[00:11:47] *** tfennelly has quit IRC
[00:15:00] *** tfennelly has joined #jbosstools
[00:15:01] *** ChanServ sets mode: +o tfennelly
[00:33:38] *** lzoubek has left #jbosstools
[00:39:52] *** tfennelly has quit IRC
[00:43:32] *** bfitzpat has quit IRC
[00:46:01] *** sgilda has quit IRC
[00:53:10] *** aslak has quit IRC
[00:58:59] *** sgilda has joined #jbosstools
[01:03:28] *** adietisheim has quit IRC
[01:10:48] *** Snjeza has left #jbosstools
[01:20:09] *** tfennelly has joined #jbosstools
[01:20:09] *** ChanServ sets mode: +o tfennelly
[01:30:17] *** nicoulaj has quit IRC
[01:44:10] *** tfennelly has quit IRC
[01:56:16] *** tfennelly has joined #jbosstools
[01:56:16] *** ChanServ sets mode: +o tfennelly
[02:00:38] *** digid has joined #jbosstools
[02:01:51] *** tfennelly has quit IRC
[02:26:47] *** digid has quit IRC
[02:46:13] *** fbricon has quit IRC
[03:05:01] *** koentsje has joined #jbosstools
[03:05:11] *** digid has joined #jbosstools
[03:05:22] *** Diablo-D3 has joined #jbosstools
[03:08:44] *** akazakov has quit IRC
[03:08:59] *** koentsje has quit IRC
[03:57:19] *** digid has quit IRC
[04:03:24] *** digid has joined #jbosstools
[04:04:18] *** irooskov has quit IRC
[04:18:31] *** digid has quit IRC
[04:31:37] *** bdlink has quit IRC
[04:41:08] *** dennybj has joined #jbosstools
[05:27:12] *** digid has joined #jbosstools
[05:28:27] *** digid has quit IRC
[05:46:57] *** dennybj has quit IRC
[06:11:18] *** rmartinelli has quit IRC
[06:53:58] *** rawbdor has quit IRC
[07:45:37] *** rawbdor has joined #jbosstools
[08:15:57] *** sgilda has quit IRC
[08:28:49] *** sgilda has joined #jbosstools
[08:28:56] *** fbricon has joined #jbosstools
[08:28:56] *** ChanServ sets mode: +o fbricon
[08:33:30] *** sgilda has quit IRC
[08:38:05] *** dgeraskov has joined #jbosstools
[08:38:05] *** ChanServ sets mode: +o dgeraskov
[08:42:57] *** sgilda has joined #jbosstools
[08:56:47] *** dennybj has joined #jbosstools
[09:01:18] *** mmalina has joined #jbosstools
[09:17:07] *** mgoldmann has joined #jbosstools
[09:32:06] *** koentsje has joined #jbosstools
[09:42:16] *** lzoubek has joined #jbosstools
[09:42:16] *** ChanServ sets mode: +o lzoubek
[09:42:49] *** maxandersen has joined #jbosstools
[09:42:49] *** ChanServ sets mode: +o maxandersen
[09:52:22] *** sgilda has quit IRC
[10:17:36] *** adietisheim has joined #jbosstools
[10:17:37] *** ChanServ sets mode: +o adietisheim
[10:24:07] *** dpalmer has joined #jbosstools
[10:33:28] <adietisheim> ping fbricon
[10:37:24] *** nicoulaj has joined #jbosstools
[10:48:23] <fbricon> adietisheim: pong
[10:48:49] <adietisheim> fbricon: morning fred. just noticed something odd when using m2e. was wondering if it is a bug
[10:49:13] *** dmaliarevich has joined #jbosstools
[10:49:34] <fbricon> adietisheim: explain
[10:50:04] <adietisheim> fbricon: I have a maven enabled proj in workspace. Each time I change the jsp, the proj conf gets out of date, get compliation errors. have to "Maven->UPdate proj config"
[10:50:34] <adietisheim> fbricon: have to do this on each change in the jsp. even if I just add a space
[10:50:43] <fbricon> can you send me the project?
[10:50:47] <adietisheim> sure
[10:52:23] <fbricon> rawbdor: ping
[10:52:29] <adietisheim> fbricon: http://ge.tt/88no9l6?c
[10:52:37] <rawbdor> fbricon,  pong
[10:53:08] <fbricon> rawbdor: have you seen the not-deployed .class issues with SAR?
[10:53:47] <rawbdor> fbricon,  no.  did you post it on a jira?
[10:54:03] <fbricon> I commented the SAR issue
[10:54:19] <rawbdor> fbricon,  link?
[10:54:20] <rawbdor> ;)
[10:54:46] <fbricon> rawbdor: https://issues.jboss.org/browse/JBIDE-9497?focusedCommentId=12620554&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-12620554
[10:54:52] <jbott> Feature Request: [JBIDE-9497] Add support for SAR projects [Open, Major, fbricon] https://issues.jboss.org/browse/JBIDE-9497
[10:55:24] <rawbdor> fbricon,  clearly an issue with my implementation.
[10:55:34] <rawbdor> there are tons of things going on in the background of wtp
[10:55:58] <rawbdor> i'll have to investigate. give me a few
[10:55:59] <rawbdor> :)
[10:56:38] <fbricon> rawbdor: you also need to export some packages too in the manifest (see previous comment)
[11:00:03] *** gbrown has joined #jbosstools
[11:00:26] <fbricon> adietisheim: not seeing your issue. Have you tried turning it off and on again?
[11:00:42] <adietisheim> fbricon: turning what on and off?
[11:01:19] <fbricon> adietisheim: eclipse. http://www.youtube.com/watch?v=nn2FB1P_Mn8
[11:01:23] <adietisheim> fbricon: the proj is actually shared via git. I wonder if this is somehow related
[11:01:40] <fbricon> ah, lemme try that too
[11:01:52] <adietisheim> fbricon: hehehe, brilliant :)
[11:03:50] <rawbdor> fbricon,  you mean export stuff in the sar's manifest?
[11:03:53] <rawbdor> or export from the mbean plugin?
