IRC Archive for channel #xwiki on 20 February 2012

Last modified by Vincent Massol on 2012/10/18 19:22

00:16 <SvenDowideit> has quit
00:16 <SvenDowideit> has joined #xwiki
00:29 <jvdrean> has joined #xwiki
00:34 <jvdrean> has quit
00:35 <jvdrean> has joined #xwiki
00:35 <jvdrean> has quit
00:36 <SvenDowideit> has quit
00:41 <SvenDowideit> has joined #xwiki
00:44 <SvenDowideit> has quit
00:45 <SvenDowideit> has joined #xwiki
01:40 <rrodriguez> has quit
02:35 <cjd> has quit
02:38 <Guest83825> has joined #xwiki
02:38 <Guest83825> has quit
03:02 <DrLou> has quit
03:19 <cjd> has joined #xwiki
04:10 <cjd> has quit
04:48 <npm> has quit
04:48 <npm> has joined #xwiki
04:52 <DrLou> has joined #xwiki
04:53 <DrLou> has quit
05:39 <cjd> has joined #xwiki
05:58 <Denis> has quit
05:58 <Denis> has joined #xwiki
07:21 <rrodriguez> has joined #xwiki
07:42 <mflorea> has joined #xwiki
08:10 <vmassol> has joined #xwiki
08:45 <rrodriguez> has quit
08:53 <sburjan`> has quit
08:53 <sburjan`> has joined #xwiki
09:20 <tmortagne> has joined #xwiki
09:41 <sdumitriu> has quit
09:47 <+tmortagne> vmassol: Hi
09:47 <+tmortagne> any idea what could cause$xwiki-enterprise-test-extension/362/testReport/org.apache.maven.surefire.junit4/JUnit4Provider/org_xwiki_test_ui_AllTests/ this error is new to me
09:47 <+tmortagne> Caused by: java.lang.RuntimeException: Unsupported browser name []
09:47 <+tmortagne> is browser need to be set somewhere now and has been forgotten here ?
09:57 <vmassol> tmortagne: good morning, checking
10:00 <vmassol> hmm this means that the browsername passed is a white space, need to find out why
10:04 <jvdrean> has joined #xwiki
10:04 <gdelhumeau> has joined #xwiki
10:11 <CIA-120> tmortagne master * r07c42ad / (9 files in 3 dirs): XWIKI-7549: Install state of an extension located in a subfolder of the ...
10:11 <CIA-120> tmortagne stable-3.5.x * r835401c / (9 files in 3 dirs): XWIKI-7549: Install state of an extension located in a subfolder of the ...
10:12 <vmassol> tmortagne: I'm debugging it
10:12 <sburjan> has joined #xwiki
10:13 <+tmortagne> vmassol: ok
10:13 <vmassol> if ${browser} isn't defined I'm not sure what's used, I'm checking that
10:14 <vmassol> my code does:     private static final String BROWSER_NAME_SYSTEM_PROPERTY = System.getProperty("browser", "*firefox");
10:14 <+sburjan> it defaults to FF
10:14 <vmassol> but I think the problem is that the "browser" prop is defined
10:14 <vmassol> sburjan: nope
10:14 <+sburjan> doesn't work defaulting to ff ?
10:14 <vmassol> sburjan: you're confusing what we want to have and what's happening I think :)
10:15 <+sburjan> I never tested if defaulting to FF actually works ;)
10:16 <vmassol> acutally I don't even need to debug, the answer is obvious, it's an empty string or whitespace
10:16 <vmassol> I'll change the code
10:17 <Enygma`> has joined #xwiki
10:18 <vmassol> hmm actually the correct solution is to define the profiles
10:18 <vmassol> I'll do that
10:18 <+sburjan> did we have some problems with this on jenkins ?
10:18 <vmassol> and we need to find a wayt o share those profiles somehow
10:18 <+tmortagne> sburjan: (09:49:54 AM) tmortagne: any idea what could cause$xwiki-enterprise-test-extension/362/testReport/org.apache.maven.surefire.junit4/JUnit4Provider/org_xwiki_test_ui_AllTests/ this error is new to me
10:19 <+sburjan> ok. didn't saw what you wrote .. I was in the bus
10:20 <CIA-120> Vincent Massol master * r6e1489e / xwiki-enterprise-test/xwiki-enterprise-test-extension/pom.xml : [Misc] Adding missing profiles which meant the browser system property wasn't defined ...
