15:03:38 <lincolnthree> #startmeeting
15:03:38 <jbott> Meeting started Wed Aug 27 15:03:38 2014 UTC.  The chair is lincolnthree. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:03:38 <jbott> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:03:39 <jbossbot> Title:3 MeetBot - Debian Wiki
15:03:56 <lincolnthree> #chair vineetreynolds, agoncal
15:03:56 <jbott> Current chairs: agoncal lincolnthree vineetreynolds
15:04:23 <vineetreynolds> Where's George?
15:04:32 <vineetreynolds> I think we need him for this meeting
15:05:25 <lincolnthree> He said he had to take a half day today
15:05:34 <lincolnthree> not sure if he'll be here for the meeting
15:05:37 <lincolnthree> #topic Agenda
15:05:42 <lincolnthree> #addtopic JBDS status
15:05:47 <lincolnthree> #addtopic Website status
15:05:51 <lincolnthree> #addtopic Next steps
15:05:55 <lincolnthree> other topics?
15:06:03 <agoncal> JavaOne and Devoxx BE conf ?
15:06:14 <lincolnthree> #topic JavaOne and Devoxx BE Conferences
15:06:37 <maxandersen> i'm calling koen - koen is supposed to be here.
15:08:36 <maxandersen> got voice msg, sent sms. I'll try fill in as much as I can in his abscence.
15:10:40 <lincolnthree> ok. sure
15:10:46 <lincolnthree> #nexttopic
15:10:46 <jbott> #topic JBDS status
15:10:58 <lincolnthree> Alright. Thanks maxandersen, where do we stand at the moment?
15:13:58 <lincolnthree> okey..
15:14:00 <lincolnthree> lol
15:14:05 <lincolnthree> maxandersen: ?
15:15:40 <lincolnthree> *cough*
15:15:50 <lincolnthree> vineetreynolds: would you like to give an update on your stuff?
15:15:56 <vineetreynolds> Yes sure
15:15:57 <lincolnthree> thx
15:16:55 <vineetreynolds> So I fixed the hangup issue with the Forge console in JBDS.
15:17:12 <vineetreynolds> It required some changes in the Command API which was done thanks to Koen. And stalep helped earlier with identifying the cause
15:17:18 <lincolnthree> vineetreynolds: what was the root cause (in summary) ?
15:17:25 <vineetreynolds> I'm therefore comfortable taking these types of issue sure
15:17:47 <vineetreynolds> Sure - the root cause was an unimplemented feature in the F2 Aesh console for Eclipse
15:17:57 <lincolnthree> vineetreynolds: you have a knack for figuring out the tuff stuff
15:18:26 <vineetreynolds> Technically the device (Eclipse in this case) is expected to respond to a Device Status Report with a Cursor Position Report. This wasnt done and thus the console waited and waited
15:18:53 <lincolnthree> ahhhh
15:18:54 <vineetreynolds> That's the summary. I must warn though that quite a few commands are unimplmented like this DSR
15:19:03 <lincolnthree> DSR?
15:19:03 <vineetreynolds> Like Cursor Up/Down etc
15:19:12 <vineetreynolds> DSR = Device Status Report
15:19:13 <lincolnthree> ah
15:19:25 <lincolnthree> and those can't be fixed on eclipse really
15:19:32 <lincolnthree> because the console is crap
15:19:42 <vineetreynolds> oh
15:19:46 <lincolnthree> i think?
15:20:20 <vineetreynolds> Well we can find out how to do this better, using say the Remote Systems Explorer plugin (iirc that's the name)
15:21:00 <vineetreynolds> http://www.eclipse.org/tm/
15:21:00 <jbossbot> Title:3 Target Management Home
15:21:01 <lincolnthree> hm
15:21:15 <vineetreynolds> It does have a console feature from what I recall
15:21:44 <lincolnthree> vineetreynolds: http://stackoverflow.com/questions/1562600/is-there-an-eclipse-plugin-to-run-system-shell-in-the-console
15:21:44 <jbossbot> Title:3 Is there an Eclipse plugin to run system shell in the Console? - Stack Overflow
15:21:51 <vineetreynolds> Yeah
15:22:28 <vineetreynolds> This is the related project with console capabilities: http://eclipse.org/tcf/
15:22:28 <jbossbot> Title:3 Eclipse TCF
15:22:42 <lincolnthree> ah cool
15:22:50 <lincolnthree> vineetreynolds: what's your next focus?
