IRC Archive for channel #xwiki on 23 June 2014

Last modified by Vincent Massol on 2014/06/23 23:21

<abusenius> has quit
03:49 <Inge-_> has joined #xwiki
03:49 <Inge-> has quit
05:05 <polx> has joined #xwiki
05:08 <polx> has quit
05:56 <polx> has joined #xwiki
06:02 <polx> has quit
06:07 <polx> has joined #xwiki
06:14 <polx> has quit
06:19 <polx> has joined #xwiki
06:21 <polx> has quit
06:44 <polx> has joined #xwiki
07:06 <polx> has quit
07:08 <polx> has joined #xwiki
07:09 <polx> has quit
07:34 <Lyes> has joined #xwiki
07:40 <evalica> has joined #xwiki
07:42 <Enygma`> has joined #xwiki
07:50 <Enygma`> has quit
07:58 <Enygma`> has joined #xwiki
08:00 <Enygma`> has quit
08:04 <Enygma`> has joined #xwiki
08:10 <msmeria> has joined #xwiki
08:21 <msmeria> has quit
08:22 <msmeria> has joined #xwiki
08:29 <tmortagne> has joined #xwiki
08:36 <Denis> has joined #xwiki
08:41 <vmassol> has joined #xwiki
08:42 <Enygma`> has quit
08:45 <Enygma`> has joined #xwiki
08:57 <vmassol> good morning
08:57 <vmassol> so today is release day for 6.1RC1
08:57 <vmassol> do we have a RM? :)
08:58 <vmassol> BTW and FTR I'm leaving early this afternoon to travel to do a training in the south of France till Wednesday night so I won't be much available at all during this time
09:01 <woshilapin> has joined #xwiki
09:04 <mflorea> has joined #xwiki
09:09 <gdelhumeau> has joined #xwiki
09:15 <ClemensR> has joined #xwiki
09:17 <tmortagne> vmassol: looks like we have 2 blockers on 6.1RC1
09:18 <vmassol> ok, we still need a RM to make things progress and ensure the release is done ASAP
09:19 <tmortagne> ClemensR: don't forget to put N/A in documentation filed when closing bug jira issue that don't really need to be documented
09:19 <ClemensR> ah, yes, I forgot ... have not been sure if I should document it and where, and then I forgot ..
09:20 <tmortagne> vmassol: I can take case of this if there is nobody else I guess
09:20 <ClemensR> ah, no, it's in the bugs - there I plain forgot it, sorry
09:21 <vmassol> thanks tmortagne, if someone has the time to do it, let us know. So I guess you're on FTM :) Thanks
09:22 <tmortagne> but someone will need to look at XWIKI-10497
09:22 <tmortagne> as for XWIKI-10446 IMO we can't really do much than waiting for signing all that
09:22 <tmortagne> it's not like it was a 6.1 regression
09:23 <tmortagne> we should not be far from it since Denis implemented lots of stuff already
09:34 <evalica> tmortagne: pls let me know when you're starting the release, maybe I still have time to commit something, if not it will go in the next release. Ty :)
09:35 <tmortagne> evalica: the release is stuck for now, need someone to look at XWIKI-10497
09:35 <tmortagne> so you can go
09:35 <tmortagne> jenkins is in a better shape than usually before a release :)
09:37 <tmortagne> checking init.vm related complains of escaping tests
09:53 <gdelhumeau> guys
09:54 <gdelhumeau> I have a very strange problem
09:54 <gdelhumeau> I have modified annotations-ui recently
09:54 <gdelhumeau> when I build a new XE, annotations are not displayed anymore by default
09:54 <gdelhumeau> If i re-import annotatiouns-ui.xar, it works again
09:56 <tmortagne> gdelhumeau: you mean if you take latest jetty/hsqldb snapshot ans start it annotations  don't work ?
09:56 <tmortagne> s/ans/and/
09:56 <gdelhumeau> the annotations button isn't displayed at all
09:57 <gdelhumeau> I think it is because .Script needs ProgramlAnnotationCodeming Rights
09:57 <gdelhumeau> I think it is because AnnotationCode.Script needs Programing Rights
09:58 <gdelhumeau> If I just re-save the page, it's ok
09:58 <tmortagne> gdelhumeau: id you updated recently ? Denis broke standard PR and then fixed it last week
09:58 <tmortagne> s/id/did/
09:58 <gdelhumeau> last thursday
09:58 <gdelhumeau> and sorry, I have built XE this morning
09:59 <tmortagne> I just downloaded latest jetty/hsqldb, lets see
10:00 <tmortagne> note that annotations tests (if any but I think there is a few) don't fail on jenkins
10:00 <gdelhumeau> yes, it is strange indeed
10:02 <tmortagne> gdelhumeau: I reproduce on last jenkins
10:02 <gdelhumeau> good
10:02 <tmortagne> I mean last snapshot
10:02 <gdelhumeau> now, just resave AnnotationCode.Script
10:03 <tmortagne> yes I did and now it works
10:03 <gdelhumeau> what could be the cause?