10:22 <florinciu> has joined #xwiki
10:22 <+sburjan> Rookie question, but how come this profiles are't/can't be inherited for a parent pom ?
10:22 <vmassol> ask google!
10:23 <vmassol> :)
10:23 <vmassol> or better, ask the maven guys
10:23 <+sburjan> afaik we already do inheritance
10:23 <+sburjan> for other stuff
10:23 <vmassol> they probably had a good reason
10:23 <vmassol> actually I think it's even documented
10:23 <florinciu> has quit
10:24 <vmassol> guy today is 3.5 final release day
10:24 <vmassol> s/guy/guys/
10:24 <florinciu> has joined #xwiki
10:24 <vmassol> mflorea: you're releasing today?
10:25 <florinciu> has left #xwiki
10:25 <+mflorea> vmassol: sure, but there is one blocker XWIKI-7549
10:25 <vmassol> AFAICS all tests are passing for 3.5, right?
10:25 <+mflorea> yes
10:26 <florinciu> has joined #xwiki
10:27 <vmassol> mflorea: this one has been fixed by thomas
10:27 <vmassol> hope it's been tested enough though....
10:28 <+sburjan> is he bot stuck ?
10:28 <+sburjan> XWIKI-7549
10:28 <+mflorea> oh, I don't know why I saw it as open.. anyway, this means we're good. I want to commit some tests to AppWithinMinute that I have locally for some time, and then I'll proceed with the release
10:28 <vmassol>
10:33 <djdirks> is now known as <joris>
10:35 <+sburjan> yes vmassol, you are right. Profiles are not inherited. Period
10:35 <vmassol> I'm fixing that now
10:36 <+sburjan> curious how ;)
10:45 <+florinciu> Hi
10:45 <+florinciu> Is there a known performance issue with the activity stream on 3.4?
10:45 <vmassol> hi florinciu
10:45 <vmassol> yes there is
10:45 <+florinciu> is there a fix for it?
10:46 <+florinciu> I'm trying to pitch XWiki as a collaboration tool in my team..the loading time of the home page doesn't help :(
10:46 <vmassol>
10:47 <vmassol> indeed this is a good point
10:47 <vmassol> we need to decide if we roll back this commit from jerome for 3.5 final
10:47 <vmassol> I had asked jerome to look at it before 3.5 final but he hasn't apparenerly :(
10:47 <vmassol> I'll ping him now
10:48 <vmassol> IMO we should rollback
10:52 <+florinciu> ok I see. thanks for the info
10:52 <+florinciu> I guess I'll remove the AS from the homepage and try to put something else in place
10:55 <vmassol> mflorea, tmortagne, others: what do we do? Do we need a vote? I don't think so since this is an importabt regression introduced and a lazy commit and according to our bylaws we can ask to revert any change not explicitly voted for
10:55 <vmassol> first, do you agree that we should remove it for 3.5 final?
10:56 <vmassol> s/remove/rollback/
10:57 <evalica> has joined #xwiki
10:57 <+mflorea> reading issue description to understand the problem..
10:57 <+tmortagne> vmassol: reading
11:01 <+tmortagne> the speed difference is indeed impressive...
11:01 <vmassol> has quit
11:02 <+sburjan> and this happens only on MySQL ? because I don't see perf issues on hsqldb build
11:02 <vmassol> has joined #xwiki
11:02 <+tmortagne> but removing something in 3.5 that has been introduced in 3.2 is a bit tough
11:02 <+tmortagne> now I agree that we don't have much time for 3.5
11:02 <vmassol> well there isn't much choice
11:02 <vmassol> on one hand you have a non working software
11:02 <vmassol> OTOH you remove one not so large feature
11:03 <+tmortagne> I did not really seen what was feature was yet :)
11:03 <+tmortagne> ok to remove it for me
11:04 <vmassol> FTR I've left a message on jerome's voicemail too
11:09 <vmassol> ok I've spoken with jerome
11:09 <+florinciu> sburjan: yes. I also saw the issue only on my production machine and not when I first tested with hsqldb. maybe the issue is not that severe on the test distribution.
11:09 <vmassol> he's ok to rollback too
11:09 <vmassol> what we could do is:
11:09 <vmassol> - rollback for 3.5 final
11:09 <vmassol> - put it back in 3.5.1 if someones fixes the issue
11:09 <vmassol> - ofc we need to document this in the RN for 3.5 since it means we remove an exsiting feature
11:10 <+sburjan> does this break backward compatibility ?
