18:01:25 <sochotni> #startmeeting #startmeeting Java SIG -- https://fedoraproject.org/wiki/SIGs/Java
18:01:25 <zodbot> Meeting started Wed Feb 23 18:01:25 2011 UTC.  The chair is sochotni. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:01:25 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:01:29 <sochotni> #meetingname java-sig
18:01:29 <zodbot> The meeting name has been set to 'java-sig'
18:01:38 <sochotni> #topic roll-call
18:01:43 <pingou> .fas pingou
18:01:43 <zodbot> pingou: pingou 'Pierre-YvesChibon' <pingou@pingoured.fr>
18:01:57 <akurtakov> .fas akurtakov
18:01:57 <zodbot> akurtakov: akurtakov 'Alexander Kurtakov' <akurtako@redhat.com>
18:02:07 <overholt> .fas overholt
18:02:08 <zodbot> overholt: overholt 'Andrew Overholt' <overholt@redhat.com>
18:02:10 <sochotni> .fas sochotni
18:02:13 <zodbot> sochotni: sochotni 'Stanislav Ochotnicky' <sochotni@redhat.com>
18:02:31 <sochotni> never thought of using this before :-)
18:02:34 <orionp> .fas orion
18:02:34 <zodbot> orionp: drchrishobson 'Chrid Hobson' <chris.hobson@orionhealth.com> - orion1979 'Alexey Rubinov' <orion1979@yandex.ru> - orionjob 'Orion McNyset' <orionjob@gmail.com> - orion 'Orion Poplawski' <orion@cora.nwra.com> - macabroso17 '' <orion_17889@yahoo.es>
18:02:57 <sochotni> hehe, long info
18:03:23 <sochotni> #info overholt akurtakov orionp pingou sochotni attending
18:04:16 <sochotni> #topic Changes to guidelines
18:04:44 <sochotni> I'd like to propose tweaking of current guidelines to have maven 3 in templates/recommendations
18:04:53 <sochotni> with maven2 as fallback in case of problems
18:05:08 <sochotni> and those problems should be filed as bugs against maven package
18:05:26 <sochotni> I already found one problem (maven-ant-tasks, yet to investigate)
18:05:42 <akurtakov> we should put this really strong that Maven 2 will not be supported soon so people can feel the need to move to maven 3
18:06:04 <sochotni> I can do the changes to draft guidelines, you can just verify them
18:06:06 <akurtakov> at least I don't plan any work on maven2 unless it causes troubles for maven 3
18:06:30 <akurtakov> sochotni: what about the macro changes?
18:06:39 <sochotni> well upstream doesn't seem eager to work on 2.X anymore either so it's natural..
18:06:47 <akurtakov> it's good to get them in one go through the FPC
18:07:05 <sochotni> akurtakov: talking about %add_maven_depmap ?
18:07:30 <akurtakov> sochotni: yes, tweaking the guidelines every second month is not a good thing either
18:08:01 <sochotni> that is not in jpackage-utils yet, so I don't know...I don't think there is any other way to make it work easily
18:08:21 <akurtakov> overholt: would you please push dbhole to look at jpackage-utils(he is the maintainer)
18:08:47 <overholt> akurtakov: sure
18:08:50 <akurtakov> or he can give up maintainership which will effectively mean that we are forking jpackage-utils
18:09:05 * overholt assumed we had already forked it
18:09:13 <sochotni> it's bug https://bugzilla.redhat.com/show_bug.cgi?id=658809
18:09:40 <akurtakov> overholt: well, we haven't but I think it's time to
18:09:58 <overholt> I'll speak with Deepak about it tomorrow
18:10:20 <sochotni> I am not gonna pretend I understand all of the jpackage history...but it sure would be nice to have ways to simplify work of packages easily
18:10:48 <sochotni> #action overholt discuss with dbhole about bug #658809 and possible "forking" of jpackage-utils
18:11:08 <akurtakov> sochotni: not knowing the history is a good thing because you will have a cleaner look at the situation :)
18:11:53 <sochotni> OK, I guess no one is agains us moving to maven 3...so I'll start working on draft changes
18:12:12 <akurtakov> the sooner maven2 dies the better
18:12:15 <sochotni> overholt: let me know what happens with jpackage and I'll possibly add that to the guidelines
18:12:20 <pingou> hm what's the target release for maven3 ? F15 or F16 ?
18:12:28 <sochotni> pingou: f15
18:12:32 <akurtakov> pingou: we have it in f15
18:12:41 <sochotni> but not as a replacement for maven2
18:12:46 <overholt> sochotni: yup
18:12:55 <sochotni> in f15 maven 3 actually depends on maven2 package
18:13:33 <akurtakov> sochotni: please also take a look at the other scripts in jpackage-utils and incorporating them in the guidelines
18:13:42 <sochotni> #action sochotni add changes to guidelines draft that use maven 3 as default
18:13:52 <akurtakov> e.g. clean-binary-files
18:13:53 <sochotni> akurtakov: that was my intention
18:14:39 <sochotni> #action add misc functions from jpackage into guidelines (clean-binary-files, verify jpackage_script macro etc)
18:14:48 <sochotni> #undo
18:14:48 <zodbot> Removing item from minutes: <MeetBot.items.Action object at 0x2b025e9ff850>
18:14:53 <sochotni> #action sochotni add misc functions from jpackage into guidelines (clean-binary-files, verify jpackage_script macro etc)
18:15:06 <sochotni> love the python output :-)
18:15:30 <sochotni> I'd go to JNI problems now...