15:23:11 <vineetreynolds> Well I was looking at another issue today, and maxandersen would like to chip in with something I guess
15:23:18 <vineetreynolds> JBIDE-17656
15:23:18 <jbossbot> jira [3JBIDE-17656] Forge crashed Eclipse [10Open (Unresolved) Bug,7 Major,6 Vineet Reynolds] https://issues.jboss.org/browse/JBIDE-17656
15:23:34 <vineetreynolds> I've finished analysis and I suspect the crash is due to us using System.gc to clean up file handles
15:23:59 <vineetreynolds> I'm just not sure that we should be so dependent on it given what it has manifested in (seriously high GC activity on Windows)
15:24:45 <vineetreynolds> lincolnthree, thoughts? Max and I tend to agree that System.gc is not required to cleanup the handle
15:24:53 <vineetreynolds> But we dont really have the context in which this was used
15:25:00 <lincolnthree> hmm
15:25:20 <lincolnthree> well, System.gc typically is the only way to ensure it gets cleaned up "now" afaik
15:25:25 <lincolnthree> i could be wrong
15:25:31 <lincolnthree> this is old functionality
15:26:12 <vineetreynolds> We can't rewrite this in any way? I think we should, because the number of times it gets invoked is well not nice
15:26:56 <vineetreynolds> Take for instance, the new project command. Every time the POM is manipulated (to set artifactid, groupId, version, etc. etc.) GC is hinted to the JVM
15:27:23 <vineetreynolds> It's also the reason why Forge on Windows appears to be slow
15:27:41 <vineetreynolds> Well, at least when you run a script like the one agoncal gave us
15:28:01 <agoncal> vineetreynolds I have tried it on the latest Forge release BTW
15:28:16 <lincolnthree> vineetreynolds: really? i thought system.gc only gets called on delete?
15:28:18 <agoncal> vineetreynolds sorry *haven't* tried
15:28:19 <vineetreynolds> agoncal, Does it crash anymore?
15:28:30 <vineetreynolds> lincolnthree, nope, this is also invoked on setContents() :/
15:28:35 <vineetreynolds> agoncal, ah
15:28:38 <lincolnthree> vineetreynolds: holy shit… ok… that's not rght
15:28:43 <vineetreynolds> lincolnthree, exactly
15:29:37 <lincolnthree> ok. thank you.
15:29:44 <lincolnthree> koentsje: welcome! any update on JBDS status?
15:29:52 <lincolnthree> koentsje: we were about to move on, but now that you're here :)
15:30:01 <lincolnthree> koentsje: vineet just updated about the console issues
15:30:24 <koentsje> lincolnthree: yes sorry to be late, i am just coming back from a doctor’s appointment with my son
15:30:31 <lincolnthree> koentsje: that's ok! no worries!
15:30:45 <maxandersen> sorry - got dragged into another fire.
15:30:52 <koentsje> lincolnthree: yes i saw the last few comments of vineet, this is about forge making eclipse crash
15:31:30 <maxandersen> vineetreynolds: cool about the console working.
15:32:03 <maxandersen> lincolnthree: about system.gc() what are the file handles that should be closed by that ? stale classloaders?
15:32:04 <vineetreynolds> Guys, it's just one issue :)
15:33:14 <koentsje> i have solved JBIDE-17628 and was verifying JBIDE-17958 but i got into weird issues i think related to not having the proper snapshot
15:33:25 <lincolnthree> maxandersen: windows jvm doesn't really close file handles until GC, which means files can't be deleted
15:33:26 <jbossbot> jira [3JBIDE-17628] Ctrl+4 does not open the default Forge runtime, but always 2.6.0 [10Resolved (Done) Bug,7 Major,6 Koen Aers] https://issues.jboss.org/browse/JBIDE-17628
15:33:29 <jbossbot> jira [3JBIDE-17958] Verifying connection blocks UI in Generate Entities Wizard [10Open (Unresolved) Bug,7 Critical,6 Unassigned] https://issues.jboss.org/browse/JBIDE-17958
15:34:01 <lincolnthree> awesome! glad to solve CTRL-4!
15:34:06 <lincolnthree> the last one seems interesting
15:34:08 <koentsje> i ran into a problem that adding jpa was not available on a newly created project
15:34:24 <lincolnthree> koentsje: I think that might be the refresh issue
15:34:40 <lincolnthree> koentsje: where Eclipse doesn't refresh after forge modifies a project
15:34:56 <koentsje> really? i have solved that ages ago?
15:35:02 <lincolnthree> koentsje: it still happens
15:35:05 <koentsje> regression maybe?
15:35:16 <koentsje> is there an issue about it?
15:35:21 <lincolnthree> koentsje: i thought so. let me check
15:35:44 <vineetreynolds> JBIDE-18071
15:35:44 <jbossbot> jira [3JBIDE-18071] Newly created JPA Entity did not appear in the Project Explorer [10Open (Unresolved) Bug,7 Major,6 Unassigned] https://issues.jboss.org/browse/JBIDE-18071
15:36:05 <lincolnthree> vineetreynolds: koentsje, yep that one sums it up
15:37:28 <koentsje> weird, i’ll take a look at that one first
15:37:52 <lincolnthree> koentsje: okay great. this one has happened to me in pretty much every presentation i've given recently :)
15:38:11 <koentsje> that sucks
15:38:24 <lincolnthree> koentsje: hopefully you can figure it out!