[11:04:00] <rawbdor> you mean mbean plugin, right?
[11:04:01] <adietisheim> fbricon: seems like works even better with french accent: http://www.youtube.com/watch?v=MJXww8aizAM&feature=related
[11:05:58] *** kkhan has joined #jbosstools
[11:06:11] <fbricon> rawbdor: yes
[11:06:33] <fbricon> adietisheim: doesn't sound french *at all*
[11:07:00] <adietisheim> fbricon: well, guess british trying to sound french ;)
[11:08:53] <fbricon> adietisheim: this guys does it better http://www.youtube.com/watch?v=dABo_DCIdpM&feature=player_detailpage#t=287s
[11:09:25] *** rawbdor has quit IRC
[11:09:27] <adietisheim> fbricon: awesome :D
[11:09:56] <fbricon> adietisheim: the whole stuffis excellent :-)
[11:13:04] *** rawbdor has joined #jbosstools
[11:13:45] <maxandersen> rawbdor: adietisheim: sorry got stuck discussing marketing during concierge meeting with chocolate croissants
[11:14:02] <rawbdor> maxandersen,  i thought meeting is in 1 hour?
[11:14:06] <rawbdor> or 45 minutes?
[11:14:06] <maxandersen> rawbdor: adietisheim: could you guys go on my conf call ? forgot my headset ;(
[11:14:09] <maxandersen> rawbdor: no its
[11:14:10] <maxandersen> now
[11:14:17] <rawbdor> ok well i have a huge problem
[11:14:26] <rawbdor> whenever i get these meeting emails
[11:14:29] <maxandersen> which is  ?
[11:14:29] <rawbdor> they all look blank
[11:14:46] <maxandersen> rawbdor: and they dont show up in your zimbra?
[11:15:10] <adietisheim> maxandersen: yep, mail is blank, and I only had the appointment in zimbra this morning
[11:15:13] <maxandersen> everyone else but you and adietisheim seem to be able to read them so wondering what is going on.
[11:15:40] <maxandersen> adietisheim: well i sent it out last night so if you didnt check mail before then you wouldnt have it in there before that.
[11:15:45] <adietisheim> maxandersen: the zimbra entry has no conf call details, thought it was on skype as usual
[11:16:02] <maxandersen> adietisheim: it was supposed to but I forgot my headset (fail!)
[11:16:09] <adietisheim> maxandersen: nope, checked when I sent you the request "when is it?". Had no zimbra entry then
[11:16:10] <rawbdor> http://oxbeef.net/tmp/crap.png
[11:16:13] <rawbdor> this is what it looks like
[11:16:15] <rawbdor> in my mail box
[11:16:18] <rawbdor> seriously
[11:16:26] <adietisheim> rawbdor: maxandersen: dito here
[11:16:26] <rawbdor> i have to do ctrl+U to see anything resembling details
[11:16:34] <rawbdor> and even then it's all cryptic
[11:16:40] <maxandersen> rawbdor: get a better mail client ;)
[11:16:46] <maxandersen> rawbdor: how is it looking in your zimbra?
[11:16:58] <rawbdor> i dont know i almost never log into zimbra its slow as balls for me
[11:17:25] <rawbdor> anyway, how long will this meeting take?
[11:18:14] <maxandersen> i got a hard stop 12ish so no longer than that.
[11:18:20] <maxandersen> its sync up for openshift
[11:18:44] <maxandersen> fbricon: I sent you a fake invite - could you screnshot that and let me know how  it looks?
[11:19:03] <adietisheim> maxandersen: rawbdor: here's how it looks in zimbra http://yfrog.com/picsy
[11:19:59] <rawbdor> ........ adietisheim   i just get an image of a frog in an ice cube
[11:20:01] <rawbdor> ...........
[11:20:40] <fbricon> rawbdor: seems you have a few unread emails ;-P
[11:20:46] <rawbdor> fbricon,  where?
[11:21:18] <fbricon> rawbdor: thunderbird (http://oxbeef.net/tmp/crap.png)
[11:21:51] *** rruss has quit IRC
[11:22:18] <rawbdor> maxandersen,  i joined what i think is your con number
[11:22:33] <rawbdor> fbricon,  please tell me which folders i should be reading ;)
[11:22:36] <maxandersen> rawbdor: cool im yrying to dial in
[11:22:42] <rawbdor> eng-china is spam
[11:22:45] <rawbdor> hiring is also spam
[11:22:45] <maxandersen> rawbdor: everyone - otherwise unsubscribe ;)
[11:22:57] <rawbdor> some of these i can't unsubscribe from lol
[11:23:03] <maxandersen> rawbdor: why not?
[11:23:11] <rawbdor> things like hiring i have no idea who to ask to remove me
[11:23:14] <rawbdor> i've asked 3x before
[11:23:17] <rawbdor> nobody ever acted on it
[11:23:21] <rawbdor> so now i get china resume spam
[11:23:33] <rawbdor> eng-china is mostly crap but sometimes find out some cool gossip
[11:23:44] <rawbdor> commits is usually crap but i often find myself checking it to see who broke something
[11:23:56] <rawbdor> wtp i read all the subject lines but most don't apply to me
[11:24:49] <fbricon> maxandersen: got your fake invite
[11:25:03] <fbricon> cause it's fake right?
[11:25:58] <maxandersen> fbricon: yes - just wanted to see if you get it ?
[11:26:29] <fbricon> maxandersen: looks good to me
[11:26:39] <maxandersen> adietisheim: can you join in ?
[11:26:50] <maxandersen> fbricon: yeah - have an ics attachement with details riht?
[11:26:54] <adietisheim> maxandersen: what's the number?
[11:26:57] <maxandersen> fbricon: what mail client are you using?
[11:27:03] <maxandersen> adietisheim: the normal one for destudio
[11:27:08] <adietisheim> maxandersen: ok
[11:27:17] <maxandersen> adietisheim: got it?
[11:27:32] <adietisheim> maxandersen: yup, dialing
[11:27:43] *** pjain has joined #jbosstools
[11:29:17] <fbricon> maxandersen: thunderbird. git the .ics, but it wants to open it w/ evolution.