11:10 <+tmortagne> probably, removing something always break backward compatibility
11:10 <vmassol> it's not used by us directly though
11:11 <vmassol> (at least jerome doesn't think so)
11:11 <vmassol> jerome had an external need when he added it
11:13 <CIA-120> Vincent Massol master * r0956a6a / (7 files in 7 dirs): [Misc] Put browser profile definitions in the test top level POM but continue defining default browser in children POM since profiles aren't inherited in Maven -
11:13 <vmassol> sburjan: check that commit
11:14 <+sburjan> checking .. are you sure it works ?
11:14 <lucaa> has joined #xwiki
11:14 <vmassol> sburjan: if you can test it too that would be good, it seems to work for me
11:14 <+sburjan> I will test after I finish with M1 and final.
11:16 <+sburjan> afaik we already had this in xwiki-enterprise-test/pom.xml and didn't worked, you had to specify the profiles in ui-tests and selenium-tests. what triggers now the inheritance ?
11:16 <vmassol> there's no inheritance
11:16 <vmassol> profiles are not inherited perdio
11:16 <vmassol> :)
11:16 <+sburjan> oh, you felt the ff profile definition there
11:16 <+sburjan> *left
11:17 <+sburjan> ok, I understand
11:17 <vmassol> but you can trigger parent profiles when building children poms
11:17 <+sburjan> may I ask how ?
11:18 <vmassol> errr
11:18 <vmassol> by using profiles
11:19 <+sburjan> i know that, but if there is no inheritance, and no profile defined for chrome in ui-tests, how can I trigger it from the parent pom. This is what I don't understand
11:19 <vmassol> using profiles
11:19 <vmassol> stadnard maven
11:19 <vmassol> -P<myprofiles>
11:19 <vmassol> how else would you trigger a prpofile?
11:20 <vmassol> I think you should read up on profiles
11:20 <+sburjan> ok.
11:20 <vmassol> this is good read:
11:20 <vmassol> (note that I already explained this to you when I saw you IRL last time ;))
11:21 <+sburjan> yes, and made sense since we declared our profiles in the ui-tests pom :)
11:22 <+sburjan> I'll have to read the URL you gave me, maybe I can understand
11:22 <vmassol> mflorea: so wdyt about XWIKI-7520
11:23 <vmassol> BTW we have 2 subchoices for rolling it back: full rollback or only partial: "I found this came from and that by removing the <map> mapping from activitystream.hbm.xml the previous speed was recovered."
11:23 <vmassol> we could do the partial rollback and explain how to put the feature back in the Rlease notes
11:24 <vmassol> (by adding some lines to activitystream.hbm.xml)
11:24 <vmassol> partial = just remove the <map> from activitystream.hbm.xml
11:27 <vmassol> ok I'm going to perform the partial rollback now on both master and 3.5 branch + create a new issue to put it back
11:27 <vmassol> shout quickly if you don't agree
11:27 <vmassol> :)
11:34 <CIA-120> Ecaterina Moraru (Valica) stable-3.5.x * r77d68eb / xwiki-platform-core/xwiki-platform-wysiwyg/xwiki-platform-wysiwyg-client/src/main/resources/org/xwiki/gwt/wysiwyg/public/stylesheets/toolbar.css : XWIKI-7532 - UI Issue on WYSIWYG Editor when editing Application Class ...
11:36 <vmassol> evalica and everyone: please don't commit on 3.5 without discussing it
11:36 <vmassol> today is release day
11:36 <vmassol> since otherwise it means your changes won't be tested
11:36 <+mflorea> vmassol: I'm fine with partial rollback. I agree with Jerome this is a nice feature and we should find the real problem
11:37 <vmassol> and this introduces the  risk of introducing a regression/problem
11:37 <vmassol> mflorea: ok doing it
11:37 <+mflorea> vmassol: evalica discussed it with me
11:38 <vmassol> please discuss here then
11:38 <vmassol> :)
11:45 <+evalica> vmassol: it was a cherry pick
11:47 <CIA-120> Vincent Massol master * r63db8fd / xwiki-platform-core/xwiki-platform-activitystream/src/main/resources/activitystream.hbm.xml : XWIKI-7520: Activity Stream performance regression -
11:48 <CIA-120> Vincent Massol stable-3.5.x * rcb7343a / xwiki-platform-core/xwiki-platform-activitystream/src/main/resources/activitystream.hbm.xml : XWIKI-7520: Activity Stream performance regression -
11:50 <vmassol> sburjan: would be great if you could test with this change in your build
11:50 <vmassol> (test 3.5 I mean)
11:52 <+sburjan> I guess I have to test on a MySQL build, right ?
