chewi       15/02/27 22:38:50

  Added:                java-project-meeting-log-20150227.txt
  Log:
  First Java project meeting in years!

Revision  Changes    Path
1.1                  
xml/htdocs/proj/en/java/meeting-logs/java-project-meeting-log-20150227.txt

file : 
http://sources.gentoo.org/viewvc.cgi/gentoo/xml/htdocs/proj/en/java/meeting-logs/java-project-meeting-log-20150227.txt?rev=1.1&view=markup
plain: 
http://sources.gentoo.org/viewvc.cgi/gentoo/xml/htdocs/proj/en/java/meeting-logs/java-project-meeting-log-20150227.txt?rev=1.1&content-type=text/plain

Index: java-project-meeting-log-20150227.txt
===================================================================
--- Log opened Fri Feb 27 20:54:52 2015
21:00 <@fordfrog> ok, so it seems the meeting has started
21:00 <@ercpe> what do you think about putting a summary into the gentoo wiki?
21:00 <@fordfrog> it was discussed above
21:00 <@ercpe> whoops
21:00 <@fordfrog> :-)
21:00 <@ercpe> what was the result?
21:00 <+_flow_> who chairs (if any)?
21:01 <@fordfrog> ercpe, Chewi will log the session
21:01 <@ercpe> the... erm... project lead
21:01 <@ercpe> :P
21:01  * Chewi hides
21:01 <@fordfrog> well, i doubt we have here anyone capable of leading the 
project atm so i suggest we are all equal for now ;-) any objections? :-)
21:02 <@ercpe> i would say we just get through the points. Any time-limit (per 
topic and in general)?
21:02 <+monsieurp> half an hour per topic
21:03 <+monsieurp> no ok joking
21:03 <+monsieurp> 5 min max
21:03 <@fordfrog> lets keep it under 60 minutes, that is till 22:00 utc (23:00 
monsieurp time :-P )
21:03 <@ercpe> ok
21:03 <@Chewi> it's just a collection of the stuff we've been moaning about for 
the past few months so we should be familiar with much of it anyway
21:03 <@ercpe> first topic:
21:03 <@ercpe> Bugzilla is littered with old bugs (2006 onwards), 1/4 of 
packages out of date, how do we deal with them? (monsieurp)
21:03 <@fordfrog> any suggestions?
21:03 <@Chewi> monsieurp seems to want to kill them with fire ;)
21:04 <@ercpe> imho we should drop as much as possible to reduce the overall 
count in dev-java/* to a more maintainable set
21:04 <@Chewi> I think we should hold back a little. there's no prizes for 
blindly closing bugs.
21:04 <+_flow_> Is there any need for action?
21:04 <@fordfrog> imo we should close only those that are really outdated
21:04 <@Chewi> bugs requesting new packages for clearly dead projects can 
obviously go
21:05 <@fordfrog> that is the package or version is not in the tree anymore
21:05 <+monsieurp> guys take a look at euscan as well: 
http://euscan.gentooexperimental.org/herds/java/
21:05 <+monsieurp> many many packages are out of date
21:05 <+monsieurp> how do we deal with these?
21:05 <@ercpe> i think that are two distinct points: load of bugs and outdated 
packages
21:05 <@Chewi> bugs about packages that are in the tree but otherwise dead 
should be considered to go but rdeps need to be dealt with of course
21:05 <+monsieurp> you think very well ercpe I must say
21:06 <@ercpe> bugs: close which are obsolete or cant reproduce
21:06 <+monsieurp> +1
21:06 <+monsieurp> or HOMEPAGE is dead
21:06 <@fordfrog> that makes sense
21:06 <+monsieurp> if homepage dead -> close, OBSOLETE -> next
21:06 <@ercpe> we have a lot of packages with broken HOMEPAGE on dev.java.net 
but still used everywhere
21:07 <@Chewi> I don't want a mass cull of packages yet because I'm hoping that 
the work I will do eclasses/infrastructure will enable us to bump/add new 
package versions more quickly
21:08 <@ercpe> what about removing outdated packages, which aren't part of a 
dep tree?
21:08 <@Chewi> a lot of packages will be effectively rewritten but removing 
packages is more involved than just bumping them
21:08 <@Chewi> I mean... if you're going to readd them later
21:09 <@fordfrog> only dead upstream packages should be removed, if nothing 
depends on them
21:09 <@Chewi> agreed
21:09 <+monsieurp> sure
21:09 <+monsieurp> what about those laying around in the overlays?
