IRC Archive for channel #xwiki on 05 September 2008

vmassol - (05/09/2008 11:50): jvdrean: hudson is still unstable :(
cypromis - (05/09/2008 11:51): what problem do you have with hudson ?
jvdrean - (05/09/2008 11:51): vmassol: again the svn up problem
cypromis - (05/09/2008 11:53): you should recreate your subversion repository
cypromis - (05/09/2008 11:53): and make sure you use fsfs
cypromis - (05/09/2008 11:53): otherwise it locks up with hudson
cypromis - (05/09/2008 11:53): I had that too and that solved my problem
cypromis - (05/09/2008 11:54): had the same problem with my subversion mirrors
jvdrean - (05/09/2008 11:56): xipe: is it something we could think about ? :)
jvdrean - (05/09/2008 11:57): vmassol: I remember now (a bit late) why I had separated enterprise from enterprise-test .. because of the installers vs. maven 2.0.9 issue
vmassol - (05/09/2008 11:59): oh we're not using fsfs?
vmassol - (05/09/2008 11:59): that's bad I guess... I have heard lots of issues when not using fsfs in the past AFAIR
vmassol - (05/09/2008 11:59): jvdrean: can you check with xipe what we are using and if it's not fsfs create a jira issue for him in SysAdmin jira project?
vmassol - (05/09/2008 11:59): thanks cypromis for the idea :)
cypromis - (05/09/2008 12:10): if you have it ready let me know
cypromis - (05/09/2008 12:10): and I do a german mirror with a git gateway
cypromis - (05/09/2008 12:10): svnsync rocks
vmassol - (05/09/2008 12:22): nice
vmassol - (05/09/2008 12:36): lunch time, bbiab
jvdrean - (05/09/2008 12:38): vmassol: I might be wrong but I feel like being extensible will always have a cost
jvdrean - (05/09/2008 12:39): with our generic properties tables we'll always have to make joins and I they'll always be slower than a single table query
jvdrean - (05/09/2008 12:39): no ?
vmassol - (05/09/2008 13:05): jvdrean: doing a join is not slower normally AFAIK (provided the right indexes are in place)
vmassol - (05/09/2008 13:06): jvdrean: being extensible always has a cost but we've already decided we wanted to be extensible in xwiki
jvdrean - (05/09/2008 13:06): vmassol: indexes that we can't provide by default AFAIK
vmassol - (05/09/2008 13:06): that's why we have this notion of objects
vmassol - (05/09/2008 13:07): jvdrean: we can provide them
jvdrean - (05/09/2008 13:07): vmassol: sure but being extensible doesn't imply that we use those mechanism everywhere
vmassol - (05/09/2008 13:07): for all our standard objects at least
vmassol - (05/09/2008 13:07): but in any case we can provide them for all I think since we have generic tables, don't we?
vmassol - (05/09/2008 13:08): anyway that's a different discussion
jvdrean - (05/09/2008 13:08): vmassol: different discussion that has been discussed in the past
jvdrean - (05/09/2008 13:08): vmassol: IIRW Sergiu looked at how we could provide indexes with hibernate and it was not possible (at least ftm)
jvdrean - (05/09/2008 13:08): anyway
jvdrean - (05/09/2008 13:09): the discussion is about "should anything be an object" :)
vmassol - (05/09/2008 13:09): not anything :)
jvdrean - (05/09/2008 13:09): XWiki.XWikiAuthor, XWiki.XWikiLastModification
vmassol - (05/09/2008 13:09): but metadata for documents
jvdrean - (05/09/2008 13:09): XWiki.XWikiHidden
jvdrean - (05/09/2008 13:09): :)
vmassol - (05/09/2008 13:10): right now we're not logical at all
vmassol - (05/09/2008 13:10): some are objects some are not
vmassol - (05/09/2008 13:10): and there's no real logic
vmassol - (05/09/2008 13:11): jvdrean: I think conceptually we want objects for all doc metadata. Now we just need to check if we can achieve this or not
vmassol - (05/09/2008 13:11): do you agree?