11:52 <vmassol> why?
11:52 <vmassol> you do test on various DBs right?
11:53 <vmassol> it's always important to test on various DBs
11:53 <+sburjan> I can't test in 2 days 4 browsers and 3 db's
11:53 <+sburjan> I haven't got the time to test on other db's
11:54 <+sburjan> but I did not see any perf issue on AS using hsqldb
11:54 <+sburjan> this is why I am asking if I need a mysql install to compare the difference
11:57 <vmassol> the perf issue should be on all DBs
11:57 <vmassol> see
11:58 <vmassol> sburjan: you could check on oracle
11:58 <vmassol> since on oracle it takes a lot of time
11:58 <vmassol> maybe the difference is more visible
11:58 <vmassol> so ideally you should compare home page load with 3.5M1 and 3.5SNAPSHOT (after my change has been built) and compare
11:58 <+sburjan> ok. I'll check M1 vs final .. and see if there is any diference when using activity entries="5"
11:59 <vmassol> make sure you have some activity in the stream first
11:59 <vmassol> don't change the home page
11:59 <+sburjan> ofc
11:59 <vmassol> just make some activity in the wiki
11:59 <+sburjan> ok
11:59 <vmassol> mflorea: I'd like to add the note in the RN for 3.5. Could you prepare the 3.5 RN page or let me know how I should prepare it
12:02 <+sburjan> which is the best way to see if the build containing your commit has been built ? because each commit triggers a new build, right ?
12:02 <vmassol> yes
12:02 <vmassol> you can see it in jeknins
12:03 <vmassol> for ex:
12:03 <vmassol> see "changes"
12:03 <vmassol>
12:03 <vmassol> hmm is stuck again
12:04 <vmassol> mflorea: btw are jenkins jobs triggered when there are changes to 3.5 branch?
12:05 <+sburjan> ok, so it is still building with tmortagne 's fix of EM .. yours must be in the queue
12:05 <vmassol> yes except I don't see it in the queue
12:05 <vmassol> I'll force it
12:05 <+sburjan> me neither .. and I'm confused :|
12:06 <vmassol> seems github did a push though:
12:06 <vmassol> (but not for my commit It hink)
12:06 <vmassol> ok I've added a build in the queue for platform 3.5
12:07 <+mflorea> vmassol: it should trigger. We discussed about this Friday. It's strange because I'm pretty sure it used to be triggered by other commits.
12:08 <vmassol> yes that's why I'm asking because on Friday it didn't trigger
12:08 <+mflorea> but how is it possible that Thomas' commit triggered a build and yours not..
12:09 <+mflorea> or maybe tmortagne forced the build too
12:09 <vmassol> well it just says: "Started on 20 f?vr. 2012 11:34:15"
12:09 <vmassol>
12:09 <vmassol> check on master
12:09 <vmassol> checking
12:09 <vmassol> hmm
12:09 <vmassol>
12:09 <vmassol> a lot more information
12:11 <+mflorea> indeed
12:12 <vmassol> going for lunch, bbl
12:13 <+mflorea> xwiki-platform and xwiki-platform-3.5 have the same config, expect the git branch, master vs. stable-3.5.x so it should work fine for all commits
12:14 <CIA-120> tmortagne stable-3.5.x * r94999c5 / (2 files in 2 dirs): XWIKI-7547: Version parsing error when loading core Extension at startup ...
12:14 <CIA-120> tmortagne master * r419d191 / (2 files in 2 dirs): XWIKI-7547: Version parsing error when loading core Extension at startup -
12:15 <+tmortagne> I did not forced any build today
12:16 <+tmortagne> mflorea: I would like to fix XWIKI-7544 before 3.5, the fix itself should be easy but what's long is to setup what I need to test it
12:17 <+mflorea> tmortagne: np, I have to commit some functional tests for awm
12:18 <+mflorea> so I'm not starting the release right away
12:24 <mflorea> has quit
12:29 <CIA-120> Eduard Moraru feature-annotations-merged-with-comments * r13f9fff / (4 files in 3 dirs): XWIKI-7540: Merge Annotations with Comments. ...