21:09 <@ercpe> thats another topic
21:09 <+monsieurp> we will discuss the overlay topic a bit later on but there's 
many cruft in there as well
21:10 <@Chewi> it may not be that many but as I said, this isn't what's holding 
us back. it just looks a little messy.
21:10 <@ercpe> ok, so we trying to bump as much as possible and dropping 
obsolete, outdated and/or abandoned packages
21:10 <@ercpe> ?
21:10 <+monsieurp> sounds good to me
21:10 <+zxiiro> +1
21:10 <@Chewi> yes. but give me time. ;)
21:10 <@fordfrog> agreed, in sane manner ofc :-)
21:11 <@Chewi> I'm not going to write the next Maven eclass overnight :P
21:11 <@ercpe> :D
21:11 <@fordfrog> Chewi, why not? :-(
21:11 <@ercpe> anything else on this topic?
21:11 <+monsieurp> nop let's move on
21:11 <+zxiiro> I'll try to help go through the open bugs too. (just need to 
find some free time ;)
21:11 <+monsieurp> new recruits in sight
21:11 <@ercpe> next: new java devs/recruits: WE NEED MORE PEOPLE TO GET ON 
BOARD! (monsieurp)
21:11 <+zxiiro> regarding version bumping
21:11 <@fordfrog> zxiiro, cool! :-)
21:12 <@fordfrog> any ideas wrt that recruting topic?
21:12 <+monsieurp> first of all: zxiiro, gnu_andrew, do you guys want to become 
dev at some point?
21:12 <@Chewi> well you may have noticed the late edit to the list of potential 
recruits ;)
21:12 <@ercpe> [email protected] received a few mails over the last few month 
regarding helping with java on gentoo
21:12 <@fordfrog> Chewi, yes, we all welcome you hare! :-)
21:13 <+zxiiro> monsieurp: i sure do
21:13 <@ercpe> getting new people on board depends on the documentation > wiki 
migration
21:13 <@ercpe> the docs on g.org are really, really outdated
21:13 <@Chewi> I've mentioned this before but I worry that if we rush to get 
people on board who aren't familiar with the wider picture, they may get 
frustrated because they can't package most things now due to Maven.
21:14 <@Chewi> but I don't want to hold things back in that regard either
21:14 <+monsieurp> ercpe: we'll get to this ;)
21:14 <@Chewi> ercpe: yeah and that too
21:14 <@fordfrog> well, generally, these are my suggestions: first we should 
express our need for new fresh blood at wiki, second, we should ecnourage ppl 
to step up, and third, we should ask recruiters what they could do for us
21:14 <+monsieurp> I've also talked to gienah a bit this afternoon and since 
he's taken care of writing most of the dev-lang/scala ebuilds, he knows his 
around around the java eclasses
21:14 <+monsieurp> he said he's willing to give us a hand
21:15 <+monsieurp> he's in this channel (hi gienah ! :))
21:15 <+zxiiro> I think making it easier for people to actually contribute 
would be helpful
21:15 <+monsieurp> *his way around
21:15 -!- mode/#gentoo-java [+v gienah] by ChanServ
21:15 <@Chewi> does anyone echo my sentiment? has anyone other than me actually 
tried to package anything new lately? :P
21:15 <+monsieurp> he's already a Gentoo dev btw
21:15 <+zxiiro> for example submitting patches to bugzilla is very outdated in 
my opinion
21:16 <+monsieurp> Chewi: I've packaged dev-java/scala-bin and it was a bit 
frustrating
21:16 <@Chewi> zxiiro: we'll come to that later on the github topic
21:16 <@ercpe> but patches in bugzilla are more than nothing
21:16 <+zxiiro> with things like GitHub and Gerrit, contributing to an OS 
project is so easy but Gentoo's infra's holding us back 
21:16 <+zxiiro> ok
21:16 <@Chewi> monsieurp: and that was only a -bin!
21:16 <+monsieurp> Chewi: yeah
21:16 <+monsieurp> Chewi: take a look at dev-lang/scala ..
21:16 <+gnu_andrew> Chewi: well for my part, the only packages I'm interested 
in are JDKs ;)
21:17 <+zxiiro> sorry missed that topic item :)
21:17 <+monsieurp> gnu_andrew: you haven't answered my question about whether 
you wanna become a dev 
21:17 <+monsieurp> zxiiro: ^ :p
21:17 <@Chewi> gnu_andrew: fair enough. and you do a very good job!
