IRC Archive for channel #xwiki on 10 January 2012

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

00:07 <jvdrean> has quit
00:44 <DrLou_> has quit
00:45 <rrodriguez> has quit
00:45 <CIA-121> Sergiu Dumitriu master * r2bbf7a8 / xwiki-platform-core/xwiki-platform-extension/xwiki-platform-extension-ui/src/main/resources/XWiki/ExtensionManagerMacros.xml : XWIKI-7290: Extension manager UI use {{warn}} instead of {{warning}} ...
02:16 <tekzilla> has quit
02:18 <tekzilla> has joined #xwiki
03:06 <CIA-121> Sergiu Dumitriu master * r7046ead / xwiki-platform-core/xwiki-platform-extension/xwiki-platform-extension-ui/src/main/resources/XWiki/ExtensionManagerMacros.xml : XWIKI-7258: Authors ending with ")" are displayed wrongly in the Extension Manager UI ...
03:47 <CIA-121> Sergiu Dumitriu master * rb496ec1 / (5 files in 2 dirs): XWIKI-7289: Make Extension Manager UI localizable ...
06:00 <Denis1> has joined #xwiki
06:00 <Denis> has quit
07:28 <mflorea> has joined #xwiki
08:36 <tmortagne> has joined #xwiki
08:52 <Denis1> is now known as <Denis>
09:09 <lucaa> has joined #xwiki
09:12 <jvdrean> has joined #xwiki
09:15 <CIA-121> Sergiu Dumitriu master * r2780273 / xwiki-platform-core/xwiki-platform-extension/xwiki-platform-extension-ui/src/main/resources/XWiki/ExtensionManagerMacros.xml : XWIKI-7259: Long text overlaps action button in Extension Manager UI ...
09:23 <evalica> has joined #xwiki
09:28 <vmassol> tmortagne: hmm I don't see the jboss cache data
09:28 <vmassol> in the jmx console
09:29 <vmassol> maybe infinispan doesn't register the jmx mbeans by default
09:32 <sburjan> has joined #xwiki
09:32 <+tmortagne> vmassol: don't know, never checked it
09:34 <Enygma`> has joined #xwiki
09:35 <vmassol> found doc, reading:
09:36 <vmassol> tmortagne: do we have an xml file for the infinispan config?
09:36 <vmassol> (how do we configure the CacheManager)
09:36 <Enygma`> has quit
09:37 <Enygma`> has joined #xwiki
09:39 <vmassol> ok I think you're doing it programmatically in InfinispanCacheFactory
09:40 <+tmortagne> vmassol: yes we have a conf file
09:41 <vmassol>  /WEB-INF/cache/infinispan/config.xml
09:41 <vmassol> ?
09:41 <vmassol> ah yes we fallback to programmatic only if it doesn't exist, cool
09:41 <jvdrean> has quit
09:41 <+tmortagne> yes
09:42 <+tmortagne> not exactly
09:42 <@sdumitriu> tmortagne: Any ETA on the ability to get all extensions from the searchable repositories, for XWIKI-7246 ?
09:43 <vmassol>  cool the file already exists
09:43 <vmassol> and it has the jmx config
09:43 <vmassol> we just need to uncomment it
09:43 <vmassol> trying it
09:45 <+tmortagne> sdumitriu: you already can get all extensions, just put a empty pattern
09:46 <vmassol> hmm doesnet work
09:46 <@sdumitriu> tmortagne: OK, will try
10:03 <vmassol> tmortagne: ok found the problem
10:03 <vmassol> container is null when InfinispanCacheFactory is initialized
10:04 <vmassol> ahah funny
10:04 <vmassol> you set it after you get the config file
10:04 <vmassol> but in the code to get the config file you check if container is null
10:04 <vmassol> :)
10:04 <vmassol> so right now the infinispan config file is not taken into account
10:06 <gdelhumeau> has joined #xwiki
10:08 <vmassol> fixing
10:15 <+tmortagne> back
10:16 <+tmortagne> ok
10:18 <vmassol> tmortagne: does it mean that the named cahces defined in config.xml were not actually defined?
10:18 <vmassol> does it mean that there was no cache for image and equation?