jvdrean - (05/09/2008 13:13): theoriticaly yes :)
jvdrean - (05/09/2008 13:14): I may suffer from the 
jvdrean - (05/09/2008 13:14): "fear of change"
vmassol - (05/09/2008 13:15): well there isn't much change
vmassol - (05/09/2008 13:15): since this is already something we do
vmassol - (05/09/2008 13:15): anyway right now we just need to decide for the hidden metadata not for all
jvdrean - (05/09/2008 13:16): yes, and we know the drawbacks of putting stuff in xclasses/xobjects
jvdrean - (05/09/2008 13:16): painful updates for example
vmassol - (05/09/2008 13:16): wdym?
jvdrean - (05/09/2008 13:18): I mean hardcoded classes / objects in java
jvdrean - (05/09/2008 13:18): s/in java/creation and updates/
jvdrean - (05/09/2008 13:18): ok let me rephrase
jvdrean - (05/09/2008 13:18): I mean hardcoded classes/objects creation and updates in our java code
vmassol - (05/09/2008 13:19): you mean the answer to the 3 questions I've asked?
vmassol - (05/09/2008 13:19): - should they be created automatically by xwiki core if they don't exist (class + object),
vmassol - (05/09/2008 13:19): - should they not be created by have default values configured in xwiki's configuration
vmassol - (05/09/2008 13:19): - should they be created on first query if not found?
vmassol - (05/09/2008 13:19): this?
jvdrean - (05/09/2008 13:20): yes
jvdrean - (05/09/2008 13:20): but more generally I don't like the fact that we have java code depending on wiki data
jvdrean - (05/09/2008 13:20): like our tag feature
vmassol - (05/09/2008 13:21): yep, me neither
jvdrean - (05/09/2008 13:21): and I wouldn't like to see more and more of this
vmassol - (05/09/2008 13:21): this is where I don't agree
vmassol - (05/09/2008 13:21): for me either we don't have any or we have all of them 
vmassol - (05/09/2008 13:22): it's ok for me to have them separated
vmassol - (05/09/2008 13:22): since this exactly what configuration is
vmassol - (05/09/2008 13:22): we just need some internal way to represent defaults
vmassol - (05/09/2008 13:22): and only have the data be overrides
vmassol - (05/09/2008 13:30): thomasmortagne: have you been able to start on the table implementation?
vmassol - (05/09/2008 13:30): (said differently are you confident for having it in 1.6M2?)
thomasmortagne - (05/09/2008 13:32): vmassol: not yet as I have issue with ldap tests I don't understand (it's working well in runtime), I was about to comment the faulty unit test and go to table
thomasmortagne - (05/09/2008 13:32): so i'm not sure about the time I need to finish table
vmassol - (05/09/2008 13:34): I think you need 1 day max for the simplest possible implementation
thomasmortagne - (05/09/2008 13:34): ok
vmassol - (05/09/2008 13:41): thomasmortagne: we need to introduce the generic notion of parameters in the new rendering too. I'll work on that
vmassol - (05/09/2008 13:41): (this is what wikimodel calls WikiParameters fyi)
vmassol - (05/09/2008 13:41): for example look at http://code.google.com/p/wikimodel/wiki/WikiSyntaxCommon
vmassol - (05/09/2008 13:41): in common syntax it's {{param1=value1 param2=value2}}
vmassol - (05/09/2008 13:41): in our syntax it's (% param1=value... %)
vmassol - (05/09/2008 13:42): we could generalize it to lots of events as wikimodel is doing
vmassol - (05/09/2008 13:42): paragraphs, lists, tables, etc
thomasmortagne - (05/09/2008 13:42): yes could be usefull
vmassol - (05/09/2008 13:42): I'm just worried about the doxia bridge
vmassol - (05/09/2008 13:43): I need to check what doxia event we would generate for those
thomasmortagne - (05/09/2008 13:43): anyway we need to do it at least for table, at worst we could generate nothing for doxia, right ?