21:17 <+monsieurp> +1 definitely thanks for your help andrew
21:18 <@ercpe> anything else we can do on this topic?
21:18 <@fordfrog> well, could someone make a summary of the resolution of this 
topic?
21:18 <@Chewi> probably not, we all agree we need new blood but maybe not the 
highest priority
21:18 <+gnu_andrew> monsieurp: it would make things easier. The main 
showstopper last time was that some Java quiz was wanted. I'm not really 
interested in packaging Java stuff.
21:19 <@ercpe> The summary is basically: "well, generally, these are my 
suggestions: first we should express our need for new fresh blood at wiki, 
second, we should ecnourage ppl to step up, and third, we should ask recruiters 
what they could do for us"
21:19 <+gnu_andrew> it has become a little silly having to point people at an 
overlay for over a month...
21:19 <@ercpe> :)
21:19 <@fordfrog> ah, ok, someone already made it :-P
21:19 <@fordfrog> so lets move on
21:19 <@ercpe> possible new recruits in sight: gnu_andrew, zxiiro (monsieurp)
21:19 <@ercpe> can we skip this?
21:19 <@Chewi> yeah
21:19 <+monsieurp> yes
21:19 <@ercpe> ok
21:19 <@ercpe> 1.8 is out. many stuff to update/version bump/get rid of 
(eclasses, packages, etc). (monsieurp)
21:20 <@ercpe> we dont even have a stable 1.7 :/
21:20 <@fordfrog> monsieurp, can you be more specific?
21:20 <@Chewi> stable 1.7 is further down the list
21:20 <+monsieurp> well basically, how do we deal with stuff 1.8 related?
21:20 <@Chewi> I get the impression that nothing much depends on 8 yet
21:21 <@ercpe> tc8 iirc
21:21 <@Chewi> so even if we bump all our other stuff, it's not essential
21:21 <@Chewi> but no doubt users will ask for it
21:21 <+monsieurp> do we say "no" or do we say "not for now, wait a bit" ?
21:21 <@ercpe> wait a bit
21:21 <+monsieurp> ok cool
21:22 <@Chewi> someone needs to investigate what issues may arise. 7 didn't 
present many new ones but it's not something we should blindly walk into.
21:22 <@ercpe> if all goes well, we will have a stable 1.7 at some point
21:22 <+monsieurp> sounds reasonable
21:22 <@fordfrog> well, 1.7 is blocked just by few bugs now
21:22 <@fordfrog> so we should focus on those bugs that block it
21:22 <@fordfrog> as java is the core of we do, we should keep up to date with 
that imo
21:22 <@ercpe> i would like to put the 1.8 topic way after "bumping outdated 
packages" and stable 1.7
21:22 <@Chewi> gnu_andrew: is there anything you can add about 8?
21:23 <@fordfrog> ercpe, bumping outdated packages is never ending story :-)
21:23 <+zxiiro> +1 focus on getting 1.7 stable and then do 1.8 after. For what 
it's worth though my company does do some of our builds with 1.8 as we are 
adding support to some of our code for it
21:23 <@Chewi> zxiiro: that sounds more like making it work with 8 rather than 
using 8's new features?
21:24 <+zxiiro> Chewi: yes, step one is to get stuff to compile with 8
21:24 <@fordfrog> exactly, it can be 1.4 source and target, but it should 
compile with 8
21:24 <@Chewi> gnu_andrew: any icedtea 8 ebuilds on the horizon, for example?
21:25 <+gnu_andrew> Chewi: there's one in overlay. I'm actually working on 
updating it at the moment.
21:25 <+gnu_andrew> Hope to finally have a 3.0.0 next month
21:25 <+monsieurp> nice
21:25 <@ercpe> great
21:25 <@fordfrog> that sounds good :-)
21:25 <+monsieurp> looking forward to it
21:25 <+gnu_andrew> it'll be u40 based so following that upstream release
21:25 <@Chewi> gnu_andrew: must have missed that
21:26 <+gnu_andrew> Chewi: it's not happened yet, it's due next month
21:26 <@Chewi> ok
21:26 <@fordfrog> ok, can we close that 1.6, 1.7 and 1.8 topic?
21:26 <+gnu_andrew> bump I'm about to do will give us 3.0.0pre03 using u40b21.
21:26 <+monsieurp> sure let's move on
21:26 <@ercpe> fordfrog: i think so
21:26 <+monsieurp> OVERLAYS!