[11:29:21] <fbricon> *got
[11:29:52] <maxandersen> ok
[11:32:17] <rawbdor> maxandersen,  adietisheim   really really weird
[11:32:26] <rawbdor> if skype is open and i stay silent, input volume stays high
[11:34:25] <rawbdor> fbricon,  can you svn up and check it?
[11:34:27] <rawbdor> the osgi?
[11:34:35] <fbricon> rawbdor: ok
[11:40:18] <rawbdor> https://issues.jboss.org/browse/JBIDE-9479
[11:40:21] <rawbdor> maxandersen, here
[11:40:22] <jbott> Enhancement: [JBIDE-9479] Enhance custom classpath containers for additional support such as as7's modules tag [Open, Major, (JBossAS), rob.stryker] https://issues.jboss.org/browse/JBIDE-9479
[11:40:55] *** tfennelly has joined #jbosstools
[11:40:56] *** ChanServ sets mode: +o tfennelly
[11:57:31] *** sgilda has joined #jbosstools
[11:57:51] <fbricon> maxandersen: when registering to Openshift, there's a link : Already have a redhat.com or RHN account?
[11:58:06] <fbricon> maxandersen: does it work with our SSO?
[11:58:10] <maxandersen> no
[11:58:57] <fbricon> maxandersen: so we just register / sign in like the regular folks :-)
[11:59:29] <fbricon> any perks for being red hatters? (bigger machines / CPU / RAM / features) ?
[12:10:08] *** rawbdor has quit IRC
[12:10:38] *** rawbdor has joined #jbosstools
[12:10:43] <rawbdor>  https://bugs.eclipse.org/bugs/show_bug.cgi?id=326526
[12:24:24] * fbricon is on a cloud
[12:27:08] *** aslak has joined #jbosstools
[12:27:09] *** aslak has joined #jbosstools
[12:55:16] *** Snjeza has joined #jbosstools
[12:55:16] *** ChanServ sets mode: +o Snjeza
[12:58:04] <rawbdor> adietisheim,
[12:58:11] <adietisheim> rawbdor: yup?
[12:58:16] <rawbdor> i seem to have a compilation error in as.core
[12:58:59] <rawbdor> PublishCopyUtil.copyFile(etc) seems to call "StreamUtils"
[12:59:15] <rawbdor> StreamUtils is a class in an actual JBoss jar. The JBoss jar is only on the classpath of as.management,  not as.core
[13:00:11] <rawbdor> why is as.core have the line "import-package"  for as.management.as7  ??
[13:04:01] <rawbdor> adietisheim,  something seems very wrong here
[13:04:27] <adietisheim> rawbdor: ?
[13:04:36] <adietisheim> rawbdor: ah
[13:04:45] <adietisheim> rawbdor: nope, this is perfectly correct
[13:05:07] <adietisheim> rawbdor: what should be wrong with that?
[13:05:08] <rawbdor> .........?
[13:05:21] <rawbdor> because why is as.core "importing" a plugin  as.management.as7?
[13:05:30] <rawbdor> there should be no dependency here
[13:05:36] <rawbdor> and you should not be using classes from those jars
[13:05:53] <rawbdor> the only reason this plugin is "imported" is to use the StreamUtils  class once.
[13:06:20] <rawbdor> you added a dependency on as7.management that other users may not want. They may be content to download most of the as packages but not the as7 shit and instead just use the web poller
[13:06:22] <adietisheim> rawbdor: why shouldn't it? it needs the interfaces of the service
[13:06:46] <rawbdor> adietisheim,  where are the service interfaces?
[13:06:51] <adietisheim> rawbdor: in that package
[13:06:53] <rawbdor> the service interface should live in as.core
[13:07:03] <rawbdor> the other plugins should be using that interface
[13:07:41] <rawbdor> remember, if AS7 accidentally breaks api, and we need to use their jars again, we'll need to create an additional plugin  (as.management.as71) or something
[13:07:56] <rawbdor> because we can't keep two copies of the same jar names in the same plugin
[13:08:03] <rawbdor> the service interface should live in as.core
[13:08:20] <rawbdor> and also, when i import all the shit into my workspace, it doesn't compile
[13:08:23] <rawbdor> that's problem number 1
[13:08:38] <adietisheim> rawbdor: if you do that you'd end up with a service that exposes an interface from another plugin which is pretty much weird
[13:08:41] <rawbdor> i get a big red error where you're using StreamUtils
[13:08:55] <rawbdor> adietisheim,  as.core is the plugin USING the service
[13:09:03] <adietisheim> rawbdor: ??? compile error? compile's perfectly well here
[13:09:09] <rawbdor> adietisheim,  not fine for me
[13:09:30] <adietisheim> rawbdor: as.core uses the service, but the service impl must expose itself with this interface
[13:10:00] <rawbdor> adietisheim,  so what happens if we have an as71 that needs to use the same interface?
[13:10:19] <rawbdor> as.core should not depend on the management.7 and management.71 and management.72 plugins
[13:10:19] <adietisheim> rawbdor: we'll extract the interface to it's own bundle
[13:10:34] <rawbdor> that's only one part of the problem
[13:10:36] <adietisheim> rawbdor: it does not, it imports whoever exports this package
[13:10:44] <rawbdor> the other part is as.core using classes from as7.management
[13:10:49] <rawbdor> from the jboss runtime jars
[13:11:11] <rawbdor> either that or my eclipse is really really confused
[13:11:13] <rawbdor> ..................
[13:11:33] <rawbdor> it's trying to pull StreamUtils from   the jboss-as-protocol-7.0.0.Final.jar
[13:11:37] <adietisheim> rawbdor: agree on the streamutils
[13:11:53] <adietisheim> rawbdor: but not on the other part
[13:11:57] <adietisheim> rawbdor: let me check
[13:12:22] <rawbdor> adietisheim, anyway i think service interfaces can be done either way. I understand your point, but in this case, since as.core is the only one who will be consuming that interface, it seems reasonable for me that they declare it
[13:12:31] <rawbdor> "hey guys, if you want to provide management to as7, use this interface"
[13:12:36] <rawbdor> 7.0 says "ok"
[13:12:38] <rawbdor> 7.1 says "ok"
[13:12:45] <rawbdor> then they expose a service using the interface as.core told tem to
[13:12:50] <rawbdor> both make sense to me
[13:13:46] <adietisheim> rawbdor: dont follow you "that they declare it"?
