13:05:24 #startmeeting Java SIG -- https://fedoraproject.org/wiki/SIGs/Java 13:05:24 Meeting started Wed Sep 21 13:05:24 2011 UTC. The chair is jcapik. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:05:24 Useful Commands: #action #agreed #halp #info #idea #link #topic. 13:05:34 #meetingname java-sig 13:05:34 The meeting name has been set to 'java-sig' 13:05:44 .fas jcapik 13:05:46 .fas pingou 13:05:48 jcapik: jcapik 'Jaromír Cápík' 13:05:52 pingou: pingou 'Pierre-YvesChibon' 13:06:02 well..would have been nice to have #topic first but OK :-) 13:06:05 .fas sochotni 13:06:05 sochotni: sochotni 'Stanislav Ochotnicky' 13:06:12 * mgoldmann is here! 13:06:18 .fas akurtakov 13:06:18 akurtakov_: akurtakov 'Alexander Kurtakov' 13:06:30 .fas tradej 13:06:31 tradej: tradej 'Tomas Radej' 13:06:39 .fas mef 13:06:40 mefoster: mef '' - mfrisch 'Mike Frisch' 13:06:53 mefoster: cool, welcome! 13:07:15 Hi guys ... just sitting in for now, want to get back into the Fedora/Java world 13:07:35 ok ... let's consider the list complete 13:07:38 mefoster: sure thing, I am happy anyone actually being a bit interested :-) 13:07:42 #info jcapik pingou sochotni mgoldmann akurtakov tradej mef present 13:07:53 jcapik: mefoster actually :-) 13:08:13 mefoster: sorry :D 13:08:13 or...ah, fas accout is mef? 13:08:20 sochotni: Yup 13:08:20 .fas arg 13:08:21 mull: ricardo 'Ricardo Arguello' - smriti 'Smriti Garg' - h12345 'Zak Hargreaves' - laudares 'Fernando Laudares Camargos' - malhargan 'Mohammad alhargan' - prarit 'Prarit Bhargava' - jaswanthbhargav 'Jaswanth Bhargav' - jhargrave 'Jordan (4 more messages) 13:08:27 mull: :-) 13:08:34 split personality 13:08:39 ummmm 13:08:50 ? 13:08:53 #info mull present too 13:09:27 jcapik: ok, bndlib I guess.. 13:09:32 #topic aqute-bndlib update (0.0.363 -> 1.43.0) 13:10:26 Ok guys .... we have a small issue here 13:10:26 We suddenly need to have two different/incompatible versions of aqute-bnd(lib) in the repository. 13:10:37 aqute-bndlib 1.43.0 is needed for the maven-plugin-bundle update (2.0.0 -> 2.3.5) 13:10:37 aqute-bnd(lib) 0.0.363 is needed for critical packages like maven-shared 13:10:55 Here's a problem overview listing groups of classes in the mentioned versions 13:10:56 #link http://jcapik.fedorapeople.org/files/aqute-bndlib/aqute-mess.txt 13:11:02 how much is it used in maven-shared ? 13:11:20 any possibility to port maven-shared to the new one 13:11:41 akurtakov_: good question .... it's in the action list at the bottom 13:11:50 akurtakov_: number 4 13:12:13 Please read it ... 13:12:13 Since I'm not 100% confident with the proposed approach, any comments are welcome. 13:12:14 ah, ok 13:12:22 jcapik: I am wondering if the classpaths (package names) are the same for both 13:12:55 seems new bndlib contains more things, perhaps with few things removed 13:13:15 jcapik: you had the API compatibility report somewhere right? 13:13:25 sochotni: the link is in the TXT 13:13:31 ah, yeah 13:13:38 #link http://jcapik.fedorapeople.org/files/aqute-bndlib/0.0.363_to_1.43.0/compat_report.html 13:13:39 #link http://jcapik.fedorapeople.org/files/aqute-bndlib/0.0.363_to_1.43.0/compat_report.html 13:13:41 :-) 13:13:43 :D :D 13:14:43 jcapik: so the idea is to get bnd in, and rebuild bndlib against it 13:14:47 or vise-versa 13:15:24 akurtakov_: I'm unsure if I understood You correctly 13:15:54 akurtakov_: the idea is to "rename" current aqute-bndlib to aqute-bnd and keep it for the current packages 13:15:57 jcapik: point 3 from your list 13:16:30 I thought that one of them embeds the other, am I wrong ? 