21:26 <@ercpe> too many overlays: java-experimental, java-overlay? we must 
define a workflow if we want to keep (all of) them or nuke them (monsieurp) and 
document the rest on the wiki (fordfrog)
21:26 <+gnu_andrew> on 1.6
21:26 <@Chewi> we should talk about 1.6 I think
21:26 <+gnu_andrew> what was the deal with getting rid of it?
21:26 <@Chewi> yeah
21:27 <@fordfrog> 1.6 should go when done with it
21:27 <@ercpe> we need a stable 1.7 to get rid of 1.6
21:27 <@Chewi> especially in the light of security vulns
21:27 <@fordfrog> yes, the bug takes care of all the deps
21:27 <@Chewi> I guess that's the only blocker really
21:27 <@ercpe> and we have to ensure that EVERY pacakge builds with 1.7
21:27 <+gnu_andrew> it's only Oracle's that's obsolete. OpenJDK/IcedTea is 
still supported.
21:27 <@Chewi> good to know
21:27 <@fordfrog> ercpe, those are the blockers for 1.7 stabilization
21:27 <+gnu_andrew> you could drop Oracle's insecure binaries now if they are 
still present and point people at IcedTea 1.13.6
21:28 <@ercpe> fordfrog: plus the ones that havent discovered yet :)
21:28 <@ercpe> gnu_andrew: we don't have oracle-jdk in 1.6
21:28 <@fordfrog> ercpe, true :-) but those should be reported when occur and 
we should deal with them :-)
21:28 <@ercpe> yep
21:28 <@Chewi> I would take a look at those blockers but I probably have enough 
to do. can I delegate to fordfrog? ;)
21:28 <+gnu_andrew> ercpe: oh ok so what 1.6 are you referring to? The bug 
linked to was about oracle-jdk 1.6
21:29 <@fordfrog> Chewi, i can't promise anything, i already cleaned some but 
do not have much time atm, to much real life stuff here
21:29 <@Chewi> fordfrog: understood
21:29 <@ercpe> gnu_andrew: i havent started the discussion about removal of 1.6 
:)
21:30 <+gnu_andrew> ercpe: oh ok I was confused by what fordfrog said
21:30 <@ercpe> bug #483018 is about oracle-jdk 1.6
21:30 < willikins> ercpe: https://bugs.gentoo.org/483018 "[Tracker] removal of 
1.6 JDK's"; Gentoo Linux, Java; CONF; tomwij:java
21:30 <+gnu_andrew> <fordfrog> ok, can we close that 1.6, 1.7 and 1.8 topic?
21:30 <@fordfrog> gnu_andrew, the bug mentioned above
21:30 <@ercpe> are there any problems with keeping icedtea 1.6 around for some 
time?
21:30 <+gnu_andrew> ercpe: Oracle 1.6 JDK should go immediately. It has 
hundreds of security vulnerabilities.
21:30 <@fordfrog> ercpe, we do not have to remove it if it's supported by 
upstream
21:31 <@ercpe> [I] dev-java/oracle-jdk-bin
21:31 <@ercpe>      Available versions:  
21:31 <@ercpe>      (1.7)  ~*1.7.0.60^fs 1.7.0.76^fs
21:31 <@ercpe>      (1.8)  ~1.8.0.31^fs
21:31 <+gnu_andrew> ercpe: not while it's still supported
21:31 <@ercpe> we dont have oracle-jdk in 1.6
21:31 <@Chewi> removing it just means we don't need to care about 1.6 at all 
any more which would be nice
21:31 <@fordfrog> we have sun-jdk
21:31 <@fordfrog> which is already masked
21:31 <+gnu_andrew> yeah that's what it was called
21:31 <@ercpe> holy sh*t
21:31 <@Chewi> yeah there never was oracle-jdk-1.6 lol
21:32 <@ercpe> didn't know that beast is still alive
21:32 <+gnu_andrew> yeah RH had the same thing I think, java-1.6.0-sun
21:32 <@ercpe> well, more half-dead
21:32 <@Chewi> okay, that's really got to go
21:32 <+gnu_andrew> ercpe: it shouldn't be. They stopped public sec. updates 
years ago
21:32 <@Chewi> who wants the pleasure? ;)
21:32 <+gnu_andrew> I do think weening packages away from 1.6 is a good idea
21:32 <+monsieurp> how difficult is it?