[13:14:07] <adietisheim> rawbdor: so you would put the interface to as.core?
[13:14:34] <rawbdor> "I am the only user of the management service for as7.  I declare it. If you want to provide management functionality to me, please provide a service under (naming scheme) and the service should implement this interface"
[13:14:40] <rawbdor> yes... i would put the interface in as.core
[13:14:52] <rawbdor> anyway, PublishCopyUtil has the following line:
[13:14:52] <rawbdor> import org.jboss.ide.eclipse.as.core.util.StreamUtils;
[13:14:55] <rawbdor> there is no such class
[13:15:08] <rawbdor> in fact i don't see a stream utils anywhere
[13:15:16] <adietisheim> rawbdor: hum, i have this class
[13:15:45] <rawbdor> aaaaaaand now it works.         and now i will go away
[13:15:50] <rawbdor> sorry to bother you
[13:15:52] <rawbdor> talk more monday
[13:15:53] <rawbdor> ;)
[13:16:40] <adietisheim> rawbdor: it's there: http://anonsvn.jboss.org/repos/jbosstools/trunk/as/plugins/org.jboss.ide.eclipse.as.core/jbosscore/org/jboss/ide/eclipse/as/core/util/
[13:16:56] <adietisheim> rawbdor: ok :)
[13:17:12] <adietisheim> rawbdor: will think once more about the interface to be in core, but I think it should not
[13:20:47] *** rawbdor has quit IRC
[13:32:13] *** pjain has quit IRC
[13:34:49] *** pjain has joined #jbosstools
[13:41:49] *** pjain has quit IRC
[13:42:14] *** pjain has joined #jbosstools
[13:43:21] *** rruss has joined #jbosstools
[13:43:55] *** myarboro has joined #jbosstools
[13:44:14] *** rruss has quit IRC
[13:55:57] *** imeikas has joined #jbosstools
[14:07:35] *** pjain has quit IRC
[14:07:53] *** imeikas has quit IRC
[14:07:58] *** pjain has joined #jbosstools
[14:31:33] *** bfitzpat has joined #jbosstools
[14:31:33] *** ChanServ sets mode: +o bfitzpat
[14:36:03] *** kkhan is now known as kabir_lunch
[14:42:03] *** maxandersen has quit IRC
[14:44:29] *** danflo has joined #jbosstools
[14:44:39] *** jwulf has quit IRC
[14:48:25] *** dennybj has quit IRC
[14:57:17] *** tfennelly has quit IRC
[15:01:42] *** tfennelly has joined #jbosstools
[15:01:42] *** ChanServ sets mode: +o tfennelly
[15:08:13] *** kabir_lunch is now known as kkhan
[15:20:34] *** rmartinelli has joined #jbosstools
[15:22:42] *** rruss has joined #jbosstools
[15:37:57] *** maxandersen has joined #jbosstools
[15:37:57] *** ChanServ sets mode: +o maxandersen
[15:38:57] <fbricon> maxandersen: jboss sar support + jmx console FTW!
[15:39:06] <maxandersen> fbricon: ?
[15:39:11] <maxandersen> SUCCESS ?
[15:39:20] <fbricon> yup
[15:40:01] <fbricon> the problem I had yesterday with exploded SARs was a bug in rawblem's code
[15:41:19] <fbricon> now you can publish exploded SARs, use the JMX console to connect to them,  play with eclipse debugger
[15:43:43] <fbricon> maxandersen: embedding jars in lib/ is supported differently between AS versions. So there's not one recipe to make them work, but the tooling support it anyway
[15:43:53] <fbricon> *supports
[15:44:06] <maxandersen> supported differently between as versions ? awesome
[15:44:17] <maxandersen> its it like different every major or ?
[15:45:22] <fbricon> AS 4 can read jars from /lib, AS 5 needs a jboss-deployment.xml for it, for AS6 and AS7 couldn't make them work. It's not documented
[15:45:56] <maxandersen> funny
[15:46:13] <maxandersen> well AS7 doesnt support SAR's afaik.
[15:46:23] <maxandersen> surprised AS6 would be different than AS5 on this.
[15:46:26] <fbricon> it support legacy sars
[15:46:32] <fbricon> AS7 I mean
[15:47:08] <fbricon> anyway. The tooling works
[15:52:54] <fbricon> erratum : it's not jboss-deployment.xml but jboss-structure.xml
[16:10:00] *** maxandersen1 has joined #jbosstools
[16:10:21] *** jpav has joined #jbosstools
[16:10:30] *** maxandersen has quit IRC
[16:41:28] <maxandersen1> adietisheim: have you verified that rob's fix for detecting multiple running servers work ? i.e. he says it should now show a dialog saying it detected another server - do you really want to start. but I don't see that happen.
[16:42:54] <adietisheim> maxandersen1: hum, nope, not yet.
[16:43:02] <adietisheim> maxandersen1: ah
[16:43:13] <adietisheim> maxandersen1: guess I know what you mean now
[16:43:16] <maxandersen1> could you try that quickly ? just luanch two servers (copy one of them)
[16:43:18] <adietisheim> maxandersen1: yes, I saw it
[16:43:37] <adietisheim> maxandersen1: a dialog with 2 radios (start anyway/dont start)
[16:43:43] <adietisheim> maxandersen1: shall I screenshot?
[16:43:55] <maxandersen1> i simply can't make that happen
[16:44:02] <adietisheim> maxandersen1: hum
[16:44:26] <adietisheim> maxandersen1: guess you have a server running and try to launch another on in eclipse?, right?