vmassol - (05/09/2008 13:44): right
vmassol - (05/09/2008 13:44): but I think it'll be better than taht
vmassol - (05/09/2008 13:44): since doxia now has the notion of SinkEventAttributes
vmassol - (05/09/2008 13:45): which is very similar
vmassol - (05/09/2008 13:45): so that should be good
vmassol - (05/09/2008 13:45):  * All sink methods that produce some presentation-level output should have at least
vmassol - (05/09/2008 13:45):  * one form that allows to pass in a Set of SinkEventAttributes. For instance in
vmassol - (05/09/2008 13:45):  * <pre>void text( String text, SinkEventAttributes attributes );</pre>
vmassol - (05/09/2008 13:45):  * the <code>attributes</code> parameter can be used to specify some text styling
vmassol - (05/09/2008 13:45):  * options, or other optional parameters.
vmassol - (05/09/2008 13:45): so that's cool :)
thomasmortagne - (05/09/2008 13:47): ok
xwikibot - (05/09/2008 15:05): Plugins.WysiwygPlugin was modified by xwiki:XWiki.jvdrean (deleted document) - http://code.xwiki.org/xwiki/bin/view/Plugins/WysiwygPlugin
xwikibot - (05/09/2008 15:06): Plugins.WysiwygPluginDownloads was modified by xwiki:XWiki.jvdrean (deleted document) - http://code.xwiki.org/xwiki/bin/view/Plugins/WysiwygPluginDownloads
vmassol - (05/09/2008 15:30): does someone know how to test if there's a next element when using JDK 1.5 for(Object o: ...) construct? I keep bumping into this and using the old Iterator and while() loop
lucaa - (05/09/2008 15:33): vmassol: it might not be possible and it kindof makes sense not to be able to: for( ) is supposed to be syntactic sugar to give you each element one by one, if you want "advanced features", you need to get the iterator 
lucaa - (05/09/2008 15:33): I usually do iterators in for () to make it nicer syntax
vmassol - (05/09/2008 15:34): yeah I don't think it's possible but I thought I should ask  just in case...
_REM_ - (05/09/2008 15:45): hi.  i am trying to ProxyPass Xwiki behind Apache and while it does proxy the page, it doesn't load the images.  I have proxypass working with other sites.  can anyone poing me in the right direction??
vmassol - (05/09/2008 15:52): well if you had stayed a few minutes longer you would have had the answer...
vmassol - (05/09/2008 15:52): ah those users who want an immediate answers to all their needs ;)
vmassol - (05/09/2008 15:52): for anyone listening here's the answer: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Performances
thomasmortagne - (05/09/2008 16:02): vmassol: what about moving XWikiSyntaxRenderer in org.xwiki.rendering.renderer.xwiki like XHTML renderer is in org.xwiki.rendering.renderer.xhtml ?
vmassol - (05/09/2008 16:03): hmmm
vmassol - (05/09/2008 16:03): I'd prefer the following:
vmassol - (05/09/2008 16:04): move the classes in .xhtml package to the internal package since they are not public I think
vmassol - (05/09/2008 16:04): and move back the XHTMLRenderer in the .renderer package
vmassol - (05/09/2008 16:04): (thomasmortagne)
thomasmortagne - (05/09/2008 16:04): yes better
vmassol - (05/09/2008 16:07): thomasmortagne: and btw I don't knowhow to style only a word in a given paragraph using wikimodel
vmassol - (05/09/2008 16:07): I've asked mikhail but no answer so far...
vmassol - (05/09/2008 16:08): I'll send a mail on the wikimodel list
thomasmortagne - (05/09/2008 16:08): ok
vmassol - (05/09/2008 16:23): I'm committing the support for parametrized paragraphs
vmassol - (05/09/2008 16:32): done
jvelociter - (05/09/2008 17:15): Hello guys
jvelociter - (05/09/2008 17:15): Did any of you recently managed to debug selenium tests in eclipse ?
jvelociter - (05/09/2008 17:16): I keep having a NPE because the selenium object is not set to anything
jvelociter - (05/09/2008 17:16): It used to work just fine, but I wonder if some recent changes to the framework could have led to this
jvdrean - (05/09/2008 17:17): jvdrean: reminds me a recent discussion about ldap tests :)
jvelociter - (05/09/2008 17:17): http://pastebin.com/m3d433ae2
jvdrean - (05/09/2008 17:17): jvelociter
jvelociter - (05/09/2008 17:17): :)
jvdrean - (05/09/2008 17:17): grrr ;D
jvelociter - (05/09/2008 17:17): you are lucky, jvanzyl is not there ;)
jvelociter - (05/09/2008 17:19): vmassol: thomasmortagne any clue? have you debug selenium tests recently ?