12:45 <CIA-120> tmortagne master * r793bf2f / xwiki-platform-core/xwiki-platform-extension/xwiki-platform-extension-api/src/main/java/org/xwiki/extension/version/internal/ : XWIKI-7547: Version parsing error when loading core Extension at startup ...
12:45 <CIA-120> tmortagne stable-3.5.x * r5f7d7e2 / : XWIKI-7547: Version parsing error when loading core Extension at startup ...
12:48 <jvdrean> has quit
13:00 <+sburjan> I think I found a blocker
13:04 <jvdrean> has joined #xwiki
13:05 <+sburjan> when I create a space using the default space dashboard, the creator of the document is blank
13:06 <+sburjan> I will report it as a blocker
13:18 <+tmortagne> sburjan: yep it's not very nice, should be the context user
13:18 <+sburjan> reported it
13:19 <+sburjan> I wonder how I didn't catch this on 3.4 .. because 3.4 is also affected
13:22 <+tmortagne> ok
13:37 <mflorea> has joined #xwiki
13:48 <+mflorea> tmortagne: enterprise build stuck when installing xwiki-enterprise-installer-debian-common .
13:48 <DrLou_> has joined #xwiki
13:49 <+tmortagne> mflorea: ok that's weird since it's not doing much, will look at it
13:50 <+mflorea> the build was triggered by vmassol's commit, I think, but that commits was related to functional tests so I don't see how it could affect xwiki-enterprise-installer-debian-common
13:51 <+tmortagne> yep and maven-install plugin is just supposed to put the file in the local repository
13:51 <+tmortagne> i don't see how it could get stuck whatever the project
13:59 <gdelhumeau> has quit
14:14 <gdelhumeau> has joined #xwiki
14:27 <+sburjan> there is something odd on maven, because I don't see any newer snapshots for 3.5
14:27 <+sburjan>
14:27 <+sburjan> the last one is from 16 February .. am I in the wrong place ?
14:28 <+tmortagne> maybe none has been built  by jenkins since
14:29 <+tmortagne>
14:29 <+tmortagne> sburjan:
14:29 <+tmortagne> last built Feb 16
14:29 <+tmortagne> one is pending
14:29 <+tmortagne> should start as soon as platform is done
14:29 <+sburjan> interesting.. I thought a commit in platform will trigger a build of enterprise too
14:30 <+tmortagne> it is
14:30 <+tmortagne> but enterprise will not build if any of its dependencies is building
14:30 <+tmortagne> so it was probably stuck waiting for several builds of platform
14:31 <+sburjan> i see
14:39 <+Enygma`> hi Denis
14:39 <+Enygma`> I need to implement a DB migration and I was wondering if you have 2 minutes to explain a bit how the DB versions work and how should I use them
14:40 <vmassol> would be great to have some docs on this maybe… :)
14:44 <+Denis> Enygma`: Hi, feel free to ask, there has been a vote:
14:49 <+Enygma`> Denis: so I should specify the XWikiDBVersion for 4.0 to 40000, right?
14:49 <+Denis> vmassol: well, documentation is always good, it is a matter of time and the security module will be a priority in comparison; I have commented in the javadoc as much as possible, and there is a very complete example with the id migration I have recently commit (maybe a little complex however).
14:49 <+Enygma`> and if someone else needs to do a DB migration too, the next version for 4.0 would be 40.001
14:50 <+Denis> right
14:50 <+Denis> 40000 is taken alread
14:50 <+Denis> y
14:50 <+Denis> 40000 is the id migration
14:50 <vmassol> Denis: it depends how many times you want to answer questions about it :)
14:51 <+Denis> vmassol: there should not be much, since the early begin, there have been less than ten migrations
14:51 <vmassol> it's already costing you more than if it were documented...
14:51 <vmassol> anyway :)
14:52 <vmassol> since you spent so much time on it I think you should think about docuetnign it
14:52 <vmassol> actually it coul dhave been nice to make a separate module for it
14:52 <+Denis> I would, if I could, as I said, the new security component is first on my list
14:52 <+Enygma`> Denis: and what about shouldExecute(XWikiDBVersion startupVersion) ? From what I understand, I need to check the given startupVersion to see if the DB is at the version that I want it to be before my migration is applied, right?