21:33 <@Chewi> well it just needs the usual last-rites process I guess
21:34 <@Chewi> check for anything that insists on sun-jdk, there should be few 
if any
21:34 <+monsieurp> ok I'll take care of it
21:34 <@ercpe> great
21:34 <@ercpe> can we move to the overlays topic?
21:34 <+monsieurp> yes
21:34 <@ercpe> too many overlays: java-experimental, java-overlay? we must 
define a workflow if we want to keep (all of) them or nuke them (monsieurp) and 
document the rest on the wiki (fordfrog)
21:34 <@fordfrog> so we keep icedtea and virtual 1.6 for now, right?
21:34 <@ercpe> fordfrog: yes
21:34 <@Chewi> monsieurp: I don't think I want that responsibility so soon in 
my tenure so thanks :)
21:34 <@fordfrog> so the related bug should be handled
21:34 <@Chewi> fordfrog: yes
21:35 <@fordfrog> as it's obsolete now :-)
21:35 <+gnu_andrew> Chewi, monsieurp: well no-one should be using sun-jdk-1.6 
and there's a very strong reason to get rid of it.
21:35 <@Chewi> I'm happy to keep overlays as they are
21:35 <@Chewi> I think they have a well defined purpose
21:35 <+gnu_andrew> I do need java overlay for now as I've nowhere else to 
commit...
21:35 <@Chewi> I still intend to use java-overlay as a staging ground
21:35 <+monsieurp> gnu_andrew and fordfrog: let's discuss this topic offline
21:35  * ercpe has never used the overlays
21:36 <@Chewi> there seems to be some confusion about java-experimental but 
it's always been clear to me
21:36 <@ercpe> i have no use for them
21:36 <@Chewi> half-baked work that's not good enough for overlay but someone 
else may be able to fix later
21:36 <+monsieurp> well, my pet peeve as far as overlay are concerned is
21:36 <@Chewi> I have used it in the past
21:36 <@ercpe> every now and than a bugs pops up at b.g.o
21:36 <+monsieurp> it's extremely difficult for newcomers to contribute to 
Gentoo
21:37 <@ercpe> since we dont have the manpower to cover so many places:  i 
would vote for removal
21:37 <+monsieurp> I'd move some stuff to Github so that we get some exposure 
and maybe get people to contribute 
21:37 <+_flow_> I would say it's extremely difficult. But yes it took me nearly 
half a year till someone allowed me to commit my ebuilds
21:37 <@ercpe> monsieurp: thats the next topic
21:37 <@Chewi> monsieurp: they can be mirrored on github, but not actually moved
21:37 <+_flow_> *wouldn't
21:37 <+monsieurp> Chewi: ok sure whatever
21:37 <@Chewi> _flow_: I'm sorry it took so long
21:38 <+_flow_> but it was not a technical problem, I guess it was more a 
manpower problem
21:38 <@Chewi> I think now there's renewed interest, we will be more receptive 
to newcomes
21:38 <@ercpe> so, do we actually NEED those overlays?
21:38 <+gnu_andrew> ercpe: yes
21:38 <@ercpe> meh
21:38 <+monsieurp> the Gentoo KDE team has some of their overlays on github and 
it's very much active
21:38 <+_flow_> well the java overlay is clearly needed
21:38 <+monsieurp> just see for yourself: https://github.com/gentoo/kde
21:38 <+zxiiro> I'd be +1 on making our development happen on GitHub, let 
people submit PR's instead of attaching them to bugs. easier to review and 
easier to merge
21:38 <@Chewi> the overlays allow people like _flow_ (and gnu_andrew, and me!) 
to contribute *more8 easily
21:38 <+_flow_> not sure about java-experimental, it just adds more complexity 
it appears
21:38 <+gnu_andrew> Chewi: to contribute at all...
21:38 <+zxiiro> easier for drive-by contributors too
21:39 <+zxiiro> sometimes someone just wants to contribute a quick patch
21:39 <+monsieurp> look at the number of commits..
21:39 <+gnu_andrew> Chewi, ercpe: really it's dependent on first solving the 
whole recruitment issue
21:39 <@ercpe> wait, guys. The topic is about the number of overays. Where they 
live is the nexxt topic
21:39 <@Chewi> if you don't like experimental, just ignore it ;)
21:39 <@Chewi> it's not intended for end users
21:39 <@Chewi> it's not in layman
21:39 <+monsieurp> if we want to revamp Java and get fresh blod and etc., 
github *is* definitely the place for showing our work to the world
21:39 <+gnu_andrew> if you don't have sufficient 'official' overlays, people 
are just going to create their own ad-hoc ones
21:39 <+monsieurp> *blood
21:40 <+gnu_andrew> that's what I did before I had java overlay access
21:40 <@ercpe> The topic is: "too many overlays: java-experimental, 
java-overlay? we must define a workflow if we want to keep (all of) them or 
nuke them (monsieurp) and document the rest on the wiki (fordfrog)"
21:40 <@ercpe> NOT the github topic
21:40 <@ercpe> so: do we keep them?