10:03 <Denis> should be my fault, for sure
10:03 <tmortagne> but that seems to be the only document that lost PR
10:03 <tmortagne> other documents requiring PR that I can think of are working well
10:04 <Denis> when I look the content author of that document, I see XWiki.superadmin (I have not the latest snapshot)
10:04 <tmortagne> and there is no way the content of this document is anything else than Admin at startup
10:04 <tmortagne> you mean in the xml ?
10:04 <Denis> yes
10:04 <tmortagne> I don't think this is supposed to be taken into account
10:05 <tmortagne> but if that it then it means superdamin lost PR I guess
10:05 <gdelhumeau> it can't be related to the commit I just pushed?
10:06 <tmortagne> gdelhumeau: I doubt it since it really seems related to the document author
10:06 <tmortagne> bu I don't understand why jenkins does not have any issue
10:06 <tmortagne> I'm pretty sure there is annotations tests
10:07 <tmortagne> testing some stuff with superadmin
10:09 <tmortagne> hmm $hasProgramming gives me true in a document saved with superadmin
10:09 <tmortagne> maybe an init bug then
10:10 <tmortagne> gdelhumeau: any idea where this button come from ? jsx ? uix ?
10:10 <gdelhumeau> the annotations button: jsx
10:10 <tmortagne> so that would be a not registered jsx
10:11 <gdelhumeau> yes, because the JSX is "always on the wiki"
10:11 <tmortagne> I tough registered jsx for a whole wiki was only requiring admin right
10:11 <gdelhumeau> it is
10:11 <gdelhumeau> ah sorry
10:11 <gdelhumeau> read to fast
10:11 <gdelhumeau> *too
10:12 <tmortagne> if none of the jsx were registered pretty much nothing would work I guess so I doubt that that's the real issue
10:12 <lucaa> has joined #xwiki
10:12 <Denis> there is a small change in the way PR are tested now
10:13 <tmortagne> Denis: I just saved the document with superadmin and I still have the button so I don't think it's a superadmin/PR issue
10:13 <Denis> before, we were asking PR on the wiki
10:13 <Denis> now we are asking PR on the document itself
10:14 <Denis> for superadmin, this should not have any impact
10:14 <tmortagne> restarting the wiki
10:15 <tmortagne> yep it's something at init
10:15 <tmortagne> after restarting the wiki, the button is not there anymore
10:17 <Denis> let me check without this small diff
10:18 <tmortagne> (and whatever the author at init, I tried with both superadmin and Admin as content author at init)
10:18 <tmortagne> gdelhumeau: any jsx in mind that would work ?
10:18 <tmortagne> to see if there is something special about the annotation one
10:19 <vmassol> bb in 10mn
10:24 <tmortagne> Denis: I cnofirm that the problem does not exist in 6.1M2 at least
10:24 <tmortagne> so it's only in snapshot
10:24 <gdelhumeau> tmortagne: only annotations use JSX with "always on the wiki'
10:24 <Denis> could be that small diff, I am in the meeting now, I try to check that in //
10:24 <gdelhumeau> but AnnotationCode.Settings too
10:25 <tmortagne> gdelhumeau: really ? everything else is on demand ?
10:25 <gdelhumeau> http://hastebin.com/uyiyuxedig.vm
10:26 <tmortagne> ok so that could be a general "always on wiki" issue at init then indeed
10:26 <gdelhumeau> yes
10:26 <tmortagne> checking what skinx does at init
10:33 <tmortagne> Denis: I think there is something wrong with XWikiCachingRightService#hasProgrammingRights
10:34 <tmortagne> contextualAuthorizationManager.hasAccess(Right.PROGRAM, doc.getDocumentReference());
10:34 <tmortagne> means testing context user, rigth ?
10:34 <tmortagne> which it should test document user in this method
10:34 <tmortagne> document content author I mean
10:35 <tmortagne> (that's what XWikiRightServiceImpl is doing)
10:35 <tmortagne> and what skinx is expecting from what I understand in the code
10:36 <Denis> no
10:36 <Denis> the backward compatibility check is done in the contextual manager
10:38 <tmortagne> ok
10:39 <Denis> I have initially only done it there, but it was wrong
10:39 <Denis> (there = in the caching right service)
10:40 <Denis> I have check fixing the small diff mention above, but it does not really fix the issue
10:41 <tmortagne> hmm I don't reproduce the issue anymore
10:41 <tmortagne> testing several times
10:41 <Denis> ?!?