[16:44:32] <adietisheim> maxandersen1: that's how I got it
[16:44:51] <maxandersen1> adietisheim: I launch two from within eclipse
[16:45:01] <adietisheim> maxandersen1: hum, not sure I tried this
[16:45:06] <adietisheim> maxandersen1: let me try it
[16:45:24] <adietisheim> maxandersen1: need some secs to finish my current task. found out why it still does not work for rob (egit)
[16:45:39] <adietisheim> maxandersen1: what a mess.. only because those guys have deprecated instructions/team proj set
[16:47:28] <adietisheim> maxandersen1: ok, found the reason. testing your stuff
[16:49:33] <adietisheim> maxandersen1: my usecase (running server & start new one in eclipse) works
[16:49:39] <adietisheim> maxandersen1: testing your usecase
[16:50:54] <adietisheim> maxandersen1: hum, works here.. running from trunk
[16:52:33] <adietisheim> maxandersen1: http://imageshack.us/photo/my-images/641/2servers.png/
[16:52:52] <maxandersen1> i get nothing here
[16:52:54] <maxandersen1> it just starts
[16:54:55] *** gbrown has quit IRC
[16:55:03] *** dgeraskov has quit IRC
[16:55:30] <maxandersen1> not getting anything.
[16:55:32] <maxandersen1> fbricon: ping
[16:58:44] <maxandersen1> Snjeza: klong
[16:58:48] <maxandersen1> dgolovin: ping
[17:00:09] <fbricon> maxandersen1: boiiing
[17:00:53] <maxandersen1> hi fbricon. talking with Snjeza about bringing in her remote attach debugger feature
[17:01:31] <fbricon> yeah, the question is what module ? :-)
[17:01:51] <maxandersen1> fbricon: and with your "materialize classpath" thing and a few other goodies that really don't have a good "home" anywhere in jbt I was going to suggest her to create that jdt module we talked about on the mailing list
[17:02:16] <fbricon> maxandersen1: the one nick was not found of
[17:02:48] <maxandersen1> oh ?actaully just adding it as a separate plugin and feature in common is ok - I just dont want it to be part of common.all.feature.
[17:03:03] <maxandersen1> since they have zero relationship
[17:06:27] <fbricon> maxandersen1: yeah we can probably create a new feature sitting in common, completely isolated from common.all.feature
[17:08:24] <maxandersen1> org.jboss.tools.common.jdt could work
[17:08:28] <maxandersen1> long package
[17:08:39] <maxandersen1> but could be moved out if needed.
[17:09:24] <maxandersen1> (need to go on a quick call with rruss brb)
[17:09:59] *** dmaliarevich has quit IRC
[17:10:52] <maxandersen1> Snjeza: would the above suggesitons work for you ?
[17:20:42] <Snjeza> maxandersen1: I can add the remote debug feature wherever you want and name it whatever you want. There are two plugins and a feature.
[17:39:17] <maxandersen1> back
[17:39:57] <maxandersen1> okey Snjeza - then I would suggest we keep it simple and just add it as org.jboss.tools.common.jdt.debug.* ?
[17:40:16] <maxandersen1> and then create a org.jboss.tools.common.jdt.feature that includes those two ?
[17:40:58] <maxandersen1> and if nick doesnt like that there there are two features in one module then it can be extracted out to a common.jdt module.
[17:45:05] *** koentsje has quit IRC
[17:45:14] <maxandersen1> fbricon: dgolovin: Snjeza: any objections about ^ ?
[17:46:00] <fbricon> maxandersen1: ok for me
[17:46:44] *** dpalmer has quit IRC
[17:49:58] <Snjeza> maxandersen1: org.jboss.tools.common.jdt.feature or org.jboss.tools.common.jdt.debug.feature?
[17:51:12] <maxandersen1> org.jboss.tools.common.jdt.feature
[17:51:46] <maxandersen1> Snjeza: then fbricon's materiailize classpath containre can be added into this feature too.
[17:52:09] <maxandersen1> unless you have a good case against that ?
[17:53:47] *** amitev has quit IRC
[17:56:26] <Snjeza> maxandersen1: the feature I have made has only Eclipse dependencies. If the materialize classpath container feature includes the same, it would be ok that these two feature are placed in the same feature.
[17:57:03] <maxandersen1> what deps are in your feature beyound listing your plugin?
[17:57:16] <maxandersen1> the rest is to calculated from the plugin right ?
[17:57:33] <Snjeza> Eclipse jdt and debug dependencies
[17:58:33] <fbricon> maxandersen1: Snjeza I don't plan on having dependencies other than Eclipse's
[17:59:18] <maxandersen1> Snjeza: thats the whole point of this jdt module - to be only having jdt dependnecies and stuff that shuold be in JDT :)
[17:59:44] <Snjeza> then ok
[18:03:09] <maxandersen1> Snjeza: jira'ed at https://issues.jboss.org/browse/JBIDE-9545
[18:03:13] <jbott> 5: [JBIDE-9545] Add remote debug attach to jboss tools [Open, Major, (JBossAS), snjeza] https://issues.jboss.org/browse/JBIDE-9545
[18:04:54] <Snjeza> maxandersen1: ok
[18:05:12] <maxandersen1> and i've answered nicks quetion about this back to july 22nd ;)
[18:05:56] <maxandersen1> fbricon: and thus you now also have a place to put your stuff.
[18:06:01] <maxandersen1> once you get back ;)
[18:06:13] <maxandersen1> 6hrs left until I disconnect from the Borg
[18:06:25] <fbricon> 6Hrs???
[18:07:05] <fbricon> you're planning on working till midnight?
[18:12:15] <Snjeza> maxandersen1: fbricon: you are planning to spend your holiday together?
[18:12:53] <maxandersen1> fbricon: I want to bring my inbox to zero so I can relax ;)
[18:13:14] <fbricon> Snjeza: yeah I'm folding maxandersen1 in my luggage
[18:13:29] <maxandersen1> fbricon: and I had a few too many hours this week I just didn't get to do anything sensible becuase of my ankle so got some work energy I need channeled out ;)
[18:13:34] *** maxandersen1 is now known as maxandersen
[18:13:34] *** maxandersen has joined #jbosstools
[18:13:34] *** ChanServ sets mode: +o maxandersen
[18:13:59] <maxandersen> fbricon: reminds me I need to start on my stretching excercises to fit  in.