21:40 <+zxiiro> I'd vote to remove them, if we can keep as close to 1 overlay 
as possible, it makes it less confusing
21:41 <+monsieurp> java-experimental is useless IMHO
21:41 <@Chewi> rename java-experimental to chewi's-craphole if you like ;)
21:41 <+monsieurp> lol
21:41 <+zxiiro> too many repos means people have to figure out where to put 
their contribution
21:41 <+monsieurp> +1
21:41 <@ercpe> i agree with zxiiro 
21:41 <@Chewi> until recently, java-experimental wasn't even publicly visible 
so I'm not sure why you hate it so much :P
21:41 <@fordfrog> well, instead of java-experimental we could create new 
overlay at github for those that are not yet granted access to java-overlay
21:41 <+monsieurp> and +1 for renaming java-experimental to chewi's craphole 
ahah :> (it did crack me up)
21:42 <@Chewi> fordfrog: I don't think that's the way to go
21:42 <@Chewi> github should have a mirror of java-overlay
21:42 <@ercpe> Ok, java-experimental will be removed, java-overlay stays
21:42 <+_flow_> fordfrog: why not mirror? That is, if you really want github
21:42 <@Chewi> that's more effective
21:42 <+monsieurp> +1
21:43 <@ercpe> next topic?
21:43 <@Chewi> wait a sec
21:43 <@Chewi> let's see exactly what's in there
21:43 <@Chewi> to see if it's really worth keeping
21:43 <@Chewi> 351 ebuilds
21:43 <+gnu_andrew> I haven't used experimental for years.
21:44 <@Chewi> 332 packages
21:44 <@Chewi> that's a lot of stuff to just throw away
21:44 <+monsieurp> the thing that bothered me a lot the first time I cloned 
java-experimental was the README content
21:44 <@Chewi> seems clear to me!
21:44 <+monsieurp> basically it says "well.. it might work but don't expect it 
to"
21:45 <@ercpe> for the lazy: that is in java-overlay: 
https://paste.ercpe.de/java-overlay.txt
21:45 <+zxiiro> ercpe: thanks
21:45 <@Chewi> monsieurp: right! that's the point lol
21:45 <+zxiiro> can't we just hardmask "experimental" stuff?
21:46 <@ercpe> some of the libs should be moved to gentoo-x86 because they are 
dependencies
21:46 <+zxiiro> why need a separate overlay
21:46 <@ercpe> e.g. wagon, plexus, etc.
21:46 <@Chewi> I think that would be worse
21:46 <@Chewi> look I don't want to be a stick in the mud so
21:46 <@Chewi> I'm happy to keep a local copy
21:47 <@Chewi> most of it is popular way out of date and needs Mavenising anyway
21:47 <@Chewi> *probably
21:47 <+monsieurp> ok sure 
21:47 <+monsieurp> we can keep it for a while if you want
21:47 <+monsieurp> *BUT*
21:47 <@Chewi> who needs to throw that switch? infra?
21:47 <+monsieurp> yes
21:48 <@Chewi> who's going to tell them? ercpe seems to hate it the most. :P
21:48 <+monsieurp> *BUT* ! we should definitely have a java-experimental mirror 
on github
21:48 <@Chewi> huh?
21:48 <@ercpe> what?
21:48 <+monsieurp> sorry
21:48 <+monsieurp> java-overlay I meant
21:48 <@Chewi> yes
21:48 <+monsieurp> typo :\
21:48 <@Chewi> I think we're all in agreement there
21:48 <+monsieurp> okay
21:48 <@fordfrog> yes
21:49 <+zxiiro> I feel like it shoud,n't be a mirror, it should be primary
21:49 <@Chewi> zxiiro: I'm not sure that would fly with Gentoo
21:49 <@ercpe> OK, to sum this up: dropping experimental, keeping overlay?