10:42 <Slashman> has joined #xwiki
10:42 <tmortagne> forget that, I'm still on M2..
10:44 <tmortagne> Denis: where is that hack for PR done exactly ?
10:45 <tmortagne> ha it's in getCurrentUser
10:45 <Denis> in org.xwiki.security.authorization.internal.DefaultContextualAuthorizationManager#getCurrentUser
10:46 <Denis> but there is also other aspects
10:46 <tmortagne> in any case not sure what wrong in details but the issue is that context.getWiki().getRightService().hasProgrammingRights(doc, context) return false
10:46 <Denis> I am testing some stuff manually, let me some time
10:47 <tmortagne> comes from checkPreAccess(right)
10:47 <tmortagne> hmm xcontextProvider.get().hasDroppedPermissions() return true
10:47 <tmortagne> so the issue is not excatly in context.getWiki().getRightService().hasProgrammingRights(doc, context) but before that it seems
10:49 <Denis> it is an issue in the stubcontext ?
10:49 <KermitTheFragger> has joined #xwiki
10:50 <tmortagne> DefaultXWikiStubContextProvider does not call dropPermission no
10:51 <tmortagne> it's called by:
10:51 <tmortagne> * contentview.vm
10:52 <tmortagne> and that's all
10:53 <tmortagne> maybe some order changed in view.vm
10:53 <tmortagne> like registering jsx after rendering the content
10:53 <tmortagne> gdelhumeau: any idea related to that ?
10:53 <tmortagne> when you save a jsx it's automatically added to some map for later so that's why resaving it works
10:54 <tmortagne> so it's not exactly an init issue
10:54 <lucaa> has quit
10:56 <lucaa> has joined #xwiki
10:56 <vfzevaze> has joined #xwiki
10:57 <Enygma`> has quit
10:59 <tmortagne> I confirm that dropPermission is called before first call to getAlwaysUsedExtensions()
11:00 <gdelhumeau> dropPermission is here for age
11:00 <gdelhumeau> hmm
11:00 <gdelhumeau> haven't changed view.vm of colibri recently
11:00 <tmortagne> (by the way it does not make any sense that something like getAlwaysUsedExtensions() is influenced by dropPermission, it should directly test right for the content author using and bypass the context)
11:01 <tmortagne> gdelhumeau: yes I know dropPersmission is old but now it's executed before initializing jsx
11:01 <Denis> in fact
11:01 <Denis> before hasProgrammingRight(context) called it
11:01 <Inge-_> is now known as <Inge->
11:01 <Denis> but hasProgrammingRight(doc, context) does not
11:02 <tmortagne> Denis: I'm not talking about security API
11:02 <tmortagne> hasProgrammingRight should take care of the droppermission
11:02 <tmortagne> but getAlwaysUsedExtensions() should not use contextual auth
11:02 <Denis> this explains why it works before
11:03 <tmortagne> because getAlwaysUsedExtensions does not care about the context, it olny care about the jsx content author
11:03 <tmortagne> ok
11:03 <tmortagne> fixing getAlwaysUsedExtensions then
11:05 <Denis> so it should not use the contextual auth manager, but the auth manager directly
11:05 <tmortagne> Denis: that's what I said yes
11:05 <Denis> just wanted to be sure (I am on a meeting in //)
11:05 <tmortagne> I guess we will have to review all calls to hasProgrammingRight(doc, context) if its behavior changed
11:06 <Denis> that was on my todo
11:06 <tmortagne> it's probably not the only think that was broken since it was the only way to test PR or an existing document o old code
11:06 <tmortagne> might have broke quite some code
11:06 <tmortagne> hope it's not going to be a massacre for others
11:06 <Denis> I got disturbed by other projects, as you know
11:07 <Denis> there was no clear diff between contextual and not contextual stuff before
11:07 <tmortagne> Denis: sure but we can't release without checking all uses of hasProgrammingRight(doc, context)
11:07 <Denis> I see it more as a bug IMO
11:07 <Denis> this could also break user code
11:08 <Denis> maybe, we could do a better approach
11:08 <tmortagne> I did not suggested to revert your fix but again that could break many code that only had this API before
11:08 <Denis> change the right service to use the  non-contextual auth manager on the second method
11:08 <evalica> has quit
11:08 <Denis> this would ensure, right service is stable
11:08 <evalica1> has joined #xwiki
11:09 <Denis> until we review all use cas
11:09 <Denis> e
11:09 <tmortagne> we probably don't have a choice for right service yes
11:09 <Denis> and only get that change applied when the right service is deprecated
11:10 <tmortagne> but no reason to change contextual auth
11:12 <tmortagne> also probably make it clear in XWikiRightService javadoc
11:12 <Denis> sure, contextualAuthMan is perfectly ok
11:13 <Denis> this is more a compatibility issue
11:13 <Denis> do you want me to do the fix ?
