IRC Archive for channel #xwiki

Last modified by Vincent Massol on 2012/10/18 19:12

sdumitriu - (00:00): K
sburjan_ joined #xwiki at 00:01
bblfish joined #xwiki at 00:02
sburjan` left at 00:04 (Ping timeout: 265 seconds
sburjan_ is now known as sburjan` ([email protected]
bblfish left at 00:08 (Quit: Leaving.
bblfish joined #xwiki at 00:45
bblfish left at 00:54 (Ping timeout: 240 seconds
sburjan_ joined #xwiki at 01:36
lucaa left at 01:36 (Ping timeout: 250 seconds
sburjan` left at 01:36 (Ping timeout: 276 seconds
sburjan_ is now known as sburjan` ([email protected]
evalica joined #xwiki at 02:00
evalica left at 02:17 (Read error: Connection reset by peer
sdumitriu left at 02:59 (Quit: Leaving.
abusenius left at 04:30 (Quit: Konversation terminated!
Baur joined #xwiki at 05:19
venkatesh joined #xwiki at 06:19
venkatesh left at 06:20 (Client Quit
venkatesh joined #xwiki at 06:20
Baur - (06:29): ping
Baur - (06:30): cjdelisle ping
Baur - (06:38): hello
Baur - (06:38): anybody is here?
Denis left at 07:01 (Read error: Connection reset by peer
Baur - (07:07): {{velocity}}$xwiki.jsx.use("XWiki.SomeJSfile") {{/velocity}} does not work
Baur - (07:08): JS script is not called
Baur - (07:08): any idea?
Denis joined #xwiki at 07:14
Baur - (07:14): ?
gsauthier left at 07:20 (Ping timeout: 272 seconds
Baur - (07:31): xipe ping
Baur - (07:31): npm ping
rrodriguez joined #xwiki at 07:44
asrfel joined #xwiki at 08:34
mflorea joined #xwiki at 08:52
silviar joined #xwiki at 08:55
silviar left at 08:57 (Client Quit
silviar joined #xwiki at 09:01
rrodriguez left at 09:08 (Quit: rrodriguez
vmassol joined #xwiki at 09:08
vmassol left at 09:09 (Client Quit
vmassol joined #xwiki at 09:09
vmassol1 joined #xwiki at 09:12
vmassol left at 09:14 (Ping timeout: 252 seconds
sburjan_ joined #xwiki at 09:27
asrfel left at 09:27 (Read error: Connection reset by peer
sburjan` left at 09:31 (Ping timeout: 264 seconds
sburjan_ is now known as sburjan` ([email protected]
asrfel joined #xwiki at 09:31
arkub joined #xwiki at 09:31
asrfel left at 09:31 (Client Quit
asrfel joined #xwiki at 09:32
rrodriguez joined #xwiki at 09:33
sburjan joined #xwiki at 09:34
rrodriguez left at 09:34 (Client Quit
vmassol1 - (09:36): good morning guys
vmassol1 - (09:37): we need to finish fixing quickly those failing tests
vmassol1 - (09:37): there are even more than yesterday :(
vmassol1 - (09:37):
vmassol1 - (09:38): sburjan: could you fix this one:$xwiki-enterprise-test-ui/72/testReport/ ?
sburjan - (09:40): vmassol1: I will try. But I ran that test on 3 different machines, without a failure. It fails only on Hudson. i guess it's a flickering
vmassol1 - (09:40): flickering must be fixed at all cost
florinciu joined #xwiki at 09:40
vmassol1 - (09:41): (it's usually not too hard to fix since it means some ajax and a missing wait in the test)
sburjan - (09:41): I know .. I am investigating now
vmassol1 - (09:42): we'"re really really bad at functional testing, we need to improve. I'll send a mail with a proposal later on
sburjan - (09:42): although It will be a blind fix .. because it happens only on hudson
sburjan - (09:42): okay .. trying to fix that test
vmassol1 - (09:42): it takes us over 3 weeks to stabilize a release that's definitely not normal
vmassol1 - (09:43): and we systematically slip by one week even though we have planned 2 weeks for stabilization so something is dead wrong
vmassol1 - (09:43): we should stabilize in less than 2 days
sburjan - (09:44): yes ..
bblfish joined #xwiki at 09:46
silviar left at 09:52 (Quit: Leaving.
sburjan - (09:53): Hop I will be able to catch the flickering .. I am a little nervous about this
sburjan - (09:53): never did this before ... especially not blindly
sburjan - (09:54): vmassol1: ui-tests on hudson are launched using mvn surefire:test or mvn clean install ?