21:49 <@Chewi> ercpe: yes
21:49 <+monsieurp> I talked to infra about it already
21:49 <@fordfrog> and mirroring java-overlay at github
21:49 <+monsieurp> zxiiro: and their answer was "do not rely too much on github"
21:49 <+_flow_> zxiiro: you definetly not want to depend your infra on some 
company
21:49 <@Chewi> sorry to keep delegating but we need to make decisions about 
who's doing stuff or it won't get done :P
21:50 <+zxiiro> right, but from what i've seen so far, can we depend on Gentoo 
infra?
21:50 <@ercpe> move overlay repositories to github under 
https://github.com/gentoo/java. benefits: more exposure, get more people to 
contribute, github infrastructure. (monsieurp)
21:50 <+zxiiro> i mean a few servers have been offline for months
21:50 <@Chewi> zxiiro: they seem to be improving again. they just brought 
archives.gentoo.org back from the dead.
21:50 <+zxiiro> ok that's good
21:50 <+zxiiro> glad to hear that
21:50 <@ercpe> while i really like github and totally see the advantages, i 
would definetly go with gentoo's infra
21:51 <@Chewi> anyway, the concept of a "primary" isn't so solid with git
21:51 <@fordfrog> ercpe, the mirror would be fine i guess
21:51 <+zxiiro> My favourite option would be Gerrit
21:51 <+zxiiro> which we can host ourselves
21:51 <@ercpe> we use git, right? Anyone can clone a git repo from gentoos 
infra and send a pull request
21:51 <+zxiiro> but you'd need infra on board with that
21:51 <@ercpe> zxiiro: they will if the primary source is on gentoos infra
21:51 <@fordfrog> ercpe, github is more user friendly and devs are often used 
to it
21:52 <+zxiiro> ercpe: yes, but merging is a pain if the place you receive PR's 
from is not the primary
21:52 <+monsieurp> here's what infra said the other day when I asked in the 
channel about moving the overlay to github
21:52 <+monsieurp> 15:21 [  monsieurp ] a simple question, out of curiosity: we 
(Java team) might move the overlays to Github. 1) Can we do it? 2) if yes, how 
can we go about doing it?
21:52 <+monsieurp> 15:26 [  Pinkbyte ] monsieurp, if you are member of github 
organization, you can just create repo, add new origin and push into it
21:52 <+monsieurp> 15:26 [  Pinkbyte ] however, as per our last discussion in 
gentoo-dev, please, do not rely on github too much
21:52 <+monsieurp> 15:27 [  Pinkbyte ] i mean - nobody can stop you to get pull 
requests from it, and it's ok. But making it primary point of contrib
21:52 <+monsieurp> ution, without mirror on our infrastructure is a bit overkill
21:52 <@ercpe> the last one is exactly my point
21:52 <+zxiiro> plus if Devs don't check the mirror then they won't know 
there's PRs to merge
21:53 <@ercpe> plus i dont want to heat up the discussion on g-dev  :P
21:53 <@Chewi> zxiiro: I know you can't just click the Merge button but I don't 
think it's that big a deal. I'm very comfortable with git on the command lin.e
21:53 <@Chewi> zxiiro: I'm quite active on github so don't worry
21:53 <+zxiiro> Chewi: yes, but I like to think of future. In this case what if 
you leave and no one else checks GitHub?
21:54 <+monsieurp> zxiiro: then it'll be you
21:54 <@Chewi> if I leave, we're screwed anyway :P
21:54 <+zxiiro> that's why I feel primary being the place you accept 
contributions is important
21:54 <@fordfrog> zxiiro, then we will delete the mirror :-P
21:54 <+zxiiro> lol
21:54 <@ercpe> We are talking about an overlay, right? i don't think that we 
should work that mouch in the overlay. We have enough work down the road in 
gentoo-x86
21:54 <@Chewi> true
21:54 <@fordfrog> ercpe, that should be starting place for newcomers
21:54 <+zxiiro> anyway, regardless mirror or not, i think it's good progress in 
the right direction
21:54 <@fordfrog> we should help them, support them and give them toys to play 
with :-)
21:55 <@Chewi> monsieurp: do you know who can give access to the organisation?
21:55 <@ercpe> open a bug
21:55 <@Chewi> ok
21:55 <+monsieurp> a3li
21:55 <+monsieurp> no ask him directly
21:55 <@ercpe> thats the way i got my permissions
21:55 <+monsieurp> yeah well
21:55 <+monsieurp> opening a bug works too
21:56 <@Chewi> moving on now, I think
21:56 <@ercpe> so what is the result of this topic?