vmassol - (05/09/2008 17:20): jvelociter: yes I know why
vmassol - (05/09/2008 17:20): it's because the selenium server is now started only once
vmassol - (05/09/2008 17:20): and we do this with a test setup or test suite (don't remember which one off the top of my head)
thomasmortagne - (05/09/2008 17:21): jvelociter: you means by run maven install on Eclipse or run as junit tests ?
vmassol - (05/09/2008 17:21): and I guess you didn't use it...
jvelociter - (05/09/2008 17:21): thomasmortagne: debug with the method written on xwiki.org
jvelociter - (05/09/2008 17:22): vmassol: what do you mean I'm not using it ?
jvelociter - (05/09/2008 17:22): It fails for XE tests for me as well
vmassol - (05/09/2008 17:22): no XE tests work fine
vmassol - (05/09/2008 17:22): :)
jvelociter - (05/09/2008 17:22): humm..
vmassol - (05/09/2008 17:23): if they fail from Eclipse it's because you're not running the tests properly in eclipse
vmassol - (05/09/2008 17:23): as I said your tests are not wrapped with this test setup/test suite
thomasmortagne - (05/09/2008 17:23): jvelociter: I don't see what methods you speak of
vmassol - (05/09/2008 17:23): because I guess you're not running the AllTests class
jvelociter - (05/09/2008 17:24): thomasmortagne: last bullet here http://dev.xwiki.org/xwiki/bin/view/Community/Testing#HFunctionalTesting
jvelociter - (05/09/2008 17:24): I'm debugging a single test class
jvelociter - (05/09/2008 17:24): as in "find the test you wish to run, put breakpoints and run it in debug mode"
jvelociter - (05/09/2008 17:24): am I doing wrong ?
vmassol - (05/09/2008 17:24): jv new XWikiSeleniumTestSetup(new XWikiTestSetup(suite));
vmassol - (05/09/2008 17:25): jvelociter:  new XWikiSeleniumTestSetup(new XWikiTestSetup(suite));
vmassol - (05/09/2008 17:25): this in AllTests.java
jvelociter - (05/09/2008 17:25): ok, but I still don't see the point
thomasmortagne - (05/09/2008 17:25): it should work, it works me me
thomasmortagne - (05/09/2008 17:25): at least for ldap tests
thomasmortagne - (05/09/2008 17:25): but i'm running maven install with M2Eclipse
jvelociter - (05/09/2008 17:25): I did that in XWS, as it's done in XE, but I can't debug either XE's tests or XWS ones
vmassol - (05/09/2008 17:26): jvelociter: what don't you understand?
jvelociter - (05/09/2008 17:26): maybe I'm doing something wrong
vmassol - (05/09/2008 17:26): yes
vmassol - (05/09/2008 17:26): yes
vmassol - (05/09/2008 17:26): yesy
vmassol - (05/09/2008 17:26): :)
thomasmortagne - (05/09/2008 17:26): you diod not forget <background>false</background> ?
jvelociter - (05/09/2008 17:26): I don't understand the point in "new XWikiSeleniumTestSetup(new XWikiTestSetup(suite));" regarding the debug method on xwiki.org
vmassol - (05/09/2008 17:26): ok let me explain then
jvelociter - (05/09/2008 17:26): (I'm trying to debug XE tests)
jvelociter - (05/09/2008 17:27): not XWS
vmassol - (05/09/2008 17:27): XWikiTestSetup: this starts and stops xwiki
vmassol - (05/09/2008 17:27): XWikiSeleniumTestSetup: this starts and stops the selenium proxy
vmassol - (05/09/2008 17:27): so if you start xwiki manually that's ok but you still haven't started the selenium proxy
jvelociter - (05/09/2008 17:27): thomasmortagne: no, I did it
vmassol - (05/09/2008 17:27): so you'll get a NPE
thomasmortagne - (05/09/2008 17:28): vmassol: he does not start manually, he is using mvn install
jvelociter - (05/09/2008 17:28): vmassol: ok, now I understand that http://dev.xwiki.org/xwiki/bin/view/Community/Testing#HFunctionalTesting is not enough any longer :)
vmassol - (05/09/2008 17:28): yes it needs to be updated 
vmassol - (05/09/2008 17:28): we need to find how to do it in IDEs now
jvelociter - (05/09/2008 17:28): so how do we do to debug single tests ?