10:20 <+tmortagne> vmassol: probably yes
10:20 <+tmortagne> well
10:20 <+tmortagne> actually no
10:20 <+tmortagne> because there is caches before theses
10:21 <+tmortagne> hmm
10:21 <+tmortagne> yes I think it was not taken into account
10:22 <+tmortagne> vmassol: did you started modifying stuff or can I take care of it ?
10:22 <vmassol> I started
10:22 <vmassol> I'm wrting the unit test
10:22 <vmassol> almost ther
10:22 <vmassol> e
10:23 <+tmortagne> ok, basically it's just about moving this.container lookup at the top of init ?
10:23 <vmassol> yes
10:24 <vmassol> I'd like to add a comment as to why we want to make this code work even when there's no container?
10:24 <vmassol> can you explain?
10:29 <+tmortagne> vmassol: well because it's not really critical to have a confuration to have a working cache
10:29 <+tmortagne> configuration
10:30 <+tmortagne> and sometime you don't want to setup a container but still need a cache
10:30 <+tmortagne> was thinking of when we will move the cache to commons
10:30 <vmassol> ok unit test working
10:32 <vmassol> btw you could use a Provider for the container now ;)
10:32 <jvdrean> has joined #xwiki
10:48 <vmassol> tmortagne: pushed, can you verify it's ok before I merge it in 3.3.x and 3.2.x?
10:48 <vmassol> err infinispan is only in 3.3.x? or was it in 3.2.x too?
10:49 <CIA-121> Vincent Massol master * r2e30d4a / (2 files in 2 dirs): XWIKI-7365: Infinispan configuration is not taken into account -
10:49 <+tmortagne> not sure, checking
10:49 <+tmortagne> I think it's 3.3
10:50 <+tmortagne> vmassol: 3.3
10:50 <vmassol> k
10:50 <vmassol> give me your greenlight and I'll merge
10:51 <+tmortagne> vmassol: seems ok to me
10:52 <CIA-121> Vincent Massol stable-3.3.x * r24f46b7 / (2 files in 2 dirs): XWIKI-7365: Infinispan configuration is not taken into account -
11:11 <vmassol> hmm I don't see how to get the cache content in the JMX console
11:19 <evalica> has quit
11:25 <+mflorea> guys, did we change something on the xwql implementation recently, after the 3.3 release? I have queries that worked on 3.3 and don't work anymore on 3.4-SNAPSHOT, this is really bad.. again, someone is making changes without checking the CI for failing tests..
11:25 <vmassol> I think jerome chaged something
11:25 <@sdumitriu> Yes, there's a "distinct" now in the query
11:26 <+mflorea> I haven't written the app within minutes tests for nothing..$xwiki-enterprise-test-ui/852/testReport/
11:26 <+mflorea> are failing for 43 builds..
11:27 <vmassol> yep I've been amazed at how little we cared about failing builds
11:27 <vmassol> I pinged aobut this but nobody cared
11:28 <vmassol> (didn't have time to debug it myself)
11:28 <+mflorea> I'm going to revert jerome's change
11:32 <@sdumitriu> vmassol: I cared a little...
11:43 <sdumitriu> has quit
11:48 <CIA-121> cjdelisle feature-store-attachments-newstore * r0903ac7 / (7 files in 6 dirs): Got newstore filesysten attachments working. -
11:55 <CIA-121> Marius Dumitru Florea master * r0e59b3c / (2 files in 2 dirs): Revert "XWIKI-7273 XWQL short form queries should distinct on document fullname." because it breaks existing queries. ...
11:57 <evalica> has joined #xwiki
11:58 <CIA-121> cjdelisle master * re479bf9 / xwiki-platform-core/xwiki-platform-oldcore/src/main/java/com/xpn/xwiki/ : XWIKI-6757: XWikiException stack traces do not show wrapped exception. ...
12:04 <sburjan> has quit
13:03 <mflorea> has quit
13:20 <sburjan> has joined #xwiki
13:27 <sburjan> has quit
13:40 <sburjan> has joined #xwiki
13:47 <sburjan> has quit
13:51 <sburjan> has joined #xwiki
13:59 <CIA-121> Marius Dumitru Florea master * r5a1d39e / xwiki-enterprise-test/xwiki-enterprise-test-selenium/src/test/it/org/xwiki/test/selenium/ : XWIKI-7355: Use PNG silk icons instead of GIF ...