sburjan - (09:54): because on my personal machine, I get different results when running these 2 commands
sburjan - (09:54): and it seems odd
vmassol1 - (09:56): you should not call surefire:test
Baur left at 09:56 (*.net *.split
npm left at 09:56 (*.net *.split
vmassol1 - (09:56): you can call mvn test but this won't run integration tests, the phase runs after
silviar joined #xwiki at 09:59
vmassol1 - (09:59): mflorea: fyi there's a warning in the gwt dom project build: [WARNING] You should not declare gwt-dev as a project dependency. This may introduce complex dependency conflicts
tmortagne joined #xwiki at 09:59
mflorea left at 10:00 (Ping timeout: 264 seconds
jvdrean joined #xwiki at 10:05
asrfel left at 10:06 (*.net *.split
arkub left at 10:06 (*.net *.split
Denis left at 10:06 (*.net *.split
DragosRusu left at 10:06 (*.net *.split
cjdelisle left at 10:06 (*.net *.split
sburjan` left at 10:06 (*.net *.split
nuvolari left at 10:06 (*.net *.split
sburjan left at 10:06 (*.net *.split
venkatesh left at 10:06 (*.net *.split
sburjan joined #xwiki at 10:11
sburjan - (10:11): god damn splits
sdumitriu joined #xwiki at 10:12
sdumitriu - (10:15): Hi guys
sdumitriu - (10:15): Starting the release
vmassol1 - (10:15): good morning sdumitriu
mflorea joined #xwiki at 10:16
vmassol1 - (10:16): sdumitriu: you sure? there are some failing tests, isn't that a pb?
sdumitriu - (10:16): Let me check
sburjan - (10:17): although I see that mvn clean install runs the tests with surefire
vmassol1 - (10:17): sburjan: you need to read up on maven phases
vmassol1 - (10:17): let me get you the url
vmassol1 - (10:17): here:
sburjan - (10:18): thanks.
vmassol1 - (10:21): sburjan: when you call mvn install it'll run all the phases that come before the install phase and hence all the plugins that have mojos bound to those phases will get executed too, whereas if you run mvn surefire:test only the test mojo of the surefire plugin will execute, nothing else
vmassol1 - (10:22): (when you choose a projet's packaging - ie <packaging> in the POM) then you get automatically a mapping between phases and mojos)
sburjan - (10:23): I understand
sburjan - (10:24): thanks you ... now I understand
sburjan - (10:24): *thanks
sburjan - (10:24): *thank
tmortagne1 joined #xwiki at 10:26
sburjan - (10:27): I run mvn clean install from enterprise/trunk .. done ... run mvn clean install from ui-tests .. ran all the tests. I still got different results as hudson . And my failing tests that fails on hudson, passes on local machine. I will have to try to fix this
tmortagne left at 10:28 (Ping timeout: 240 seconds
evalica joined #xwiki at 10:31
sdumitriu - (10:33): These work locally:$xwiki-enterprise-test-ui/72/testReport/
sdumitriu - (10:34): This failed because of this log message: INFO:  seeing JVM BUG(s) - cancelling interestOps==0
sdumitriu - (10:34):$xwiki-enterprise-test-webstandards/72/testReport/
sburjan - (10:36): so they pass on you local machine too ?
sburjan - (10:36): I am interested in UserProfileTest
sdumitriu - (10:36): Yes, they pass
sburjan - (10:37): so what can I do so UserProfileTest to pass on Hudson too ?
vmassol1 - (10:37): find why they don't pass and fix it :)
sburjan - (10:38): sad part is that I can't reproduce locally :)
vmassol1 - (10:38): look at the error
vmassol1 - (10:38): read the test code
vmassol1 - (10:38): deduce how the error can happen
lucaa joined #xwiki at 10:39
vmassol1 - (10:42): fyi I've generated a clover report for the whole platform and the TPC is 28%. Now adding functional tests to see how it improves
sburjan - (10:43): fingers crossed
sburjan - (10:43): :)
vmassol1 - (10:43): so this means our unit tests only cover 28%
vmassol1 - (10:43): that's quite low
mflorea - (10:49): sdumitriu: the content edit menu is displayed above the WYSIWYG editor dialogs, probably a big z-index. Should we fix this on the WYSIWYG stylesheet or in the XE skin?
sdumitriu - (10:50): WYSIWYG
mflorea - (10:50): ok, I'll fix it
sdumitriu - (10:51): sburjan: I think that this line fails: getDriver().switchTo().frame(2);
sdumitriu - (10:51): In ProfileEditPage#setUserAddress
sburjan - (10:54): I will try to fix this so I won't have to use that
DragosRusu joined #xwiki at 10:59
nuvolari joined #xwiki at 10:59
cjdelisle joined #xwiki at 10:59
venkatesh joined #xwiki at 10:59
arkub joined #xwiki at 10:59
asrfel joined #xwiki at 10:59
Enygma` joined #xwiki at 10:59
Baur joined #xwiki at 10:59
npm joined #xwiki at 10:59
Baur - (11:20): ping
Baur - (11:20):  $escapetool.xml(${pdoc.getRenderedTitle('plain/1.0')}) » $escapetool.xml(${pdoc.getRenderedTitle('plain/1.0')}) » ???? ???????? Onlinebank
Baur - (11:21): for some reason breadcrumbs does not appera correct
cjdelisle - (11:21): are you getting things like &#123;
cjdelisle - (11:22): ?