13:16:35 akurtakov_: oh yes .... I'd like to replace it with 1.43 13:16:49 akurtakov_: I think that's a mistranslation. It was supposed to be "rebuild packages depending on current aqute-bndlib with aqute-bnd" 13:17:05 or no, that's v2 13:17:15 akurtakov_: well .... new bndlib partially embeds former bnd and bndlib 13:17:39 akurtakov_: but the compatibility report shows, that both versions are incompatible 13:17:51 akurtakov_: and the TRUNK version is a total mess 13:18:08 jcapik: at least there is trunk nowadays :) 13:18:13 akurtakov_: particular grups of classes were moved .... base64 to libg, etc. 13:18:39 akurtakov_: and that seems to be absolutely incompatible with neither of the former versions 13:18:50 jcapik: ok, what are the showstoppers for your plan? 13:19:11 akurtakov_: I'm not aware of any yet 13:19:36 snice the maven gid:aid is different I don't think it's gonna cause problems similar to plexus-container-default 13:19:44 so I'd say your plan is OK 13:19:53 yup, I think so too 13:20:22 Any other ideas? 13:20:26 create a new package aqute-bnd, rebuild those few packages that depend on it with new BR/R and then update 13:20:26 jcapik: but if you have cleaning ideas, please talk to upstream 13:20:42 jcapik: it would be good to prevent further mess 13:20:46 well ideally we would be able to move packager from older version 13:20:48 sochotni: akurtakov_: sure .... 13:20:55 *packages 13:21:27 If upstream guys still use the old version, then I'd like to cooperate with them in order to convince them for update 13:21:57 jcapik: usually when you can compile with new version of a dep they will accept patches 13:22:18 Ok ... 13:23:01 so .... anything else regarding this topic? 13:23:21 nope 13:23:23 I'd say #action & #agreed 13:24:16 jcapik: I think 2-3 of packages needing rebuilds are mine, it would be good to let other owners know why we are changing this 13:24:48 #agreed Actions mentioned in http://jcapik.fedorapeople.org/files/aqute-bndlib/aqute-mess.txt will be performed by jcapik, ... 13:25:27 #topic Maven 2 vs. 3 status and explanation (if this is needed) 13:25:37 this is mine... 13:25:38 so.. 13:25:45 #chair sochotni 13:25:45 Current chairs: jcapik sochotni 13:26:03 as noob, I would love status and explanation 13:26:19 so would I 13:26:49 ok so...We used to have only maven2 package, then some time ago we added maven package (maven-3.0.3) 13:26:59 jcapik: me 2 13:27:21 since we want to get rid of maven2 (upstream is basically unmaintained now)... 13:27:35 and we don't want to support two versions unless we have to 13:27:53 we basically got rid of all BR/R on maven2 in rawhide 13:28:06 and f16 as well if I am not mistaken 13:28:09 but.. 13:28:25 maven2 contained around 15 jar files, not just the main maven 13:28:42 some of those jars are needed by maven plugins that are currently being used 13:28:53 especially during build time 13:29:12 so we split monolithic maven2 package into several sub-packages 13:29:40 current list is: maven-artifact-manager, maven-error-diagnostics, maven-model, maven-monitor, maven-plugin-repository, maven-profile, maven-project, maven-toolchain 13:29:54 these are packages that are not provided by maven-3.x 13:29:57 what about maven-core and maven-compat? 13:30:06 that's and interesting thing 13:30:41 since we no longer get maven-core and other jars from maven2 some packages might complain about missing packages etc 13:30:54 comment: they do 13:31:11 this problem can usually be solved by modifying pom.xml and adding maven-core and/or maven-compat to dependencies 13:31:31 this is because maven-3.x moved some classes around in jars 13:32:31 proper fixes for this should be upstreamable I believe 13:32:49 so if I understood you correctly yesterday (we talked with sochotni about the matter), we should test building the package with maven3 and convince upstream to move to maven3 as well 13:32:51 in any case: avoid BR/R: maven2 13:33:29 yes, it usually involves chainging mavenVersion property in pom and adding/changing dependencies a bit 13:33:39 good, thanks 13:33:39 bah, changing 13:34:17 note that some upstreams might not be willing to accept such patches 13:34:33 akurtakov_: i can imagine that 13:34:43 e.g. maven-site-plugin went through few iterations before being able to work under maven2 with maven 3.x deps 13:35:39 akurtakov_: wait a sec, you adapted its pom.xml to maven3 and made it compile with maven2? 