14:00 <mflorea> has joined #xwiki
14:01 <+mflorea> tmortagne: do you have any idea why wysiwyg selenium tests are not triggered anymore? I committed yesterday both in platform and in xwiki-enterprise-test-wysiwyg and the job hasn't been triggered. I've looked at the job configuration but I can't see anything special compared to other enterprise test modules.
14:01 <+mflorea> in the "Build Triggers" configuration section I see that it is set to run after xwiki-enterprise-test-selenium, and the "Build when a change is pushed to GitHub" is not checked (probably because this is checked for enterprise which then triggers the test modules)
14:06 <sburjan> has quit
14:28 <+tmortagne> mflorea: looking
14:30 <+tmortagne> mflorea: configuration seems ok, here is the triggering chain: xwiki-enterprise -> xwiki-enterprise-test-pom  -> xwiki-enterprise-test-selenium  -> xwiki-enterprise-test-wysiwyg
14:30 <+tmortagne> so building xwiki-enterprise should eventually build wysiwyg tests
14:31 <+tmortagne> I can see the conf work for xwiki-enterprise-test-pom
14:31 <+tmortagne> it's ok for xwiki-enterprise-test-selenium too
14:31 <+mflorea> could be that test-selenium was failing? it wasn't the case before, i.e. wysiwyg tests were run even if there were selenium test failyres
14:31 <+tmortagne> mflorea: see
14:31 <+tmortagne> "Started by upstream project xwiki-enterprise-test-selenium build number 902"
14:31 <+tmortagne> so it seems to work
14:32 <+mflorea> now, because I just fixed selenium tests
14:32 <+mflorea> i.e. xwiki-enterprise-test-selenium
14:32 <+tmortagne> ha yes trigering is always broken if something fail
14:32 <+mflorea> but this isn't right
14:32 <+mflorea> ah
14:32 <+tmortagne> I just looked at latest builds
14:32 <+mflorea> I wasn't aware of this
14:33 <+tmortagne> maybe it's configurable
14:33 <+tmortagne> ye it is
14:33 <+tmortagne> "Trigger only if build succeeds"
14:34 <+tmortagne> changing it for xwiki-enterprise-test-selenium
14:34 <+tmortagne> since we don't need the test to pass for the WYSIWYG
14:34 <+tmortagne> but it manily mainly mean that the thing the WYSIWYG need shouldn't be in this project
14:34 <+tmortagne> like UI test framework is actually in another project
14:35 <+tmortagne> changed for "Trigger even if the build is unstable"
14:35 <+tmortagne> I should maybe do the same for platform, not sure
14:36 <+mflorea> yep
14:36 <+mflorea> the common API should be in a separate module
14:36 <+mflorea> but since we have to move the tests to selenium 2 it not worth to make this change now
14:37 <vmassol> tmortagne: just noticed that my linkchecker is not multiwiki-aware…. I'm not sure how to fix this though since the linkchecker manager is in rendering which doesn't know about the notion of wiki...
14:37 <+mflorea> i.e. to refactor selenium tests
14:37 <+tmortagne> mflorea: I agree
14:37 <vmassol> right now it opens a privacy issue since anyone can see the link status for any subwiki
14:38 <+tmortagne> vmassol: what is not multiwiki in linkchecker exactly ?
14:38 <+mflorea> tmortagne: I remember a recent commit about case insensitive search but I can't find it. do you know it? otherwise I'll keep searching
14:38 <vmassol> tmortagne:
14:38 <vmassol> even though this is in the commons wiki you can see all links for all wikis
14:40 <+tmortagne> that page could be aware of multiwiki concept, where is the API which require multiwiki information exactly ?
14:40 <+tmortagne> mflorea: no idea, was it a commit I did ?
14:40 <+tmortagne> does not remember anything like this
14:40 <+mflorea> I don't remember, nevermind, I'll keep searching
14:41 <vmassol> tmortagne: that has 2 issues. 1) we still have the script service in rendering which can be accessed and 2) it's costly to parse all references to filter them
14:41 <+tmortagne> mflorea: search is supposed to be case sensitive from what I know unless you did not properly configured MySQL database
14:41 <vmassol> one solution I can think of would be to have a system to inject the structure
14:41 <vmassol> ideally we'd need the resources module
14:42 <vmassol> (i think)
14:42 <+tmortagne> vmassol: I don't say the solution is to filter in that page I'm asking where the multiwiki concept is really required
14:42 <vmassol> tmortagne: IMO each wiki should only list the links for itself and the main wiki should list all links
14:43 <+tmortagne> you mean duplicate them ?
