at some time this will be used by many distributions
Tobi: we don't have enough disk/time to build all
Tobi: so what kind of config would be preferred?
zerver: do we intend to support multiple boost versions?
abma: we already support multiple boost versions, but we didn't know about that :)
zerver: yes, I know that we do, but is it desirable ? :P
Tobi: I think that happens nearly automagically, as I assume boost is a bit stable :)
abma: maybe oldest + newest one are enough?
Tobi: supporting only single boost version is painful for packagers etc. if distro doesn't have that boost yet
Tobi: yes that is what I thought too
zerver: okay
abma: i guess the versions between are mostly used by us devs
Tobi: makes most sense also to look at what popular distros do I suppose
Tobi: supporting older than debian stable is pointless
abma: yep
zerver: one other alternative is to not upgrade boost very often, and when we do, never upgrade to latest
abma: spring currently doesn't build with 1.48 and i think at some time we will use that version
abma: linux builds use the system installed boost
jK: nice gentoo has now 1.48 in the portage (still masked=unstable)
abma: that was maybe the initiation of the idea:
Tobi: ok, so I'll make it build min and max
Tobi: I'm still looking what debian stable has now
Tobi: ah, 1.42
Tobi: so then, I reconfigure buildslave to make 1.42 and 1.48 builds, and we bump minimum to 1.42?
zerver: sure
abma: ubuntu 10.04 has 1.38 / 1.40
abma: hm, desktop support for ubuntu 10.4 ends at april 2013
abma: maybe we should try if it builds, if not increase version ;-)
zerver: :)
abma: minimum boost version i found in config is 1.34.0 in ./tools/ArchiveMover/CMakeLists.txt
abma: next is 1.35.0 in ./CMakeLists.txt
abma: and i think building it at a daily base is fine...
abma: can't say, the buildslave owner has to decide i think

Release plan
jK: hotkeys still need fix
zerver: okay
abma: flushing logfile after a crash seems to not work well, too...
abma: i'll set target version for mantis reports... this maybe helps a bit
zerver: I'll try fix the hotkeys then
jK: you need to add a return param and fix like 300 functions
abma: mhm damn, this updated the date :-/
zerver: 300 functions :O
zerver: if you already started investigating it, then me doing it is a really bad idead
zerver: -d
jK: might be exaggerated, still a lot ^^
jK: yeah
jK: but today I am too tired
jK: might be tomorrow or the day after
zerver: i'll check the flush infolog fail instead
abma: i try to merge kloots commit into the bugreports...

Anything else? (WVTTK)
abma: i've some questions related to luasocket... : why doesn't lua_open_lib register "socket" as global?
abma: ... I.cpp#L261
abma: to make it work "socket = require ("socket")" is needed in the lua widget
abma: i guess you jk can help here best (if you didn't felt asleep ;)
abma: ... et.lua#L12
abma: and the widget: ... socket.lua
jK: no idea, would need to check their objects, but I don't have their code on my hdd atm and I am too tired to get it
abma: ok, np
abma: (+ just do a git pull and you've the code ;)
zerver: is infolog flush fail mantised?
abma: yes
abma: maybe this comment is the interesting one:
zerver: thx
abma: np
abma: i investigated it a bit too, but i couldn't reproduce it on linux
abma: /cheat + /crash did always fully ouput the stacktrace
zerver: easy one, flushing is done in the wrong place
<< Tobi left!
abma: gn8!
<< abma left!
Post Reply

Return to “Meeting Minutes”