Baur - (11:24): nope, for some reason if take away for unregistered user to have right for viewing and commenting for XWIKI page
Baur - (11:24): the breadcrumb shows incorrectly
Baur - (11:24): I returned rights to what it was before
Baur - (11:24): and it is ok now
Baur - (11:25): i guess it is a bug
cjdelisle - (11:25): gueses need view access on the XWiki space, that's all I can think of.
cjdelisle - (11:25): *guests
Baur - (11:25): i  can not give that access right for unregistered users
Baur - (11:25): cause they can change things later
cjdelisle - (11:26): just view, not edit.
Baur - (11:26): well from xwiki space they can have access to any other page, right?
cjdelisle - (11:28): No. giving guests permission to view in XWiki space only means they can view in the XWiki space.
tmortagne1 - (11:28): Baur: XWiki space is just one space among others, when you can view it there is no reason you can view anything else
Baur - (11:32): well, when i say xwiki space i mean main space, and it contains things like sections and recent changes, and unregistered user can go to any space through the link, right?
tmortagne1 - (11:34): main space is Main (which doe snot contains recent changed but a page which shows recent changes) not XWiki and anyway clicking on a link does not give you any right to that link target
Baur - (11:37): ok, here is what i did
Baur - (11:37): when i type localhost:8080
Baur - (11:37): i go directly to my blog space
Baur - (11:37): there are some information for unregistered users
Baur - (11:38): so, if I click XWIKI link whis is located at the top left corner
Baur - (11:38): i get to the Main space
Baur - (11:38): where I can see sections and recent changes
Baur - (11:39): these links give me access to the pages where recent changes were made
Baur - (11:39): so there i can make changes to that pages
Baur - (11:39): and this is not good
cjdelisle - (11:42): ``these links give me access to the pages'' The links don't give you the access (although that's an interesting concept)  The links only point to the pages and if you didn't have access, you would get an access denied message.
tmortagne1 - (11:42): Baur: if you want to deny right for guess you need to do it in the admin, just hide links is useless
tmortagne1 - (11:43): anyone can go directly to the page typing the URL
Baur - (11:43): so you are saying that i have to adjust access rights in =every space and page
tmortagne1 - (11:44): not in every pages
tmortagne1 - (11:44): page rights inherit from space rights which inherits from wiki rights
tmortagne1 - (11:45): so just remove view right of guest at wiki level and set it for spaces you want to make public
tmortagne1 - (11:45): like Blog and XWiki here if i understood well
Baur - (11:46): the idea is that the unregistered users should have rights to view and comment on the Blog space or page, except that they should not have right to access any other space or page
Baur - (11:46): that is what i wanted to have
sdumitriu - (11:46): Then change the parent
tmortagne1 - (11:48): Baur: that's exatly what i just explained to you: remove any guest right at wiki level in the admin and give him view and comment right in Blocg space
tmortagne1 - (11:48): s/Blocg/Blog/
tmortagne1 - (11:48): and maybe XWiki space for technical reason i did not followed your dicussion with cjdelisle
sdumitriu - (11:48): Hm, actually the breadcrumbs should be OK
sdumitriu - (11:48): What version are you using?
Baur - (11:49): ok, when i remove view right from xwiki main space, the breadcrumb on the Blog space is dispayed incorrectly
Baur - (11:49): wait a minute lemme show you
Baur - (11:49):  $escapetool.xml(${pdoc.getRenderedTitle('plain/1.0')}) » $escapetool.xml(${pdoc.getRenderedTitle('plain/1.0')}) » ???? ???????? Onlinebank
sdumitriu - (11:50): Yes, seen that
Baur - (11:50): the breadcrumb is displaied this way
sdumitriu - (11:50): But it should be fixed
Baur - (11:50): that is what I am saying , it is kind of bug or error
Baur - (11:50): that is all I wanned to say
Baur - (11:50): :D
sdumitriu - (11:50): XWIKI-5362
sdumitriu - (11:51): It was fixed for 2.5M1
Baur - (11:51): well I am using 2.4.3 version
Baur - (11:54): so should I import this 2.5M1 version or what
sdumitriu - (11:54): You can use this file:
Baur - (11:55): shoudl I just copy and paste this code to hierarchy.vm file?
sdumitriu - (11:55): Yes
Baur - (11:56): ok
Baur - (11:57): now it shows only this:  Main.WebHome » ???? ???????? Onlinebank
sdumitriu - (11:58): Just remove the parent from Blog.WebHome
venkatesh left at 11:58 (Ping timeout: 240 seconds
Baur - (11:58): i guess there shoul be something different from Main.WebHome
fmancinelli joined #xwiki at 11:59
cjdelisle - (12:01): A lot of tests seem to be flickering now which didn't in the past.