14:43 <vmassol> I don't understand
14:43 <+tmortagne> I tough you were talking about link storage
14:43 <vmassol> if the same link is broken in several places yes
14:44 <+tmortagne> vmassol: you could do what I do in extension manager
14:44 <+tmortagne> there is nowhere the notion of wiki in extension manager, I use a generic "namespace"
14:44 <vmassol> indeed
14:45 <vmassol> the only problem that remains is the parsing of the links found but we could use a generic resolver and have platform bring another one
14:45 <vmassol> (to extract the namespace)
14:46 <+tmortagne> hmm
14:47 <vmassol> another soltuion is to introduce a Reference with a type
14:47 <vmassol> and then have a PageReference in platform
14:47 <vmassol> and the livetable page would check the type and cast to PageReference
14:47 <vmassol> and in PageReference you'd have a getWiki()
14:47 <vmassol> i think this is our ResourceReference idea
14:48 <+tmortagne> not exactly
14:48 <+tmortagne> in ResourceReference you have information about the target
14:48 <+tmortagne> you want information about the source where the link is
14:49 <+tmortagne> i.e. you want to know where the broken link is
14:49 <vmassol> source where the link is is itself a ResourceRefrence
14:49 <+tmortagne> not yet
14:50 <+tmortagne> it's just a string identifier in the XDOM currently I think
14:54 <vmassol> yes in Rendering ResourceReference has a reference field and it's a string
14:55 <vmassol> we would need to have EntityReference a subclass of a more primitive type
14:55 <+tmortagne> or implementing thing in platform
14:55 <+tmortagne> s/thing/this/
14:55 <vmassol> this = ?
14:55 <vmassol> linkchecker,
14:55 <vmassol> ?
14:56 <+tmortagne> the catch of the links to check
14:56 <+tmortagne> instead of rendering
14:56 <+tmortagne> linckchecker store itself can stay in commons if it provide a concept of namespace
14:56 <vmassol> well if we do this we might as well move the whole thing to platform
14:56 <vmassol> if we add the concept of namespace then we're good
14:57 <+tmortagne> for the storage but rendering alone cannot extract the namespace
14:57 <vmassol> we just need to add an interface to extract the namespace from the string
14:57 <vmassol> the impl would be in platform indeed
14:57 <vmassol> with a default impl in rendering that does nothing
14:57 <vmassol> (empty namespace)
14:58 <vmassol> I was just thinking that this namespace need is more generic than just hte link checker
14:58 <vmassol> and this could fit in the resource module we talked about
14:58 <vmassol> provided we agree to add the concept of namespace in the resource module
14:58 <vmassol> anyway it's a bit too early for such a module but the interest is growing
14:59 <+tmortagne> it's more generic that lin kchcker for sure since it already exist in EM and multi CM as I told you
14:59 <vmassol> yep
14:59 <vmassol> do you have already interfaces for extracting namespaces somewhere?
14:59 <vmassol> NamespaceResolver or something like that
15:00 <+tmortagne> I never had the need to extract the namespace from a string
15:00 <vmassol> ok
15:00 <vmassol> well
15:00 <vmassol> it could be implemented at the level of the rendering api
15:00 <vmassol> in LinkBlock
15:00 <vmassol> I mean LinkBlock could return an object having the namespace already extracted
15:01 <vmassol> ie in ResourceReference
15:01 <vmassol> not sure if it's useful or not
15:01 <vmassol> since it would mean 2 parsing….
15:02 <vmassol> one for the namespace and another one when we fully resolve the link reference
15:02 <vmassol> (when we render)
15:02 <+tmortagne> not sure, I would prefer doing something for linkchecker for now I think
15:02 <vmassol> yep
15:02 <+tmortagne> too many stuff in the pipe to think carefuly about introducing something like that
15:02 <vmassol> we need to think more about it in general
15:03 <vmassol> yep
15:08 <vmassol> hmm the script service will not be able to return only links for the currnet namespace anyway since there's no notion of current namespace there....
15:08 <vmassol> so the script service would need to move to platform for that
15:09 <CIA-121> Marius Dumitru Florea master * ref69f8d / xwiki-platform-core/xwiki-platform-extension/xwiki-platform-extension-ui/src/main/resources/XWiki/AddExtensions.xml : Fix HTML markup. -
15:11 <vmassol> unless I invent a NamespaceValidator interface...