18:16:04 <sochotni> the guidelines need to be clarified, but I am not sure anyone of us actually understands ins and outs of 64/32bit Java differences
18:16:37 <sochotni> though bug #665576 provides good information
18:16:54 <sochotni> I was hoping zodbot would help here ...
18:17:04 <sochotni> #link https://bugzilla.redhat.com/show_bug.cgi?id=665576  "build-classpath swt" fails on 64bit
18:17:11 <pingou> bug #665576
18:17:26 <sochotni> bad zodbot, bad
18:17:44 <akurtakov> I still think that using _jnidir is better than _libdir/name
18:18:32 <akurtakov> but we have to define _jnidir to be _libdir/java
18:18:40 <akurtakov> not /usr/lib/jave as it is now
18:20:25 <akurtakov> anyone having smth against it ?
18:20:53 <sochotni> I'd have to think about this properly and do some tests to REALLY understand what these changes mean
18:21:08 <pingou> + makes it a bit more readable - it's one more macro
18:21:42 <pingou> but using libdir would seem logical (although I know nothing about the 32/64 differences in java)
18:22:15 <akurtakov> ville's option B from this bug
18:22:25 <sochotni> we currently install jni libs into %{_libdir}/%{name} which is not java VM specific but host arch specific
18:22:27 <akurtakov> is what I propose
18:23:10 <sochotni> currently _jnidir is always /usr/lib/java right?
18:23:20 <pingou> %_jnidir        %{_prefix}/lib/java
18:23:26 <sochotni> yup
18:23:48 <akurtakov> overholt: opinion ?
18:24:05 <overholt> akurtakov: I think it need to be fixed
18:24:11 <overholt> but I'm not sure what the best solution is
18:24:46 <sochotni> change the macro to be %{_libdir}/java is the first step, but as is said in the bug...this will require rebuild of packages using JNI
18:24:56 <sochotni> plus there are runtime configurations as well..
18:25:15 <akurtakov> and this will match what we do in eclipse :)
18:25:20 <pingou> pity we didn't do it before the rebuild
18:26:00 <akurtakov> sochotni: pingou: it wouldn't require that much rebuilds because most packages using native parts don't use jpackage scripts due to them not working in this case
18:26:03 <sochotni> pingou: too late to cry over spilled milk I guess :-)
18:26:10 <overholt> :)
18:26:16 <akurtakov> but yes we will break a few applications :)
18:26:40 <sochotni> I am willing to do that in rawhide, but it would be good to have "preemptive" list
18:27:01 <pingou> sochotni: I was just thinking loudly, sorry :)
18:27:23 <sochotni> pingou: but you were right!
18:28:23 <sochotni> actually rationale section in our guidelines provide hints to what needs to happen
18:29:26 <sochotni> and it seems to be a bit more complicated..
18:30:24 <akurtakov> well, the best scenario is described there
18:30:30 <akurtakov> we can not get everything at once
18:31:47 <akurtakov> just fixing the location and jpackage scripts will give a lot better handling of the situation even without autofinding jni
18:31:56 <sochotni> I don't think there are that many jni packages so hopefully nothing will break beyond repair
18:32:35 <pingou> (isn't rawhide there to be broken ? :))
18:32:47 <sochotni> So I guess give green light for B) handling of #665576 ?
18:32:53 <akurtakov> +1
18:32:56 <sochotni> and we'll look after possible breakages
18:32:57 <pingou> +1
18:33:02 <sochotni> +1 from me as well
18:33:14 <pingou> overholt: ?
18:33:21 <overholt> sorry, I haven't read it recently but I'm sure you're all right
18:33:38 <akurtakov> :)
18:34:07 <sochotni> #agreed give green light for B) handling of #665576
18:34:07 <akurtakov> sochotni: please announce it on java list for comments first
18:34:33 <akurtakov> though it still depends on the jpackage-utils
18:34:36 <sochotni> #action sochotni announce changes to JNI fixes on java ML
18:34:47 <sochotni> #undo
18:34:47 <zodbot> Removing item from minutes: <MeetBot.items.Action object at 0x2b025e9ff350>
18:34:54 <sochotni> really...depends on dbhole :-)
18:35:15 <akurtakov> moving?
18:35:19 <sochotni> I'll comment on the bug and I'll make sure it gets announced in right time
18:35:24 <sochotni> yu[
18:35:30 <sochotni> yup that is :-)
18:35:36 <sochotni> #topic Tomcat 5->6
18:36:03 <sochotni> 8 packages left
18:36:28 <akurtakov> sochotni: do you have a list?