Denis joined #xwiki at 12:01
Baur - (12:05): how can i remove parent from Blog.WebHome
sdumitriu - (12:07): Edit it
sdumitriu - (12:07): There should be a Parent field in the right panel
sdumitriu - (12:07): Delete the value from there
sdumitriu - (12:07): Save
cjdelisle - (12:36): -Xmx1024 just uploaded and downloaded xwiki-enterprise-web-2.5-SNAPSHOT.war (98MB)  I'd say my efforts paid off.
cjdelisle - (12:37): One more trick up my sleeve which I haven't compiled yet.
rstavro joined #xwiki at 12:37
vmassol1 - (12:39): ui-tests brings TPC from 23% to 38.6%
cjdelisle - (12:39): I'm surprised, I expected ui-tests would bring it up more.
vmassol1 - (12:39): (I have 7 test failures btw in ui-tests)
vmassol1 - (12:39): ui-tests is quite light
vmassol1 - (12:39): now I'll run selenium-tests
cjdelisle - (12:40): I would have thought anything which runs the wiki will run through a vast majority of all the code.
cjdelisle - (12:40): This means there is code which rarely runs. Possibly a place for optimization.
vmassol1 - (12:41): I'll publish the full report once I'm done
vmassol1 - (12:41): and we can then start analyzinf it
vmassol1 - (12:42): the good news is that I'm now able to run clover on all
sburjan - (12:42): 7 fails ? Now hudson reports only one.
vmassol1 - (12:42): I'll document that
cjdelisle - (12:42): Does it list the names or code lines of the statements which run vs the statements which don't?
vmassol1 - (12:42): yes
cjdelisle - (12:43): The ones which run (if there are that few) should be our #1 target for speed optimization.
cjdelisle - (12:43): yay!
vmassol1 - (12:43): cjdelisle: you can check to familiarize yourself with the report
cjdelisle - (12:43): thanks
vmassol1 - (12:44): it's also interesting to see test code that does not run
sburjan - (12:45): I am doing this already :) Installed clover plugin on my eclipse
vmassol1 - (12:45): sburjan: and you're able to generate the full TPC for it all? what value do you get?
sdumitriu - (12:45): vmassol1: So you don't want me to release the root pom, nor the tools, right?
sburjan - (12:45): haven't ran it against all ui-tests
vmassol1 - (12:46): sdumitriu: correct
sburjan - (12:46): I'll check soon and report
sburjan - (12:46): (in a meeting atm)
lucaa left at 12:48 (Ping timeout: 252 seconds
cjdelisle - (12:48): Those reports look great for a lot of uses. I am really interested in ui-tests + selenium1 tests vs code in xwiki-core.
cjdelisle - (12:48): I'm betting a lot of what drags us down is old code which should be deprecated and retired.
cjdelisle - (12:49): bbiab c0ffee
mflorea - (12:50): sdumitriu: regarding the content edit menu bar, it scrolls with the page even it edit mode. It looks strange if WYSIWYG editor dialogs are above it when the page is scrolled..
sdumitriu - (12:52): Yes
mflorea - (12:52): hmm, I guess it's ok after all
sdumitriu - (12:54): Releasing core
mflorea - (12:55): sdumitriu: I still have to commit the fix for the z-index issue but it's in platform/web
sdumitriu - (12:55): I know
Baur left at 13:00 (Ping timeout: 265 seconds
sburjan left at 13:07 (Read error: Connection reset by peer
sburjan joined #xwiki at 13:07
mflorea - (13:09): sdumitriu: I committed my change. It's not good that the z-index is specified in Java code but I preferred to keep the change minimal before a RC release. I'll refactor that class later.
sdumitriu - (13:09): K
sdumitriu - (13:09): Thanks
mflorea left at 13:10 (Quit: Leaving.
sdumitriu - (13:10): (I spotted the problem two nights ago, but forgot about it)
vmassol1 left at 13:11 (Ping timeout: 252 seconds
Baur joined #xwiki at 13:16
vmassol joined #xwiki at 13:16
Baur - (13:17): I want to modify login page, mainly I want to remove quicklinks panel
Baur - (13:17): how can i do that?
bblfish left at 13:17 (Quit: Leaving.
Denis left at 13:31 (Quit: Leaving.
silviar left at 13:31 (Quit: Leaving.
Denis joined #xwiki at 13:31
Baur - (13:35): ping
silviar joined #xwiki at 13:38
sdumitriu - (13:39): There are several ways
abusenius joined #xwiki at 13:43
Baur - (13:54): I tried to modify startpage.vm file
Baur - (13:54): but no success
sdumitriu - (13:55): What did you do?