15:38:26 <koentsje> it’s because i haven’t given any presentations lately ;)
15:38:29 <lincolnthree> easily
15:39:09 <lincolnthree> yeah what's up with that ;)
15:39:32 <lincolnthree> koentsje: this issue also results in projects not being imported sometimes
15:40:04 <lincolnthree> i added a comment to the issue
15:40:43 <koentsje> lincolnthree: ’sometimes’ is a word that always is worrysome
15:42:25 <lincolnthree> yeah
15:42:30 <maxandersen> lincolnthree: that is not what i'm seeing. if you make sure to properly close the stream they are released. …known issue though is that java's classloaders don't relase their jar streams.
15:42:30 <maxandersen> lincolnthree: hence why i'm asking *what* file handles was running loose?
15:42:51 <lincolnthree> maxandersen: ah, just user files, project files, etc
15:44:45 <lincolnthree> maxandersen: we force close JAR files in Forge: https://github.com/forge/furnace/blob/master/container/src/main/java/org/jboss/forge/furnace/impl/modules/AddonModuleFileCache.java#L31
15:45:54 <maxandersen> vineetreynolds: its just one issue reported by three people in various contexts on the OS that majority of users are on…hence - could be pretty serious for adoptin.
15:45:54 <maxandersen> and i'm just curous to know if it is known to be classloader held file handles or just file handles in general.
15:45:55 <maxandersen> if first part - then its problematic if forge use alot of classloaders if second part then that should be easily solvable by actually ensuring file/inputstreams are closed.
15:46:43 <lincolnthree> maxandersen: i think just answered your question? :)
15:46:55 <vineetreynolds> maxandersen, it's the second. Afaik, deletion (using File.delete) still leaves a handle around so the file is not really deleted until GC
15:47:22 <vineetreynolds> Somehow this bit of code made it's way into another area meant for modifying file contents
15:48:26 <lincolnthree> vineetreynolds: yeah, modifying shouldn't need this
15:48:29 <lincolnthree> vineetreynolds: deletion does afaik
15:49:05 <vineetreynolds> http://bugs.java.com/bugdatabase/view_bug.do?bug_id=4715154
15:49:07 <jbossbot> Title:3 Bug ID: JDK-4715154 (fs) Cannot delete file if memory mapped with FileChannel.map (windows)
15:49:07 <jbossbot> jira [3JDK-4715154] (fs) Cannot delete file if memory mapped with FileChannel.map (windows) [10Closed (Won't Fix) Bug,7 P3,6 Mark Reinhold] https://bugs.openjdk.java.net/browse/JDK-4715154
15:49:20 <vineetreynolds> Strangely thats the only known bug I can find about File.delete and Windows
15:49:28 <vineetreynolds> It is specific only to memory mapped files
15:50:04 <vineetreynolds> But people report otherwise as well: http://stackoverflow.com/a/4213208
15:50:04 <jbossbot> Title:3 I can't delete a file in java - Stack Overflow
15:50:47 <lincolnthree> ok. we should move on
15:50:50 <maxandersen> lincolnthree: yeah so forced system.gc() is a bad solution for this imo…
15:50:50 <maxandersen> lincolnthree: sounds like there might be forge commands not closing stuff properly …will only be visiible on winodws (or some osx's with older filesystems)
15:51:09 <lincolnthree> sounds like we are moving on the JBDS issues which is good
15:51:17 <maxandersen> koentsje: yeah the lack of workspace refresh is really disturibng.
15:51:27 <lincolnthree> maxandersen: i disagree, we are pretty careful to close handles afaik
15:51:39 <lincolnthree> maxandersen: you can reproduce this behavior on windows in a simple test case
15:51:44 <lincolnthree> maxandersen: it's not a forge thing
15:52:16 <lincolnthree> maxandersen: hm… it's possible
15:52:26 <koentsje> maxandersen: i am surprised and totally unaware this refresh isn’t happening anymore, let
15:52:32 <koentsje> me look into it ;)
15:52:47 <maxandersen> lincolnthree: we've had tons of such issues over the years with jboss servers not closing stuff properly. its always been traced back to missing close of inputstreams and the likes.
15:53:33 <maxandersen> lincolnthree: if it was a general problem eclipse would not function on windows ;)
15:54:59 <lincolnthree1> sigh
15:55:03 <lincolnthree1> #nexttopic
15:55:06 <lincolnthree1> someone advance us
15:55:09 <lincolnthree1> and make me chair :p
15:56:06 <lincolnthree1> vineetreynolds:
15:56:09 <lincolnthree1> ^^
15:56:19 <lincolnthree1> 4 minutes left in the meeting.