13:35:42 mefoster: you might want to have a look at packaging guidelines again btw, there were quite a few changes since your hiatus 13:36:11 akurtakov_: i don't quite understand 13:36:12 tradej: upstream, not us. The plugin usually has to work with both maven2 and maven3 13:36:14 sochotni: I can imagine! It's all on/linked from the SIG wiki page? 13:36:22 tradej: it's not me, but it is compiling with maven 3.x but verified that all the deps are resolvable and it works under maven2 13:36:26 mefoster: should be 13:36:35 sochotni: akurtakov_: ah 13:36:50 heavy magic 13:36:50 mefoster: https://fedoraproject.org/wiki/Packaging:Java 13:37:15 also https://fedoraproject.org/wiki/Java_review_template might be a shorter version to get idea 13:37:20 sochotni: Thanks, will read. I also need to get up to speed on the new git-based process ... 13:37:28 oh, for sure 13:37:43 I am sure we can help with any problems so just ask.. 13:37:54 ok, I don't know what else to add to maven2->3 topic 13:38:01 sochotni: I like the review template :) 13:38:08 it's gone easier than I imagined really 13:38:19 my questions feel answered 13:38:52 just keep that transition in mind in case you encounter some weird build failure with missing maven class 13:39:13 the situation is much clearer to me now, thanks. 13:39:43 thx 13:39:50 anything else? 13:39:59 thx 13:40:05 #topic SIG tracking bug 13:40:41 I am slowly losing track because of all the review bugs :-) 13:40:53 ok, I'll join on this one :) 13:41:03 I have gone through all the bugs yesterday 13:41:11 #link https://bugzilla.redhat.com/show_bug.cgi?id=652183 tracker bug 13:41:12 * mgoldmann noticed that 13:41:20 there are a number of started almost done bugs 13:41:30 or blocked 13:41:52 there is nothing major according to me 13:41:54 that's true - I'll make sure today I'll move forward everything 13:42:17 akurtakov_: we were planning to do a "Review day" 13:42:30 maybe we can do it during FUDcon? :-) 13:42:35 +1 13:42:44 if you think it's a good idea 13:42:48 how should it look like? 13:43:06 +1 13:43:13 mgoldmann: review day? We sit down next to each other and start reviewing like crazy :-) 13:43:17 mgoldmann: you sit in the middle and a number of guys around you send you review requests :) 13:43:21 lol 13:43:21 I can try to do some reviews maybe ... 13:43:27 sounds good! 13:43:30 I'm in 13:43:59 akurtakov_: +1 13:43:59 * pingou votes that mgoldmann does as many reviews as he submits :p 13:44:07 #agreed Review day - most probably during FUDCon Milan 13:44:14 and then richard fontana comes around and everything gets blocked on licensing :] 13:44:17 are we all going there ? 13:44:26 * mgoldmann will be there 13:44:32 * sochotni as well 13:44:33 * akurtakov_ too 13:44:35 tradej: I'm not too bad at that as well ;) 13:44:49 * tradej will not be there afaik 13:44:52 so at least 4 of us 13:44:57 pingou: that would work if he submits less than a half of them 13:45:03 * vanaltj guesses you all are talking about Milan coming up soon, and will not be there. 13:45:12 jcapik: Milan ? 13:46:18 pingou: that was just a stupid comment to Your previous proposal : "***pingou votes that mgoldmann does as many reviews as he submits " 13:46:18 vanaltj: yeah https://fedoraproject.org/wiki/FUDCon:Milan_2011 13:46:40 jcapik: will you be in Milan with us ? 