Stulle joined #xwiki at 13:55
Stulle - (13:55): hi
sdumitriu - (13:55): Hi Stulle
Stulle - (13:56): i am trying to run a private XWiki, however, on opening the site i get a login screen with a greyish colorset
Stulle - (13:56): although it should be colibri with default color theme
Stulle - (13:56): am i doing something wrong or is there a bug?
sdumitriu - (13:56): The skin is configured in a wiki document
sdumitriu - (13:57): If the wiki is completely private, then it's not possible to read the skin document before being logged in
Stulle - (13:57): oh, good to know
sdumitriu - (13:57): It is a kind of bug
Stulle - (13:57): so what can i do to work around it?
Stulle - (13:57): is there a way to reload the proper skin/color theme upon logging in?
sdumitriu - (13:57): How did you make the wiki private?
Stulle - (13:58): on thw rights page
Stulle - (13:58): i set the three lower checkmarks
lucaa joined #xwiki at 13:58
Stulle - (13:59): any suggestions?
sdumitriu - (13:59): Two options
sdumitriu - (13:59): Change templates/colorThemeInit.vm to use different defaults
sdumitriu - (14:00): Or, uncheck those options and instead deny access rights to guest, and allow view on the default color theme document
mflorea joined #xwiki at 14:00
tmortagne1 - (14:01): what's weird is that right issues with skin is supposed to be fixed according to XWIKI-4378
sdumitriu - (14:02): tmortagne1: It's a different problem
Stulle - (14:02): the latter sounds good to me. however, how would i block guests without those options?
sdumitriu - (14:02): colorThemeInit tries to open the ColorThemes.Default document
tmortagne1 - (14:02): ok so it's the right skin but not the right color theme ?
Stulle - (14:02): indeed
rstavro left at 14:02 (Quit: Page closed
sdumitriu - (14:02): Stulle: From the administration, Rights, swich to Users (initially it's displayng Groups), and set the rights for Unauthenticated Users
Baur left at 14:03 (Quit: Page closed
Stulle - (14:03): ahhh, got it
Stulle - (14:03): hold on, gonna check
abusenius left at 14:04 (Ping timeout: 272 seconds
Stulle - (14:07): that did the trick!
Stulle - (14:07): thanks a bunch!!!
Stulle - (14:07): should i still report this somehow?
Stulle - (14:08): or this could possibly be FAQ content... like somebody else is bound to run into this at some point
abusenius joined #xwiki at 14:08
venkatesh joined #xwiki at 14:08
arkub left at 14:28 (Ping timeout: 276 seconds
Stulle - (14:35): well, thanks again. see you around
venkatesh left at 14:36 (Ping timeout: 240 seconds
Stulle left at 14:40 (Ping timeout: 265 seconds
florinciu left at 14:43 (Quit: Leaving.
sburjan - (15:11): did someone fixed the last ui-test failing ?
sburjan - (15:15): odd... on #74 0 failues
sburjan - (15:15): *failures
sburjan - (15:23): oh, still building
silviar left at 15:25 (Read error: Connection reset by peer
silviar joined #xwiki at 15:25
lucaa left at 15:29 (Ping timeout: 265 seconds
arkub joined #xwiki at 15:39
tmortagne1 - (15:43): sdumitriu: can I commit things on core trunk already ?
sdumitriu - (15:43): Yes
sdumitriu - (15:44): Trunk is still 2.5
tmortagne1 - (15:44): ok good, that way i will not need to merge it in 3 branches ;)
tmortagne1 - (15:44): when do you plan to create the branch ?
sdumitriu - (15:45): After the enterprise release
sdumitriu - (15:45): I'll do all the branches at once
tmortagne1 - (15:45): ok
abusenius left at 16:12 (Ping timeout: 272 seconds
asrfel left at 16:33 (Quit: Leaving.
sburjan left at 16:49 (Read error: Operation timed out
vmassol - (16:55): selenium-tests brings TPC to 40.5% (over the 38.6 by ui-tests)
tmortagne1 - (16:56): selenium-tests+ui-tests or selenium-tests alone ? (sorry if it's obvious with over i'm not sure how to understand it ;))
vmassol - (16:57): platform + ui-tests +selenium-tests = 40.5%
vmassol - (16:57): running other tests now
silviar left at 16:58 (Quit: Leaving.
vmassol - (17:02): (the clover database weight 50MB, takes 10 minutes to generate the report now)
vmassol - (17:02): (and growing)
vmassol - (17:07): cjdelisle: shouldn't the invitationmanager plugin be retired now?
tmortagne1 - (17:08): vmassol: ok
vmassol - (17:10): tmortagne1: note that for some reason I'm not sure I understand the gwt module client classes are listed too so it's possible they're dragging us down a bit. I have excluded them from cloverification so I don't understand why they're listed
tmortagne1 - (17:10): indeed if clover expect them to be executed it's not nice for the %
vmassol - (17:14): now up to 41.9% with rest-tests
sdumitriu - (17:16): cjdelisle: You shouldn't have reused XAINVITATION-12, since it was released already
vmassol - (17:17): sdumitriu: btw I hope you're releasing from a branch?