21:56 <+monsieurp> yes
21:56 <@ercpe> gh as a primary source?
21:56 <@Chewi> mirror
21:56 <+monsieurp> mirror
21:56 <+monsieurp> [22:49:23] [ @ercpe ] OK, to sum this up: dropping 
experimental, keeping overlay?
21:56 <+monsieurp> + mirror on gh
21:56 <@ercpe> that was the previous topic :)
21:56 <+monsieurp> ah shit
21:56 <@ercpe> ok
21:57 <+monsieurp> but yeah :> mirror
21:57 <@Chewi> not sure exactly how the mirroring works, maybe the repo is 
effective read-only? we'll see I guess.
21:57 <@ercpe> we have almost hit the 60 minute limit
21:57 <+monsieurp> it's ok
21:57 <@ercpe> i have to work tomorrow so i would like to stop here
21:57 <+monsieurp> :(
21:58 <@Chewi> just before all my stuff, great :P
21:58 <@fordfrog> ok, so we're done for now, the rest text time
21:58 <@fordfrog> s/text/next
21:58 <@Chewi> well I could ramble on for a while so I'd rather have your fresh 
attention some other time
21:58 <@fordfrog> just one last question, how often do we/are we able to meet 
here?
21:58 <@ercpe> 50% of all topics isn't that bad :)
21:58 <+monsieurp> twice a week 
21:58 <@Chewi> normally I'd say we should do this once a month
21:59 <+monsieurp> Chewi: everyday
21:59 <@Chewi> but obviously we have a lot to talk about right now
21:59 <@ercpe> yeah, every two or four weeks
21:59 <@fordfrog> ok
21:59 <+monsieurp> every two weeks
21:59 <@fordfrog> so at least each month and at most bi-weekly
21:59 <@Chewi> how about we pencil in next Friday to deal with the rest of 
these topics
21:59 <@ercpe> fordfrog: thats a good idea
21:59 <@Chewi> and then take it a bit more steady from there
22:00 <@ercpe> deal!
22:00 <+monsieurp> deal
22:00 <+monsieurp> Friday is good
22:00 <@Chewi> cool
22:00 <+monsieurp> zxiiro: ?
22:00 <+monsieurp> gnu_andrew: ^
22:00 -!- fordfrog changed the topic of #gentoo-java to: Java on Gentoo 
http://java.gentoo.org/ | Other Java stuff please take to ##java | Ask 
question(s) and please be patient 
http://www.gentoo.org/proj/en/java/java-devel.xml | Open Bugs: 
http://tinyurl.com/gentoojava | Ideas/Topics for next meeting at 
https://wiki.gentoo.org/wiki/Project:Java#Topics (even non-members can and are 
encouraged) | Next meeting Mar 06 @ 21:00 UTC
22:00 <@ercpe> Next meeting: Friday, Mar 6, 21:00 UTC
22:00 <+monsieurp> UTC, gotcha
22:00 <@ercpe> the important part: UTC
22:00 <@ercpe> :P
22:00 <@Chewi> fordfrog: you read my mind :D
22:00 <+monsieurp> AHAH :P
22:00 <@fordfrog> :-)
22:00 <+gnu_andrew> most Fridays work for me, but not next I'm afraid
22:01 <+monsieurp> :( so it doesn't actually work then?
22:01 <@Chewi> gnu_andrew: it'll mostly be able Maven anyway
22:01 <@Chewi> *about
22:01 <+zxiiro> monsieurp: Friday's good for me too
22:01 <+gnu_andrew> Chewi: oh that's ok, the less I hear about that thing the 
better ;)
22:01 <+zxiiro> all my meetings areo n Thursdays
22:01 <@Chewi> haha
22:01 <+zxiiro> so i'm free fridays :)
22:01 <+monsieurp> gnu_andrew: RH policies, eh..? a Gentoo meeting a month
22:02 <+gnu_andrew> monsieurp: huh?
22:02 <+monsieurp> :P
22:02 <+monsieurp> joking :>
22:02 <+gnu_andrew> monsieurp: I don't even like Ant, never mind Maven. Hence 
why I want to stay away from Java packaging ;)
22:03 <@Chewi> right, I'll call a close to this meeting for the log. thanks 
guys. :)
22:03 <+gnu_andrew> they should just use good old autoconf and make like 
everyone else
22:03 <+gnu_andrew> Chewi: thanks
22:03 <+monsieurp> thanks! very productive :)




Reply via email to