13:46:40 pingou: :D 13:46:59 pingou: should I :] 13:47:17 sochotni: so at least 5 of us :) 13:47:24 pingou: don't know 13:47:34 * mull wishes he could be there 13:47:40 jcapik: you will be in Milan with us! 13:47:44 even I'll be in Milan just thanks to me being a substitute so... 13:47:44 nice crowd I think 13:48:15 pingou: just listening? 13:48:45 jcapik: no no, you can do some reviews too ;) 13:49:15 ok, so... 13:49:20 pingou: ok ... I can do reviews, but no public talks 13:49:29 I tried to book a room for SIG meeting 13:49:36 during FUDCon 13:49:52 so if we get it we can do some hackfest in there 13:50:06 I'm pretty sure there will be some free rooms 13:50:17 in Tempe - we had a lot 13:50:24 Milan ? 13:50:35 LinuxCode: https://fedoraproject.org/wiki/FUDCon:Milan_2011 13:51:14 mgoldmann: jboss as update? 13:51:26 sure 13:51:28 #topic JBoss AS in Fedora update 13:51:56 so, as you know - I'm in the process of pushing AS 7 into Fedora 13:51:58 #link https://fedoraproject.org/wiki/JBossAS7 13:52:07 I think I can add links, right? 13:52:18 mgoldmann: yes 13:52:21 perfect 13:52:31 so, I prepared exactly 40 packages so far 13:52:45 which isn't very big amount if you look at the AS7 dependency 13:52:53 tree* 13:53:28 14 packages are already in Rawhide thanks to hard work on reviewing people like akurtakov_ sochotni tradej and others! 13:53:46 mgoldmann: you're too kind :) 13:53:48 ok 13:53:54 many reviews are blocked because it's a dependency chain 13:54:20 mgoldmann: I think it might make sense maybe to paralellize your side a bit.. 13:54:21 mgoldmann, I will try to take a few of them this week 13:54:26 I'm trying to do my best to do the fixes as fast as I can, but sometimes it takes some time 13:54:47 so that: you prepare first version of the package and someone else posts it for review, while adding you as primary maintainer in the end 13:54:57 you need to bear in mind, that I'm a lead of a project completely not related to AS7 :) 13:55:20 sochotni: that was my idea 13:55:24 are all of the reviews linked to the SIG tracker bug? 13:55:28 sochotni: https://fedoraproject.org/wiki/JBossAS7#Cleanup_queue 13:55:36 mefoster: yes 13:55:42 I think some people might scream "Murder" 13:56:08 but I was thinking about doing the initial, crappy package because I know what'e needed now 13:56:38 I cleaned the cleanup queue a month ago and since then no new package were added 13:56:55 mgoldmann: I meant even really packages you think are OK to go in... 13:56:57 mgoldmann: so you're getting no help on the packaging side ? 13:57:09 as you've seen reviews usually uncover some small issues 13:57:12 but Im now in the state that I *can* build a very slimmed down verison of AS7 based *only* on Fedora rawhide deps 13:57:30 akurtakov_: more or less correct 13:57:46 akurtakov_: I have support from the mageia guy = gil 13:57:59 I took some specs from his work 13:58:12 mull: you also have some common deps with jboss right? 13:58:19 meaning eucalyptus 13:58:24 some people asked how they can help, but it's pretty hard, so nobody really have done some work for me :) 13:58:43 sochotni, yes, hibernate especially 13:59:10 mgoldmann: having a list of first level deps ? 13:59:15 but also jboss-logging, jboss-marshalling, etc. (which I am reviewing for mgoldmann ) 13:59:26 akurtakov_: not really now 13:59:47 I think that this is the only way to get others involved 13:59:53 I noticed the wiki does not always properly put a strike through closed bugs, mgoldmann 14:00:05 so when you know they are done, you may want to update the wiki page manually 14:00:13 or some kind of idenitified missing deps - even if it's 5-6 packages or a chain of htem 14:01:01 mull: this is manual work :) 14:01:23 so, this is more or less what I have to say about AS7 14:01:38 I really appreciate the help from you guys 14:01:48 you have any questions? 