sdumitriu - (17:17): No
sdumitriu - (17:17): The branch will be made after the release
vmassol - (17:17): why not? It's the first step in the release process
vmassol - (17:18): there's a reason for that
vmassol - (17:18):
vmassol - (17:18): so now it means we have stuff committed for 2.5 that are not for 2.5 :(
tmortagne1 - (17:19): vmassol: like what ? you mean the clover things ?
sdumitriu - (17:19): Hm, I haven't seen release branches in a very long time
vmassol - (17:20): sdumitriu: that means release manager are not following the process which isn't good at all
vmassol - (17:20): it's there for a purpose
vmassol - (17:20): if RM wants to change the process fine but it needs to be discussed
vmassol - (17:21): tmortagne1: like all the commits that have happened since sergiu release core + all that could happen
vmassol - (17:21): s/release core/released core/
tmortagne1 - (17:22): vmassol: as far as i remember we never created branch just for a release and usually people ask when they can commit things when someone is releasing
vmassol - (17:22): that's not true
vmassol - (17:22): I always did
vmassol - (17:22): and it's a best practice everywhere I know
vmassol - (17:22): actually I've never seen anyone doing it the other way around
sdumitriu - (17:22): That's a partial truth
vmassol - (17:23): (unless he made a mistake ;))
vmassol - (17:23): in any case
sdumitriu - (17:23): A best practice is to do the branch very early and prepare the release for days
sdumitriu - (17:23): Creating a branch for 30 minutes is not a good practice
vmassol - (17:23): sdumitriu: not really, except for RC
vmassol - (17:24): 30 minutes?
vmassol - (17:24): it'll live on for several months
vmassol - (17:24): till 2.5 is phased out
vmassol - (17:24): BUT
sdumitriu - (17:24): ?
tmortagne1 - (17:24): vmassol: "I always did" means a very long time ago and AFAIK jvdrean never did as well as me
vmassol - (17:24): yes that'"sq wrong
sdumitriu - (17:25): What branch are you talking about?
vmassol - (17:25): I don't understand why we take the pain to document things
vmassol - (17:25): if people just do whatever they want
sdumitriu - (17:25): The 2.5 branch, or a small branch just for doing the release?
vmassol - (17:25): guys if you don't agree with something please discuss it
vmassol - (17:25): and then modify the doc
vmassol - (17:26): in any case we now have stuff  in svn that says 2.5* when they shouldn't
vmassol - (17:26): and that's bad
vmassol - (17:26): sdumitriu: there's very little activity on a branch
tmortagne1 - (17:26): vmassol: again what things ? i'm pretty sure everything took care like always
sdumitriu - (17:26): What things?
tmortagne1 - (17:27): s/everything/everyone/
lucaa joined #xwiki at 17:27
vmassol - (17:27): just hoping for the best has never been called a& strategy
vmassol - (17:28): BUT what I don't understand is why you're not creating a branch when the release process clearly says to create one
sdumitriu - (17:28): Vincent, please tell me what kind of branch
sdumitriu - (17:28): Like, if I release an application, do I make a branch for it?
vmassol - (17:29): sdumitriu: what's important is to get out of trunk
sdumitriu - (17:30): OK, I'm releasing the extensions application
sdumitriu - (17:30): Do I make a branch for it?
sdumitriu - (17:31): What's it going to be named?
sdumitriu - (17:31): What happens with the branch after the release?
vmassol - (17:31): and especially for platform/core + web and enterprise/ (where majority of commits happen) but yes normally we should do it for all releases including apps. Since we have a lot of apps it could be a pain though for apps and we currently don't maintain branches for them
vmassol - (17:32): but yes this shows we need to fine tune the branching process even more
cjdelisle - (17:32): I suppose the InvitationManager plugin can indeed be retired, I'll send a proposal.
sdumitriu - (17:32): Well, I'm -1 on that
vmassol - (17:33): you don't want to retire the invitationmanager ?
sdumitriu - (17:33): What good is having these branches, as well as their equivalent tags:
sdumitriu - (17:33): xwiki-core-2.5-milestone-1
sdumitriu - (17:33): xwiki-core-2.5-milestone-2
sdumitriu - (17:33): xwiki-core-2.5-rc-1
sdumitriu - (17:33): xwiki-core-2.5 (or -final, so that it doesn't conflict with the 2.5 maintainance branch)
sdumitriu - (17:33): xwiki-core-2.5.1
vmassol - (17:33): (or ?1 on fine tuning the bracning process?)
sdumitriu - (17:33): Will we ever make changes on the xwiki-core-rc-1 branch?
vmassol - (17:34): I don't think we need to create that branch
sdumitriu - (17:34): And do we release xwiki-core-rc-1.1 ?