14:52 <vmassol> I don't know how much links with oldcore we have
14:52 <vmassol> I thought migration was first
14:52 <vmassol> ont he list
14:53 <vmassol> since you've done it before
14:53 <+Denis> Enygma`: no, do not implement if you need a simple migration
14:53 <vmassol> right now it's in oldcore, does it mena it's throwaway code?
14:53 <+Denis> vmassol: but the security module need much more explanation
14:53 <vmassol> I think the only link should be the storage
14:54 <vmassol> so it should be pretty well compartimented and could be reusable even when we change the storage quite easily
14:54 <vmassol> but you'll have a cyclic dep
14:54 <vmassol> ....
14:54 <vmassol> ATM
14:55 <vmassol> the prereq is probably a storage module
14:55 <+Enygma`> Denis: ah, getVersion() specifies on which DB version my migration needs to be applied, and not what is the new version defined by the changes of my migration.
14:56 <+Denis> Enygma`: getVersion() give the new version you are migrating the DB to
14:56 <+Denis> migration are applied in order appropriatly
14:58 <+Denis> shouldExecute could be use to remove your migration from the list of needed migration in special case; it has been used for migration that fix migration, and that are only needed when the initial migration has been used in buggy state
14:58 <+Denis> hope you never need it !
14:59 <+Enygma`> Denis: hm... ok, I think I`ve got it :)
14:59 <+Enygma`> thanks
15:00 <+Denis> vmassol: there is some weird issue in the migration system due to the multiplicity of the stores, and the link between them
15:00 <+Denis> currently the migration system see the store as a whole, and is dependant of the kind of store
15:01 <+Denis> this is not good when we mix different kind of store (hibernate, filesystem,...)
15:02 <+Denis> so, there is more to do in that field, and my migration system is currently good from an hibernate POV (most of the migration we have done up to now)
15:02 <vmassol> ok...
15:02 <+Denis> I have not fully review it, from what it was before
15:02 <+Denis> just improvement currently, and more component sugar
15:03 <+Denis> while, the security component is more a full rewrite and expected to last longer
15:03 <vmassol> it's going to be in oldcore too?
15:03 <+Denis> no, it's fully independant and use a bridge, read the mail !
15:04 <vmassol> (yeah I need to read that mail...)
15:05 <+sburjan> this seems stuck:
15:08 <CIA-120> Marius Dumitru Florea master * r14fd395 / xwiki-platform-core/xwiki-platform-oldcore/src/main/java/com/xpn/xwiki/web/ : XWIKI-7558: Creating a space using Space Dashboard Template will result in a page with blank creator ...
15:08 <CIA-120> Marius Dumitru Florea master * rae09194 / (3 files): [misc] Preview class fields with custom display before they are added/saved to the class ...
15:08 <CIA-120> Marius Dumitru Florea stable-3.5.x * r24b18c7 / xwiki-platform-core/xwiki-platform-oldcore/src/main/java/com/xpn/xwiki/web/ : XWIKI-7558: Creating a space using Space Dashboard Template will result in a page with blank creator ...
15:08 <CIA-120> Marius Dumitru Florea stable-3.5.x * rdbc1a79 / (3 files): [misc] Preview class fields with custom display before they are added/saved to the class ...
15:13 <florinciu> has quit
15:14 <CIA-120> Marius Dumitru Florea stable-3.5.x * r007ccfa / (6 files in 2 dirs): Add functional tests for Date field (AppWithinMinutes/ClassEditor) ...
15:14 <CIA-120> Marius Dumitru Florea master * r5da298e / (6 files in 2 dirs): Add functional tests for Date field (AppWithinMinutes/ClassEditor) -
15:58 <jvelo> has joined #xwiki
16:01 <CIA-120> tmortagne stable-3.5.x * r136d113 / (5 files in 5 dirs): XWIKI-7544: Add support for standard proxy settings in Extension Manager(cherry picked from commit afba250409f3d185289d9526c07629b44c9fca19) -
16:02 <CIA-120> tmortagne master * rd481874 / (5 files in 5 dirs): XWIKI-7544: Add support for standard proxy settings in Extension Manager -
16:09 <+mflorea> vmassol: you should comment on XWIKI-7555 I remember you forced the current behavior intentionaly, by adding an object of EditModeClass to Main.WebHome
16:11 <vmassol> mflorea: yep it's intentional
16:11 <vmassol> I almost commtned on it this morning
16:13 <vmassol> mflorea: why is there a need for EditModeClass in home?