15:12 <vmassol> (implemented in platform)
15:12 <vmassol> NamespaceFilter rather than validator
15:12 <vmassol> (the default impl would consider everything ok)
15:12 <vmassol> all this is pretty complex....
15:13 <vmassol> the only alternative is to move it all to platform
15:13 <vmassol> but then people using our rendering only cannot benefit from it… it's a choice....
15:14 <+tmortagne> moving all this to platform sounds the easiest for now, we can always move back later when all is ready, we have this more generic than that in platform
15:14 <+tmortagne> s/this/things/
15:15 <vmassol> yeah I'm still hesitating
15:15 <vmassol> (it also means removing something that we made available in 3.3)
15:17 <vmassol> the good point is that the linkchecker is optional
15:18 <vmassol> and not active by default
15:18 <vmassol> so only the sysadmin can decide to activate it
15:18 <vmassol> (almost)
15:18 <vmassol> at least someone with PR
15:18 <vmassol> I'll add a warning in the RN
15:21 <vmassol> we could require farm admin rights to display the "external links" tab for now
15:27 <vmassol> ok I'm not going to work on this now but I've created an issue
15:38 <vmassol1> has joined #xwiki
15:39 <vmassol> has quit
15:55 <CIA-121> Marius Dumitru Florea master * r2207718 / (2 files): Fix access mode. -
16:03 <CIA-121> Marius Dumitru Florea master * rc550165 / xwiki-enterprise-test/xwiki-enterprise-test-rest/src/test/it/org/xwiki/test/rest/ : XE-1069: Skin and ColorTheme improvements: new colors, gradients, shadows, etc. ...
16:19 <CIA-121> tmortagne master * ra3cfaa4 / (4 files in 2 dirs): XWIKI-7348: Provide UI to import and syncronise an external extension in the repository -
16:33 <DrLou> has joined #xwiki
16:33 <vmassol1> I think we should be able to close wdyt?
16:34 <vmassol1> actually I asked the question in
16:35 <vmassol1> I'm closing it
17:18 <evalica> has quit
17:24 <Enygma`> has quit
17:38 <lucaa> has quit
17:46 <sdumitriu> has joined #xwiki
17:49 <jnsl_> has joined #xwiki
17:50 <jnsl_> how much memory dose xwiki generally use? I'm currently using confluence, but it takes 70% of all my memory so im looking for a less power hungry application
18:01 <sburjan> has joined #xwiki
18:08 <sburjan> has quit
18:25 <tmortagne> has quit
18:58 <+mflorea> sdumitriu: ping to read my last mail on devs. Please take some time to fix the remaining tests related to EM UI and Space Template so that I can release 3.4M1 tomorrow morning. Thanks
18:59 <@sdumitriu> OK mflorea
19:00 <mflorea> has quit
19:12 <gdelhumeau> has quit
19:19 <jvdrean> has quit
20:04 <CIA-121> Sergiu Dumitriu master * r6a86e93 / (5 files in 2 dirs): XWIKI-7246: Add ability to browse all available extensions ...
20:13 <CIA-121> Thomas Mortagne master * r45ccbeb / xwiki-platform-core/xwiki-platform-oldcore/src/main/java/com/xpn/xwiki/web/ : Revert again -
20:30 <mflorea> has joined #xwiki
20:56 <SvenDowideit> has quit
20:57 <SvenDowideit> has joined #xwiki
21:14 <abusenius> has joined #xwiki
21:56 <rrodriguez> has joined #xwiki
22:30 <rrodriguez> has quit
22:31 <rrodriguez> has joined #xwiki
22:33 <rrodriguez> has quit
22:35 <rrodriguez> has joined #xwiki
22:42 <abusenius> has quit
22:50 <lucaa> has joined #xwiki
22:58 <lucaa> has quit
23:12 <DrLou> has quit
23:20 <DrLou_> has joined #xwiki
23:23 <DrLou_> has left #xwiki
23:25 <DrLou_> has joined #xwiki
23:26 <DrLou_> has left #xwiki
23:43 <mflorea> has quit
23:55 <SvenDowideit> has quit
23:55 <SvenDowideit> has joined #xwiki

Get Connected