vmassol - (17:34): but we do need to create a ?2.5 branch
sdumitriu - (17:34): Yes
sdumitriu - (17:34): Agreed
tmortagne1 - (17:34): vmassol: looks like you added this rule in version 26.3 (2009/01/27) which means around XE 1.7.1 so that's probably why, i did not know since then there was this rule
sdumitriu - (17:34): But you said that we always create a branch
vmassol - (17:35): yes that's why I said we need to fine tune the process
tmortagne1 - (17:35): i don't read the releaseprocess page before  each release, you should have discuss it with jvdrean and me and warn us
sdumitriu - (17:35): Do we branch before the first milestone?
vmassol - (17:35): all I want to avoid is locking devs for long time + avoid potential commits in the wrong location
vmassol - (17:35): + really make sure that RC is clean
vmassol - (17:35): since it's a RC
sdumitriu - (17:36): If it's not a release-only branch, then we also lock the dev branch, which is as important as the trunk
vmassol - (17:36): now we can discuss the process to achieve this if we agree about those obecjtives
florinciu joined #xwiki at 17:36
vmassol - (17:37): tmortagne1: honestly I don't recall but I may have discussed it on the list
vmassol - (17:37): (I don't usually add stuff without discussing them but it can happen)
vmassol - (17:38): sdumitriu: dev branch = stable branch?
sdumitriu - (17:38): dev branch = the current definition of branches, like xwiki-core-2.4
sdumitriu - (17:39): It's not "stable"
sdumitriu - (17:39): People commit to branches
vmassol - (17:39): sdumitriu: I'm not suggesting that we develop new stuff elsewhere than on trunk
jvdrean - (17:40): (reading), I don't remember a discussion about creating a branch before each module release
jvdrean - (17:40): indeed it's in the release process (which I don't read entirely when I do a release)
vmassol - (17:41): jvdrean: sdumitriu has good points in that the strategy needs to be refined
vmassol - (17:42): to find a good balance between tedious work and control
vmassol - (17:42): is creating a branch complex and long?
vmassol - (17:42): from svn POV it should be real quick and easy
vmassol - (17:43): (and not even disk costly since it should maintain refs)
abusenius joined #xwiki at 17:43
sdumitriu - (17:44): SVN branches are not easy
sdumitriu - (17:44): We really should move togit
sdumitriu - (17:44): to git
vmassol - (17:44): svn branches are very easy
vmassol - (17:44): it's just a copy
vmassol - (17:44): (pointer copy)
vmassol - (17:44): what's hard is merge but that's not needed here
tmortagne1 - (17:44): creating a branch dureing a release means:
tmortagne1 - (17:44): - creating the branch
tmortagne1 - (17:44): - checkout it
tmortagne1 - (17:44): - then do the release on the branch
tmortagne1 - (17:44): for each module
tmortagne1 - (17:45): it's easy yes from server POV
sdumitriu - (17:46): Performing the release takes about 12G of disk space locally
jvdrean - (17:46): nice :)
sdumitriu - (17:47): Since each integration test uses a zip and an uncompressed version of the distribution
vmassol - (17:47): ok we need to think more about this and make a proposal.
vmassol - (17:47): one of them could be to move to git. I'm just worried about tools integration but maybe the support is acceptable now in IDEs?
vmassol - (17:47): and even with git we would need a strategy
jvdrean - (17:48): I don't see the point of the strategy thomas described above
vmassol - (17:48): jvdrean: do you agree with the objectives I listed?
jvdrean - (17:48): the timespan between branching and release would be a minute, or vmassol do you think about creating branches for all the modules to release, and then start the release ?
vmassol - (17:49): "all I want to avoid is locking devs for long time + avoid potential commits in the wrong location
vmassol - (17:49): + really make sure that RC is clean"
sdumitriu - (17:49): The locking timespan for applications is a few seconds
vmassol - (17:51): sdumitriu: maybe the branching rule should only be for RC
vmassol - (17:51): ?
tmortagne1 - (17:51): it's not that big either even for core actually if we are only talking about release:prepare since the perform is done on a clean tag
cjdelisle - (17:51): Just read over the release process. My first reaction is that it's so complicated that it's a virtual guarantee that something will go wrong.
jvdrean - (17:51): vmassol: difficult not to agree with that, now I think our current non-strategy works quite well, developers are warned that a release is in progress, they must hold their commits
vmassol - (17:51): jvdrean: for RC you're going to ado a branch anyway
tmortagne1 - (17:51): vmassol: note that for RC we usually created the branch before releasing it
vmassol - (17:51): so why not start by doing it?
tmortagne1 - (17:52): sdumitriu did it the other way this time but it's not critical either (it's possible i made the mistake too)
jvdrean - (17:52): yes branching before doing the release for core/web/enterprise is the way to go, that's what we usually do
sdumitriu - (17:53): Applications released
sdumitriu - (17:53): Waiting for the web modules to be uploaded
vmassol - (17:53): ok so maybe we can modify the branching rule for now by stating that the branch is created for RC? wdyt?