16:13 <vmassol> it only does an include
16:14 <vmassol> maybe it's a leftover
16:16 <+mflorea> to force edit mode instead of inline. If you don't have this object then getDefaultEditMode detects the include macro and defaults to inline if the included macro doesn't have a EditModeClass object (like in this case I think)
16:16 <+mflorea> s/included macro/included document
16:16 <vmassol> ok
16:18 <+mflorea> yep, I just tested, if you remove it then it goes to inline mode
16:24 <+sburjan> I see an odd behavior. If a Admin finishes editing a Page Rights page, and after that a normal user edits it, the user will still see the document being locked by the Admin. How is the current implementation, how xwiki know that Admin has left that page ?
16:24 <+sburjan> so it can unlock the page
16:24 <+tmortagne> sburjan: a lock is saved in the database for a document when someone start editing it
16:25 <+tmortagne> the lock is removed when saving or canceling
16:25 <+tmortagne> when you edit XWiki check if there is a lock saved for the document
16:25 <+tmortagne> also if you don't save or cancel the lock is lost after some time
16:25 <+tmortagne> don't remember exactly how much time
16:26 <+sburjan> ok..
16:26 <+tmortagne> probably something like 30 min
16:26 <+sburjan> still not very nice IMO
16:27 <+tmortagne> sure but it's not that easy to avoid collisions without some complex system
16:27 <vmassol> tmortagne: sburjan means something else
16:28 <vmassol> sburjan: can you explain what you mean by "If a Admin finishes editing a Page Rights page, "
16:30 <+sburjan> use case: vmassol denies comment rights for user sburjan  on a page. tmortagne comes and wants to deny  view page for the _same_ page as vmassol  denies comment rights. tmortagne will see the page locked, even if vmassol has set the permission, and navigated away from the page
16:30 <vmassol> "vmassol denies comment rights for user sburjan  on a page. " how?
16:30 <+tmortagne> the issue if that the rights UI does not "finish" anything
16:30 <+tmortagne> so it never remove the lock explicitely
16:31 <vmassol> yep
16:31 <+sburjan> go to page, Edit Rights
16:31 <+sburjan> Edit -> Access Rights
16:31 <+tmortagne> one of the reason why I never liked this save when you click thing
16:31 <vmassol> and this is a bug for me
16:31 <vmassol> it's a major pita and generates unnecessary revs
16:32 <+sburjan> I think this is a very old behavior I guess, right from the start, right ?
16:32 <vmassol> no
16:33 <+sburjan> but there is not much to do about it .. except putting a save button on the right edit page, I don't know. Should I report it ?
16:33 <vmassol> you can report it indeed, there might already be an issue though
16:33 <+tmortagne> pretty sure there is something about it already
16:34 <+tmortagne> no this particular use case but about the way right UI is working in general
16:34 <vmassol> but it's a good reminder of something we could work on in 4.X
16:34 <vmassol> since it's about usability
16:34 <vmassol> it wouldn't do any harm in reporting it again IMO
16:35 <+sburjan> ok, reporting again. I also found a really critical issue Andreas is working on .. so I won't report it again
16:43 <+sburjan> done
16:44 <jvelo> has quit
16:45 <+sburjan> vmassol: downloading latest snapshot to check AS slowness
16:46 <vmassol> ok
16:47 <+sburjan> gonna run in parallel a 3.5M1 and latest 3.5 SNAPSHOT on oracle .. let's hope I see a speed difference
16:55 <+mflorea> guys, any pending commits for 3.5?
16:55 <vmassol> not from me
16:55 <+mflorea> tmortagne?
16:55 <+tmortagne> mflorea: nothing blocker
16:55 <+mflorea> ok, I'll wait for CI to finish building 3.5 and start the release
16:56 <vmassol> cool, let's hope sburjan finds a speed difference
16:56 <+mflorea> I'll repare the release notes in the mean time
16:56 <+sburjan> just give me 5 minutes, unpacking zip's
16:56 <+mflorea> s/repare/prepare
17:02 <nIMBVS> has joined #xwiki
17:03 <nIMBVS> hi. after enabling filesystem store for attachments and using it for some time, some of my attachments stopped working
17:04 <nIMBVS> I didn't find a pattern for them not working anymore...