18:36:33 <sochotni> akurtakov: https://bugzilla.redhat.com/show_bug.cgi?id=640397
18:37:23 <akurtakov> most of them are insignificant
18:37:36 <sochotni> #info tomcat5 packages to go: kawa, castor, jakarta-taglibs-standard, struts, tomcatjss, gdata-java, freemaker
18:37:45 <akurtakov> we(eclipse) still haven't moved too
18:37:46 <sochotni> actually 7 because 8th is tomcat5 itself
18:38:31 <sochotni> orionp: how is your time lately? It's your bug after all :-) (although we all agree tomcat5 needs to go...)
18:40:12 <sochotni> I could look into one package at least I guess..maybe tomcatjss. Most moves are quite simple
18:41:47 <orionp> I really have no time.
18:42:06 <akurtakov> omajid is going to look into freemarker
18:42:14 <sochotni> I guessed so, reviews have stalled too :-)
18:43:00 <sochotni> OK, well...It'd be great to have it in F16 :-)
18:43:49 <sochotni> F15 is frozen anyway so I guess strange things could happen if we moved things too much
18:44:25 <akurtakov> F-15 is done deal we can not drop packages anymore
18:44:31 <akurtakov> so there is no point pushing for it
18:44:38 <sochotni> yup, exactly
18:45:11 <sochotni> Because of this I'll guess there is no reason to start tomcat 7 discussion for now?
18:45:23 <akurtakov> well, we need someone to spend time on it
18:45:28 <akurtakov> and there is no volunteer
18:45:35 <akurtakov> so let's skip
18:45:37 <sochotni> yup, and there are other things...
18:45:59 <sochotni> #info tomcat5->6 migration finalisation will wait until F16
18:46:14 <sochotni> #topic Java SIG tracker bug status
18:46:25 <sochotni> there are mostly package reviews
18:46:57 <sochotni> some of them stalled(more or less) on packager
18:47:19 <akurtakov> yep, most
18:48:08 <sochotni> what is the DEADREVIEW time? 2 weeks without response?
18:48:19 <pingou> I think a bit more
18:48:24 <akurtakov> well, I'm not sure there is some set
18:48:24 <pingou> first ping once
18:48:45 * pingou refers to tibbis' scale
18:48:53 <akurtakov> I usually close when I review my bugs and there was no response for more than 2months
18:48:56 <pingou> tibb's*
18:49:48 <sochotni> akurtakov: 2 months is quite a long time, but I am OK with that
18:50:05 <pingou> sochotni: ping after 2/3 weeks, 3/4 weeks later close
18:50:29 <sochotni> pingou: I am doing something similar I guess
18:51:03 <pingou> 2 weeks is short, some people have holidays :)
18:51:04 <sochotni> #info most SIG bugs are stalled package reviews. Try to ping packages in a few weeks, close bugs after a few more without reponse
18:51:32 <sochotni> pingou: that was with ping in middle I believe...
18:51:40 <akurtakov> there are also few FTBFS
18:51:41 <pingou> ah ok
18:52:00 <akurtakov> java-gnome doesn't work with lates libs
18:52:08 <akurtakov> work is supposed to happen upstream
18:52:51 <akurtakov> intellij-idea and jps are playing games with javac and this time there only savor is to use ecj for the build
18:53:12 <sochotni> there is hint how to solve it in intellid-idea bug
18:53:20 <sochotni> yup, ecj
18:54:14 <akurtakov> it's up to the maintainer in this case
18:54:38 <sochotni> yup, we might ping them later during f15 phase if nothing happens I guess
18:54:43 <sochotni> #topic open floor
18:55:18 <sochotni> I just wanted to let you know I am slowly starting work on packaging sonatype-tycho
18:55:31 <overholt> that's great news, sochotni
18:55:50 <sochotni> I asked for bundling exception from FPC, seems it won't be a problem
18:56:12 <sochotni> #info https://fedorahosted.org/fpc/ticket/59 sonatype-tycho bundling exception bug
18:56:28 <sochotni> that was supposed to be #link but oh well...
18:57:22 <sochotni> overholt: well....really just starting so I guess it will take me a while to get it to build properly...
18:57:34 <overholt> sochotni: understood
18:58:22 <akurtakov> sochotni: you may start pushing the additional deps for review and let us(eclipse) know what is needed from our side
18:58:58 <sochotni> akurtakov: I think there will only be 1 or 2 packages we don't yet have so I should have one of them ready on Friday hopefully
18:59:09 <akurtakov> that's cool
18:59:18 <sochotni> it was that xml-parser thingy
18:59:20 <sochotni> de.something
19:00:14 <akurtakov> I'm more concerned about the way it uses p2 repos
19:01:17 <akurtakov> because we don't ship p2 for bundles outside of eclipse-platform
19:01:18 <sochotni> yes, that's what I'll have to look into...it's quite possible we are looking into another custom resolver similar to maven/maven2 packages
19:01:38 <sochotni> and things I haven't even though about I am sure
19:02:30 <sochotni> ok, I guess that's enough
19:02:40 <sochotni> thanks for participating everyone
19:02:43 <sochotni> #endmeeting