[18:14:22] <fbricon> maxandersen: where are you going BTW?
[18:14:29] <maxandersen> about 2 meters
[18:15:01] <maxandersen> ill have a staycation
[18:16:00] <fbricon> I guess you prefer business travels :-)
[18:16:09] *** myarboro has quit IRC
[18:20:32] <maxandersen> ;)
[18:33:18] *** fbricon is now known as fbricon_out
[18:35:32] *** jpav has quit IRC
[18:42:33] *** akazakov has joined #jbosstools
[18:42:34] *** ChanServ sets mode: +o akazakov
[18:47:38] *** tfennelly has quit IRC
[18:47:58] *** tfennelly has joined #jbosstools
[18:47:58] *** ChanServ sets mode: +o tfennelly
[19:07:25] *** rruss has quit IRC
[19:19:51] *** aslak has quit IRC
[19:23:41] *** kkhan has quit IRC
[20:01:09] *** adietisheim has quit IRC
[20:07:36] <maxandersen> fbricon_out: still around?
[20:11:34] *** myarboro has joined #jbosstools
[20:16:48] *** bfitzpat is now known as bfitzpat_afk
[20:26:04] *** aslak has joined #jbosstools
[20:51:03] <fbricon_out> maxandersen: yes
[20:51:34] *** fbricon_out is now known as fbricon_around
[20:51:56] <maxandersen> hey fbricon_out, I had noted in my notes JBIDE-8208 as the issue I had to talk with Snjeza about concering m2e issue?but either that link/issue is the wrong one or my memory is just really bad.
[20:52:00] <jbott> Feature Request: [JBIDE-8208] Improve how annotation processors are setup/detected via maven/m2eclipse [Open, Major, (maven, UpStream), fbricon] https://issues.jboss.org/browse/JBIDE-8208
[20:52:02] <maxandersen> can you remind me again ?
[20:52:40] <fbricon_around> maxandersen: JBIDE-9271
[20:52:43] <jbott> Bug: [JBIDE-9271] Workspace locks/slow when Invoking 'CDI (Contexts and Dependency Injection Builder)' on 'project' [Open, Critical, (cdi, maven), scabanovich] https://issues.jboss.org/browse/JBIDE-9271
[20:54:39] <maxandersen> AH yes
[20:54:49] <Snjeza> maxandersen: Igor hasn't accepted the patch.
[20:55:42] <maxandersen> Snjeza: but the bugzilla talks about a *bunch* of patches - where as this is one of many right?
[20:56:29] <maxandersen> and im curious - isn't our builder being too agressive if it causes big delays just because classpath gets refreshed ? (akazakov<)
[20:58:23] <maxandersen> Snjeza: and from m2e side it sounds like reseting the classpath is also done overeagerly and they basically shouldn't reset it unless it actually changed - is that what the patch does or something different ?
[20:58:40] <Snjeza> maxandersen: right.
[20:59:29] <maxandersen> Snjeza: I suggest you isolate that *exact fix* and open a separate bug for it since that simply doesn't make any sense (IMO)
[21:00:03] <maxandersen> Snjeza: and from what i read in the bug Igor wasn't against that change, just not dearing to apply all the other stuff together with it ?
[21:00:13] <Snjeza> "and from m2e side it sounds like reseting the classpath is also done overeagerly and they basically shouldn't reset it unless it actually changed - is that what the patch does or something different ?" - I think that is the cause of the issue Lincoln described in JBIDE-9271.
[21:00:18] <jbott> Bug: [JBIDE-9271] Workspace locks/slow when Invoking 'CDI (Contexts and Dependency Injection Builder)' on 'project' [Open, Critical, (cdi, maven), scabanovich] https://issues.jboss.org/browse/JBIDE-9271
[21:01:10] *** fbricon_around is now known as fbricon_afk
[21:01:17] <maxandersen> yes - and this would *kill* performance for any decent project so this is rather critical to fix/handle (either by us (akazakov ? ??) or in m2e by you/igor)
[21:01:38] <maxandersen> from Igor: "If you still want to optimize BuildPathManager#mavenProjectChanged, please
[21:01:38] <maxandersen> submit new patch but make sure that all existing m2e tests pass with your
[21:01:38] <maxandersen> changes."
[21:02:14] <maxandersen> ah ..reading more now ;)
[21:02:22] <maxandersen>  i missed they were interconnected.
[21:03:32] <maxandersen> so there is two problems here.
[21:03:47] <maxandersen> 1) M2e is agressively updating classpaths even though they changed.
[21:04:04] <maxandersen> 2) Our builder is locking the whole workspace (yuck!) for doing its thing.
[21:04:14] <maxandersen> akazakov: dgolovin: can we do anything about #2 ?
[21:04:33] <maxandersen> akazakov: have you tried using the patch from the updatesite snjezana providded ?
[21:06:05] *** mmalina has quit IRC
[21:07:50] <dgolovin> maxandersen: locking workspace is to wide, I guess we can be more specific if that helps
[21:07:58] <Snjeza> maxandersen: :and from what i read in the bug Igor wasn't against that change, just not dearing to apply all the other stuff together with it ?" - Igor wrote a separate test for the patch that is, in my opinion, nonsense. m2e doesn't use  either FLAG_NONE or FLAG_DEPENDENCIES .
[21:08:51] <maxandersen> his link to the unit test fails - do you have that by any chance ?
[21:09:05] *** mgoldmann has quit IRC
[21:11:58] <Snjeza> mazandersen: https://github.com/sonatype/m2e-core-tests/commit/11a22ea34226db84c7f9cb1c8bc0e4b22ffc09bc
[21:12:19] <Snjeza> maxandersen:
[21:12:55] <maxandersen> what I dont get from the bugzilla is if he actually applied the hashset/hashlist fix ?
[21:13:08] <maxandersen> with that gone lincolns problem should be gone right?