14:01:59 nope, just glad to help 14:02:16 I think it would make sense if we split into packager-reviewer teams there, but we can talk about that later 14:02:21 or during FUDCon 14:02:32 during FUDCon makes sense 14:02:49 I'll have no time to package new things in the next 2 weeks I think 14:02:49 mgoldmann: one thing for fudcon that would be nice is having the dependencie chain 14:03:02 so that we could eventually split the chains in pieces 14:03:05 but I'll be still available for fixing reviews of course 14:03:07 so we can clean the queue by then :) 14:03:11 each team takes a bit of it 14:03:20 pingou: I'll work on it 14:03:23 mgoldmann: cool :) 14:03:27 but it'll be scary 14:03:28 mgoldmann: mvn-local might be of use for identifying what gids:aids are still missing 14:03:53 ok, so... 14:04:02 mull: have anything about eucalyptus? 14:04:12 or we move directly to open floor? 14:04:20 sochotni, not much to update right now 14:04:29 next goal is to build gwt and get it reviewed 14:04:37 tough goal 14:04:51 the debian packaging that's been done is helping 14:05:04 mull: java 7 might be the selling point 14:05:13 to convince upstream to move to jdt 3.7.1 14:05:19 which adds support for it 14:05:49 mull: anyway let me know when you get to the jdt issue 14:06:08 * mgoldmann is attending two meeting at the same time, fun 14:06:12 ok. that's all I have for now. thanks for your reviews, btw 14:06:13 mgoldmann: :-) 14:06:23 #topic open floor 14:06:43 do we have/want a mvn2spec tool ? 14:06:50 or pom2spec 14:07:02 pingou: rpmstubby 14:07:12 but I didn't cared to make it runnable from the console 14:07:18 akurtakov_: exactly :-) 14:07:20 as I don't use it that way 14:07:31 akurtakov_: yum doesn't find it :) 14:07:41 though it should be doable in a day 14:07:48 pingou: yum install eclipse-rpmstubby 14:08:13 does it follow our current guidelines ? 14:08:27 pingou: as much as I have identified 14:08:32 cool :) 14:08:40 note that I haven't packaged new maven thing for more than a month 14:09:25 that's likely going to be useful @fudcon if we do some packaging next to the reviews 14:10:02 I almost forgot...there is a rather big change to how we do JNI in Fedora now, but I haven't prepared any materials to explain except read one rather long bug thread... 14:10:05 so all of you guys refuse to use it from inside Eclipse and want me to make it a console app 14:10:52 * pingou notes: "install eclipse" 14:11:17 akurtakov_: it's not so much a case of "refusing" to me... 14:11:22 but a matter of convinience 14:11:43 I'd have to run eclipse, emacs, 2 VMs, firefox, freemind...I don't have 16GB of RAM 14:11:44 well,even if I change it I'll add a wrapper to run the same code on the console 14:11:49 akurtakov_: CLI CLI CLI CLI CLI :D 14:12:12 unfortunately I might add... 14:12:17 yeah. i mean eclipse is awesome, but occasionally i want to do a wild crazy search grep pipe thing, and these are really hard in eclipse 14:12:23 sochotni: I do on my server and one apps nicely uses more than half of it :) 14:13:05 if I was buying a desktop now...with the prices of RAM... 14:13:07 oh well.. 14:13:25 akurtakov_: anyway...would be nice to have, but I bet you have more pressing matters... 14:13:32 +1 14:13:52 I want to ask jcapik about plexus-maven-plugin - are all patches applied and send to upstream? 14:14:28 akurtakov_: gonna check .... but I believe ..... yes 14:16:02 so we can end ? 14:16:16 I guess 14:16:16 I'm cool 14:16:21 agreed 14:16:24 yep 14:16:26 jcapik: you can do the honours :-) 14:16:41 #endmeeting