vmassol - (17:54): tmortagne1: you committed stuff on trunk too right?
vmassol - (17:54): (for blockquote)
tmortagne1 - (17:54): vmassol: after asking sdumitriu if he finished svn release of core already
vmassol - (17:54): in principle this means we''re now forced to do a RC2
sdumitriu - (17:54): Why?
tmortagne1 - (17:54): so it's not going to be in the RC1
sdumitriu - (17:55): vmassol: There's no rule that there should be no commits between the RC and the final
tmortagne1 - (17:55): vmassol: that's not a new feature or something that's a bugfix
vmassol - (17:55): yes ther is
sdumitriu - (17:55): We've always included small fixes
tmortagne1 - (17:55): the point of RC is to fix bugs
vmassol - (17:55): no
vmassol - (17:55): we only started to do that quite late actually
vmassol - (17:56): I don't remember doing it when I was RM
vmassol - (17:56): oterhwise you can't guarantee
vmassol - (17:56): but
vmassol - (17:56): in any case it's a bad practice
vmassol - (17:56): normally the final is just a promotion of a RC
vmassol - (17:56): ie a rename
vmassol - (17:56): (it's not a re-release)
vmassol - (17:57): oterhwise it's a RC in disguise :)
cjdelisle - (17:57): +1
vmassol - (17:58): the only issue we have for doing this
vmassol - (17:58): (apart from setting up nexus)
sdumitriu - (17:58): +1 as well, but we're not doing this
sdumitriu - (17:58):
vmassol - (17:58): is that we have the version in
vmassol - (17:58): sdumitriu: I know we've been lax
vmassol - (17:58): but there's a difference between having been lax and establishing it as an implicit rule
vmassol - (17:59): if we agree we must fight it
vmassol - (17:59): not make it the norm
vmassol - (17:59): re version.propreties I think jason told me you could run some scripts in the promotion step
vmassol - (18:01): ok let's talk about all this after the release
vmassol - (18:01): I think I'd like to do the next release
vmassol - (18:01): since I haven't done any for a long time
vmassol - (18:01): and I need to practice again
vmassol - (18:01): I also would like to take the time to improve the process
vmassol - (18:02): (set up nexus, etc)
vmassol - (18:02): let's talk about it on the list
lucaa left at 18:02 (Ping timeout: 265 seconds
vmassol - (18:02): sorry for the disruption sergiu
fmancinelli left at 18:03 (Ping timeout: 250 seconds
vmassol - (18:03): I  guess I was just angry about myself for committing something that I didn't want to go on 2.5 and that I redirected my internal anger on you for not following the process that I thought we were following…. something like that :)
vmassol - (18:04): in any case I apologize
vmassol - (18:04): and let's take time to discuss this on the list instead
sdumitriu - (18:06): web released
tmortagne1 left at 18:16 (Quit: Leaving.
lpereira joined #xwiki at 18:41
jvdrean left at 18:51 (Quit: Leaving.
lucaa joined #xwiki at 18:53
fmancinelli joined #xwiki at 18:57
vmassol left at 18:57 (Ping timeout: 260 seconds
vmassol joined #xwiki at 18:58
sburjan` joined #xwiki at 19:14
arkub left at 19:25 (Quit: Leaving
mflorea left at 19:53 (Quit: Leaving.
vmassol - (20:45): I tried to run the wysiwyg tests locally....
vmassol - (20:45): result:
vmassol - (20:45): Tests run: 274, Failures: 2, Errors: 131, Skipped: 0
vmassol - (20:45): :)
vmassol - (20:46): I'll retry without screen saver/lock screen to see if that's the problem (would be bad)
evalica left at 20:46 (Quit: Leaving.
cjdelisle - (20:49): A lot of webdriver tests fail if you force the ff window not to pop up. I think it causes the onfocus event not to be triggered.
vmassol - (20:52): could the screen saver cause that?
cjdelisle - (20:54): I imagine it would interrupt the screensaver. If I run tests, the window periodically pops up and if I put it in a different workspace to prevent it popping up, tests fail.
florinciu1 joined #xwiki at 21:04
fmancinelli left at 21:50 (Ping timeout: 276 seconds
fmancinelli joined #xwiki at 21:52
mflorea joined #xwiki at 21:58
lpereira left at 22:46 (Quit: Leaving.
fmancinelli left at 23:00 (Ping timeout: 264 seconds
mflorea left at 23:04 (Quit: Leaving.
florinciu1 left at 23:04 (Read error: Connection reset by peer
florinciu left at 23:05 (Quit: Leaving.
jvdrean joined #xwiki at 23:11
vmassol - (23:39): not better….
vmassol - (23:39): Tests run: 274, Failures: 2, Errors: 130, Skipped: 0
vmassol - (23:40): anyway bed time….
vmassol - (23:40): nn
cjdelisle - (23:40): gnight
vmassol left at 23:42 (Quit: Leaving.

Get Connected