[21:14:51] <akazakov> maxandersen: CDI builder can be slow in very big projects or in some specific cases. we are thinking what we can do with it. but JBIDE-9271 is not the case. CDI builder should work fast enough there.
[21:14:54] <jbott> Bug: [JBIDE-9271] Workspace locks/slow when Invoking 'CDI (Contexts and Dependency Injection Builder)' on 'project' [Open, Critical, (cdi, maven), scabanovich] https://issues.jboss.org/browse/JBIDE-9271
[21:15:31] <maxandersen> so why is lincoln / snjezana experiencing the slow down ?
[21:16:06] <maxandersen> akazakov: are you saying its purety the classpath reset from m2e that is slow in it self ?
[21:17:16] <maxandersen> snjezana: is the FLAG_NONE issue check possible to remove ? what does that prevent for you ?
[21:17:17] <maxandersen> Snjeza:
[21:17:47] <maxandersen> i think igor's point is that flag_dependencies events aren't sent unless something actually changed.
[21:17:50] <akazakov> maxandersen: no exactly. I mean resetting class path causes slaw work of CDI builder.
[21:18:33] <maxandersen> akazakov: didnt you just say JBIDE-9271 is not a case of slow cdi builder?
[21:18:36] <jbott> Bug: [JBIDE-9271] Workspace locks/slow when Invoking 'CDI (Contexts and Dependency Injection Builder)' on 'project' [Open, Critical, (cdi, maven), scabanovich] https://issues.jboss.org/browse/JBIDE-9271
[21:18:38] <maxandersen> (im confused now ;)
[21:19:06] <akazakov> I'm wasn't quite correct first time :)
[21:19:08] <maxandersen> if it is made slow by this reset - what is so special about us to be so slow because of that ?
[21:19:28] <maxandersen> My guess is that instead of only updating based on what changed you trigger a full scan ?
[21:20:14] <Snjeza> maxandersen: I wasn't able to reproduce the issue. I have created a feature patch, but there is no feedback from Lincoln yet. As I know, he is the only that can reproduce the issue.
[21:20:29] <maxandersen> oh
[21:22:14] <akazakov> maxandersen: we don't re-scan the entire classpath. we re-scan only changed entries.
[21:23:04] <Snjeza> "if it is made slow by this reset - what is so special about us to be so slow because of that ?" - m2e sets project's classpath even if the classpath hasn't been changed. The CDI or KB builder recomputes the classpath when it is changed. This can be slow if a classpath is big.
[21:23:10] <maxandersen> akazakov: so even if m2e resets his classpath container with 120 entries you dont do anything if they are the same 120 entries ? or is at classpath container level?
[21:23:47] <maxandersen> yeah exactly - but seems like eclipse JDT is fine - why do we take such a big impact?
[21:25:31] <maxandersen> so snjezana - im writing a mail to igor asking for more details?and i cant see where in your patch this breaks his test ? can you point it out ?
[21:26:44] <maxandersen> Snjeza: never mind found it
[21:27:00] <maxandersen> Snjeza: igor is right
[21:27:06] <akazakov> We analyze classpath to find changed entries
[21:27:07] <maxandersen> Snjeza:  you do this:           if (hasRemoved || hasNew) {
[21:27:07] <maxandersen> 99
[21:27:07] <maxandersen>             flags = MavenProjectChangedEvent.FLAG_DEPENDENCIES;
[21:27:07] <maxandersen> 100
[21:27:07] <maxandersen>           }
[21:27:22] <maxandersen> akazakov: yes - but thats heavy operation I assume.
[21:28:12] <maxandersen> Snjeza: that hasRemoved is set to true too often.
[21:29:17] <Snjeza> only if some project has been added or removed. Without my patch, classpath is always changed.
[21:30:03] <maxandersen> Snjeza: yes I get that - but you are also setting the flags when an unrelated project gets removed - thats what is what his test is about is it not ?
[21:30:30] <Snjeza> besides, m2e doesn't use FLAG_DEPENDENCIES
[21:30:55] <maxandersen> Snjeza: that is not his point.
[21:30:59] *** scabanovich has joined #jbosstools
[21:30:59] *** ChanServ sets mode: +o scabanovich
[21:31:37] <maxandersen> Snjeza: his point is that if you remove a independent project the flag state should be NONE
[21:31:57] *** Diablo-D3 has quit IRC
[21:32:46] <maxandersen> Snjeza: I get what you say about m2e not using the specific flag - but your patch breaks the api contract of flags not being set when nothing changed.
[21:33:05] <maxandersen> Snjeza: can't we avoid setting it when its non related removes ?
[21:38:05] <maxandersen> but it seems he did apply the linkedhashset issue and *that* would the problem Lincoln has right ?
[21:39:03] <akazakov> maxandersen: ok, if there is any difference in the classpath (for example order of jars was changed) then we do a full build instead of incremental
[21:40:20] <maxandersen> akazakov: Snjeza: reordering of jars shouldn't happen in m2e should it (when its just a jave file change) ?
[21:46:27] *** myarboro3 has joined #jbosstools
[21:46:47] *** myarboro2 has quit IRC
[21:46:59] *** rmartinelli has quit IRC
[21:48:31] *** myarboro has quit IRC
[21:49:38] <Snjeza> "yes I get that - but you are also setting the flags when an unrelated project gets removed - thats what is what his test is about is it not ?" - adding and removing projects isn't an often operation. When my patch is included, classpath is reseted when   a project is removed or added and when a classpath is changed. Withou the patch, classpath is always reseted.
[21:50:37] <maxandersen> Snjeza: I *get* that - but you also fire the Flags.FLAG_DEPENDENCIES when *unrelated* project gets added/removed.
[21:50:51] <maxandersen> Snjeza: igor doesn't want that - and that makes sense IMO
[21:51:29] <maxandersen> Snjeza: my question is if we can either fix the bug to do the right thing (only change flag when its an related project that got removed/added) or simply just ignore add/removed ?
[21:52:20] *** rmartinelli has joined #jbosstools
[21:54:31] <maxandersen> Snjeza: and again, he applied the change  about linkedHash* - that seems to me to be the biggest culprit ?