15:56:24 <lincolnthree1> we have more topics
15:56:28 <agoncal> Guys, I need to go (daughter to pick up) #Bye
15:56:31 <lincolnthree1> agoncal: can you make me chair pls?
15:56:32 <vineetreynolds> #chair lincolnthree1
15:56:32 <jbott> Current chairs: agoncal lincolnthree lincolnthree1 vineetreynolds
15:56:35 <lincolnthree1> thx
15:56:40 <lincolnthree1> ok	sorry this took so long agoncal
15:56:46 <lincolnthree1> #nexttopic
15:56:46 <jbott> #topic Website status
15:57:09 <lincolnthree1> #info Should have new designs to review in a few days. Then will begin the community and documentation pages
15:57:12 <lincolnthree1> #nexttopic
15:57:12 <jbott> #topic Next steps
15:57:17 <lincolnthree1> Ok this is pretty clear.
15:57:24 <lincolnthree1> #info Work on JBDS issues!
15:57:29 <lincolnthree1> #nexttopic
15:57:29 <jbott> No next topic. Use #addtopic to add topics.
15:57:40 <lincolnthree1> #topic JavaOne and Devoxx Conferences
15:58:00 <lincolnthree1> agoncal: if you are still here and have 2m, go ahead, if not, we will catch up later, no worries
15:59:16 <lincolnthree1> ok
15:59:23 <lincolnthree1> that's all then!
16:00:02 <maxandersen> okey - so how ready are we for codefreeze of jbosstools-forge tomorrow ?
16:00:46 <lincolnthree1> maxandersen: sounds like we must fix the workspaces refresh and project not importing issues
16:00:50 <lincolnthree1> maxandersen: that's a blocker IMO
16:01:34 <vineetreynolds> maxandersen, I still havent found time for the JBDS Central wizard
16:02:42 <lincolnthree1> vineetreynolds: that would be nice to have
16:02:42 <lincolnthree1> maxandersen: this isn't the Final codefreeze is it?
16:02:50 <maxandersen> lincolnthree1: agree on the refresh one. what is the project import one ?
16:02:50 <maxandersen> vineetreynolds: which part ? do we have the wizard/cheatsheet ? the central integration I/fbricon can do once that is there.
16:03:21 <lincolnthree1> maxandersen: same issue afaik
16:03:24 <vineetreynolds> Hmm no wizard yet, I need to work on the JBDS central sources itself
16:03:27 <lincolnthree1> maxandersen: projects are sometimes not imported
16:04:33 <maxandersen> lincolnthree1: well thats the question - but realistically it wont be. too many outstanding issues
16:04:39 <maxandersen> lincolnthree1: but need to figure out if we are doing CR1 with some respins or need to add in a CR2 before GA.
16:05:25 <koentsje> maxandersen: looking at it now, i’d really opt for a CR2
16:07:53 <vineetreynolds> btw meeting is not yet over
16:08:06 <vineetreynolds> end it?
16:09:28 <maxandersen> lincolnthree1:  so if we got some specific jiras that are critcial/blockers I just want to be sure they are marked as such in jira so can tell the team we are stuck until that list is resolved.
16:09:28 <lincolnthree1> ok on it
16:09:28 <maxandersen> vineetreynolds: why?
16:09:29 <maxandersen> vineetreynolds: the whole point was no changes in central is needed for you to do the cheatsheet.
16:09:29 <maxandersen> vineetreynolds: by creating the cheatsheet and add it as an archetype i thnk was the where we ended before my PTO.?
16:09:29 <maxandersen> vineetreynolds: did I miss something ?
16:10:00 <maxandersen> lincolnthree1: on it? meaning marking the bad ones as critical/blocker.
16:10:02 <lincolnthree1> maxandersen: yes
16:10:55 <vineetreynolds> maxandersen, nope, I need to work on the archetype now I suppose
16:11:08 <vineetreynolds> But after I close out the Cordova tooling item for JBDS
16:12:10 <vineetreynolds> maxandersen, I think we were planning to use Forge to create a new project with certain dependencies. Let me go back and check why we decided on this
16:12:48 <lincolnthree1> vineetreynolds: max started talking
16:12:49 <lincolnthree1> vineetreynolds: so i let it run
16:12:49 <vineetreynolds> The same could have been achieved with an archetype, which was finalized on
16:12:55 <vineetreynolds> lincolnthree1, ok
16:14:12 <maxandersen> koentsje: ookey but we should then get *all* issues that are not important triaged *off* so we only have the list of actual work items.
16:14:12 <maxandersen> lincolnthree1: cool.
16:15:32 <koentsje> maxandersen: ok, they will go to the 4.2.x stream
16:15:43 <lincolnthree1> #endmeeting