17:04 <vmassol> a good question for cjd if he's avail… ;)
17:05 <+sburjan> ok, I have at least one odd issue. I am not able to start xwiki with custom port on windows
17:05 <+sburjan> let me investigate
17:06 <nIMBVS> but when I access the attachments I get "Error. The attachment does not exist"
17:06 <nIMBVS> but I can find the file on the filesystem. it's there...
17:06 <+mflorea> guys, do we move to 4.0M1?
17:06 <+mflorea> it is marked as blocker
17:06 <nIMBVS> nothing in catalina.out
17:07 <nIMBVS> can you advise how to enable some logging for attachments?
17:10 <+sburjan> vmassol: I have a question. I am not using the default xwiki database name, and it seems it doesn't start if the db name isn't xwiki. Do we have somewhere hardcoded xwiki database for oracle ? I am running XE, not XEM
17:13 <vmassol> you must use xwiki
17:13 <vmassol> and on oracle db == user
17:14 <vmassol> use the default db name and you'll be fine
17:14 <+sburjan> yes.. but I didn't knew we had a restriction on db name ... don't like it
17:14 <vmassol> change it
17:14 <vmassol> fix it
17:14 <+sburjan> I wanted to have both running in parallel ... so I can measure timings
17:15 <vmassol> that's bad
17:15 <vmassol> it's really the worse of things to do
17:15 <vmassol> running in // will defniitely mess up your times
17:15 <vmassol> very bad
17:15 <vmassol> just time it one after another
17:15 <+sburjan> okay.. so all for the best running them separately
17:15 <+mflorea> ok, I'll move XWIKI-7513 to 4.0M1 since there is a workaround provided in the comments
17:21 <vmassol> mflorea: hmm
17:21 <vmassol> just seen about 7513
17:21 <+tmortagne> issue with XWIKI-7513 is that it's a regression
17:21 <vmassol> it's a blocker
17:21 <vmassol> and a regression
17:22 <vmassol> I don't see how we could release without it
17:22 <+tmortagne> we released 3.4
17:22 <vmassol> because we didn't notice it
17:22 <vmassol> which means 3.4 is a broken release
17:22 <vmassol> I don't see this as good :)
17:22 <vmassol> 3.5 is supposed to be the last release
17:23 <vmassol> of the 3.x cycle
17:23 <vmassol> it would be better if it worked
17:24 <vmassol> are we using REST upload of attachments in XE?
17:25 <vmassol> this is caused by the move to restlet 2.0 right?
17:25 <+tmortagne> in XE we always set the mime type when dealing with REST whatever we do
17:25 <+sburjan> vmassol: yes, I see some slowness on 3.5M1 .. just by importing the xar which already fills a lot of AS data
17:25 <+tmortagne> yes
17:25 <+tmortagne> I think
17:25 <vmassol> seems jerome is pretty unlucky in his commits ;)
17:25 <+mflorea> :)
17:26 <+tmortagne> yea
17:26 <vmassol> if I remember well we arleady rolled back once the move to restlet 2.0
17:26 <+tmortagne> yes
17:26 <+tmortagne> but now it's going to be harder
17:26 <+sburjan> just pray I don't see the same slowness on SNAPSHOT too.. this would make unlucky too. Installing firebug so I can measure times more accurately
17:27 <+tmortagne> (and IMO we should not revert for this)
17:28 <vmassol> yes reverting seems a bit too excessive here
17:28 <vmassol> if we set the mime type internally then no feature of XE is broken then
17:28 <vmassol> it's only external code which might be broken
17:28 <+mflorea> maybe setting the content type is a best practice
17:29 <+tmortagne> and kind of crappy code since it's not that nice to upload something without indicating the mime type
17:29 <vmassol> yes maybe in which case we wouldn't have a rgression
17:29 <+mflorea> i.e. maybe the behavior or restlet 2.0 is intentional
17:29 <+tmortagne> but restlet guys did not answer to my message so i don't know if it's on purpose on not
17:30 <+tmortagne> it does not look intentional in the code, at least there is nothing explaining it's ok it behave like this
17:30 <vmassol> I'm with jerome
17:31 <+tmortagne> it get the mime type and compare it without really expecting null it seems
Created by IRCBot on 2012/02/20 00:16

Get Connected