[21:55:44] *** rmartinelli has quit IRC
[21:57:19] * maxandersen need to go eat. ill be back.
[21:58:13] <Snjeza> maxandersen: "can't we avoid setting it when its non related removes ?" we don't know whether projects are related or not when setting the flag. I set the FLAG_DEPENDENCIES when the project is removed/added. m2e sets FLAG_NONE always which also is not correct.  Since we don't know whether projects are related, I set FLAG_DEPENDENCIES that resets the classpath.
[22:01:10] *** myarboro has joined #jbosstools
[22:06:08] <maxandersen> Snjeza: im confused - if you say with your patch it reduces classpath resets, how is that change still relevant ?
[22:07:06] <Snjeza> "and again, he applied the change about linkedHash* - that seems to me to be the biggest culprit ?" - that patch doesn't work anything if the classpath patch isn't included. m2e still resets the classpath for every project change (editing java file, for instance). I was debugging kb. It recomputes its classpath when m2e resets the project's classpath.   It would be good if Lincoln would try...
[22:07:08] <Snjeza> ...to reproduce the issue with the feature patch I attached to JBIDE-9271.
[22:07:12] <jbott> Bug: [JBIDE-9271] Workspace locks/slow when Invoking 'CDI (Contexts and Dependency Injection Builder)' on 'project' [Open, Critical, (cdi, maven), scabanovich] https://issues.jboss.org/browse/JBIDE-9271
[22:08:16] <maxandersen> Snjeza: ive mailed lincoln asking him.
[22:12:14] <maxandersen> Snjeza: could you create a test that shows just saving a java file triggers unnecessary updates?
[22:15:46] <maxandersen> Snjeza: ive sent mail to m2e-dev to try figure out if there is a better way to do a patch.
[22:16:27] <maxandersen> Snjeza: but would help if we can create test that isolates the prolbem to show it just like igor showed the "bad" effects of (y)our patch
[22:18:17] <maxandersen> and im still interested in us being sure CDI doesn't handle this classpath updaate ineffective anyway...
[22:18:27] <maxandersen> akazakov: Snjeza:  ^?
[22:18:45] <Snjeza> maxandersen: I could try to create a test that would show that m2e sets FLAG_NONE when adding an related project.
[22:19:32] <maxandersen> Snjeza: hehe - that would prove another point - but sure.
[22:20:07] <maxandersen> Snjeza: the critical point is to prove that changes are still reported when *nothing* changed, right ? thats  the case that affects us (as far as we can see?)
[22:20:10] <akazakov> maxandersen: yes, we are working on it. there still may be problem with changed classpath in CDI builder.
[22:22:31] <conan> maxandersen: see pvt msg
[22:29:05] <Snjeza> maxandersen: "hehe - that would prove another point - but sure." - we don't know if the projects are related. WE can set either FLAG_NONE or FLAG_DEPENDENCIES. m2e sets FLAG_NONE. They can do it because they always resets the classpath. I have to set FLAG_DEPENDENCIES and resets the classpath when adding/removing projects because I don't know whether they are related. That resets the...
[22:29:07] <Snjeza> ...classpath when adding/removing projects which is not an often operation. However, the classpath won't be reseted when the classpath isn't changed *edited a java file, for instance). m2e will reset the classpath when editing a java file which can have a negative performance effect on projects having a big classpath.
[22:30:59] <maxandersen> Snjeza: yes - and that last section igor would agree with we jst need to have a test that proves that (ignoring all the add/remove project stuff since that is not part of that issue)
[22:31:04] <maxandersen> get that ?
[22:32:22] <maxandersen> if we have that "evidence" then your pont about using flag_dependencies as a flag to trigger reset would be something to use to convince if that is not ok then need some other way to trigger the date...
[22:47:08] <maxandersen> to trigger the change i mean.
[22:48:24] *** dpalmer has joined #jbosstools
[22:53:27] <Snjeza> maxandersen: "if we have that "evidence" then your pont about using flag_dependencies as a flag to trigger reset would be something to use to convince if that is not ok then need some other way to trigger the date..." - see http://git.eclipse.org/c/m2e/m2e-core.git/tree/org.eclipse.m2e.jdt/src/org/eclipse/m2e/jdt/internal/BuildPathManager.java (mavenProjectChanged, line 204). You can see...
[22:53:29] <Snjeza> ...that updateClasspath is always called. The following is my patch https://bugs.eclipse.org/bugs/attachment.cgi?id=194379&action=diff#a/org.eclipse.m2e.jdt/src/org/eclipse/m2e/jdt/internal/BuildPathManager.java_sec1.  You can see that updateClasspath is called only if a project is added, removed or its classpath is changed.
[22:54:55] <maxandersen> yes snjezana - I get that.
[22:55:24] <maxandersen> but we need to prove that updateclasspath is done too often.
[22:55:37] <maxandersen> i..e change of a java file shuold not trigger that updateclasspath.
[22:57:29] <maxandersen> in my mail i argue for it via words - but having a unittest for it would make the point much tranger
[22:57:34] <maxandersen> stronger ;)
[23:01:57] <Snjeza> maxandersen: change of a java file calls mavenProjectChanged for sure and updates classpath. If that doesn't happen IMavenProjectChangedListener is invalid
[23:31:33] *** bfitzpat_afk has quit IRC
[23:51:38] <maxandersen> fbricon_afk: we used m2e 0.13 in jboss tools 3.2.0 right ?
[23:51:54] <maxandersen> so users should just be able to update to 1.0, correct?
[23:51:58] <maxandersen> Snjeza: ^ ?
[23:53:20] <fbricon_afk> maxandersen: JBT  3.2.0 uses m2e 0.12.x, 3.2.1 > 1.0.x
[23:53:32] *** fbricon_afk is now known as fbricon
[23:53:49] <maxandersen> so to do that move they need to uninstall m2e then install it agian, correct?
[23:54:01] <fbricon> maxandersen: indeed
[23:58:06] <fbricon> maxandersen: that namespace conflict causes almost all the m2e-wtp installation failures, according to the Eclipse Marketplace :-(





top