vmassol - (05/09/2008 17:28): no idea
vmassol - (05/09/2008 17:28): :)
jvelociter - (05/09/2008 17:28): ah, we don't
jvelociter - (05/09/2008 17:28): :)
vmassol - (05/09/2008 17:28): right now
vmassol - (05/09/2008 17:28): what you can do
jvelociter - (05/09/2008 17:29): or we comment in AllTests
vmassol - (05/09/2008 17:29): is have a very single MyTests class
vmassol - (05/09/2008 17:29): similar as AllTests
jvelociter - (05/09/2008 17:29): all the tests we don't want
jvelociter - (05/09/2008 17:29): yep
vmassol - (05/09/2008 17:29): yes...
vmassol - (05/09/2008 17:29): you might want to remove the XWikiTestSetup
vmassol - (05/09/2008 17:29): if you have already starte xwiki manually
jvelociter - (05/09/2008 17:30): and put "new XWikiSeleniumTestSetup(suite)" ?
vmassol - (05/09/2008 17:30): yes
jvelociter - (05/09/2008 17:30): ok
jvelociter - (05/09/2008 17:30): will try that
jvelociter - (05/09/2008 17:30): thanks !
vmassol - (05/09/2008 17:31): jvelociter: 
vmassol - (05/09/2008 17:31): what you can also do
vmassol - (05/09/2008 17:31): is inside a given tescase class
vmassol - (05/09/2008 17:31): add new XWikiSeleniumTest(suite)
vmassol - (05/09/2008 17:31): in the suite() method
vmassol - (05/09/2008 17:31): and just run that test case in eclipse
jvelociter - (05/09/2008 17:32): ah ok
jvelociter - (05/09/2008 17:32): I see
jvelociter - (05/09/2008 17:33): but it's not good for the whole all tests, to have this
vmassol - (05/09/2008 17:33): I recommend the excellent book named "JUnit in Action" to understand all this ;)
jvelociter - (05/09/2008 17:33): :D
vmassol - (05/09/2008 17:33): not it's not
jvelociter - (05/09/2008 17:33): of course :)
vmassol - (05/09/2008 17:33): since the goal of this was to run the selenium proxy only once
jvelociter - (05/09/2008 17:33): ok, I get it
jvelociter - (05/09/2008 17:34): we should modify the doc until we find a solution
vmassol - (05/09/2008 17:34): yes
vmassol - (05/09/2008 17:34): if you succeed could you modify it?
jvelociter - (05/09/2008 17:34): yes
vmassol - (05/09/2008 17:34): thanks
jvelociter - (05/09/2008 17:43): Hehe, Thanks to jvdrean...
jvelociter - (05/09/2008 17:44): he's devising original threatening methods
xwikibot - (05/09/2008 17:59): AdminGuide.InstallationDB2 was modified by xwiki:XWiki.jvdrean (The file xwiki.db2.hbm.xml is shipped with XWiki) - http://platform.xwiki.org/xwiki/bin/view/AdminGuide/InstallationDB2
xwikibot - (05/09/2008 18:00): AdminGuide.InstallationMSSQL was modified by xwiki:XWiki.jvdrean (The file xwiki.mssql.hbm.xml is shipped with XWiki) - http://platform.xwiki.org/xwiki/bin/view/AdminGuide/InstallationMSSQL
xwikibot - (05/09/2008 18:26): AdminGuide.InstallationMSSQL was modified by xwiki:XWiki.jvdrean (Removed deprecated section) - http://platform.xwiki.org/xwiki/bin/view/AdminGuide/InstallationMSSQL