11:13 <tmortagne> Denis: as you prefer, if you don't have time I can take care of it
11:14 <Denis> I will
11:14 <Denis> it is easy
11:15 <tmortagne> still fixing the skinx, better anyway
11:17 <tmortagne> will avoid putting back the regression when we fix deprecated API calls ;)
11:17 <vmassol> guys, yes please try to make sure that our extensions contributed externally will still work (it's possible some are calling hasProgramminhRight I guess)
11:17 <Denis> maybe, should I only use the non contextual when the doc is not null
11:17 <Denis> wdyt ?
11:18 <Denis> vmassol: we are fixing right service to be backward compatible
11:19 <vmassol> cool
11:20 <tmortagne> Denis: IMO the safest is to have the exact same behavior than old hasProgrammingRights
11:21 <tmortagne> so never testing droppermission on matter what
11:21 <tmortagne> s/on/no/
11:21 <tmortagne> old hasProgrammingRights(XWikiDocument doc, XWikiContext context) that is
11:21 <Denis> the question is, doesn't we have an issue when hasProgrammingRights(context) != hasProgrammingRights(null, context)
11:21 <tmortagne> they were different before
11:22 <Denis> but isn't it a bug
11:22 <tmortagne> not testing droppermission in hasProgrammingRights(XWikiDocument doc, XWikiContext context) is a bug also IMO
11:22 <tmortagne> but it's too late
11:22 <Denis> not necessarily
11:22 <tmortagne> since we have a new component let keep it the same behavior it always had
11:23 <tmortagne> if someone wan't the right behavior he can move to new SS or component
11:23 <Denis> when using not null document, nothing is taken from context, so it is badly documented but it is consistant
11:24 <tmortagne> changing the behavior now is a bad idea IMO
11:26 <tmortagne> since we have a new API lets not take any risk with the old one
11:28 <vmassol> +1 for no risk :)
11:30 <vmassol> " Used to API" == Use new API?
11:30 <vmassol> :)
11:31 <tmortagne> yes written too quickly
11:38 <tmortagne> any taker for XWIKI-10497 ?
11:38 <tmortagne> gdelhumeau, mflorea ?
11:39 <gdelhumeau> could try
11:42 <tmortagne> great thanks
11:44 <Denis> tmortagne: fixed
11:45 <tmortagne> Denis: thanks :)
11:46 <Denis> it solve the issue obviously, I got the annotation button back :)
11:47 <gdelhumeau> good
11:47 <tmortagne> probably solve a few others too :)
11:59 <Denis> tmortagne: it was mainly a revert in that case. I am now looking at the first method which also have a few minor change, but that should also be considered
12:01 <Denis> as I said earlier, the check is done at document level instead of wiki level, but it should not have any impact AFAIK
12:03 <Denis> the other difference is that we now check the rendering context restriction
12:03 <Denis> if we don't want those diff, I need to also revert the change in the first method
12:04 <Lyes> has quit
12:12 <Lyes> has joined #xwiki
12:37 <evalica1> is now known as <evalica>
12:55 <lucaa> has quit
12:57 <woshilapin> has quit
12:57 <DarkKnightCZ> has joined #xwiki
13:01 <polx> has joined #xwiki
13:03 <ClemensR> gdelhumeau: as you are looking for XWIKI-10497  ... I remember I changed something in the livetable macro in XWIKI-10418  ... I am looking if that could possibly have caused the issue, but so far did not find anything
13:04 <gdelhumeau> actually, if I write "alert('hello')" in the script tag of the #livetable macro, it is not displayed with IE inside the modal box
13:11 <gdelhumeau> ClemensR: I have reverted your commit and it seems it does not fix the issue
13:11 <gdelhumeau> (locally)
13:11 <gdelhumeau> but the weird thing is that it worked in 6.0.1
13:12 <gdelhumeau> lunch time
13:12 <polx> has quit
13:20 <DarkKnightCZ> what piece of code is handling adding / modifying objects in app within minutes? (currently, date field ignores user timezone, which is not desirable, especially with applications like calendar)
13:25 <ClemensR> gdelhumeau: yes, same here, the IE9 (and only IE9, in IE9-mode) does not execute the script
13:29 <Lyes1> has joined #xwiki
13:31 <Lyes> has quit
13:32 <ClemensR> I copied over the complete "template" folder from 6.0.1 to 6.1-milestone-2 and the problem remains
13:34 <Enygma`> has joined #xwiki
13:39 <woshilapin> has joined #xwiki
13:55 <polx> has joined #xwiki
13:57 <ClemensR> gdelhumeau:  there is one difference I found between 6.0.1 and  6.1.-SNAP ad that is script tags seem to get prefixed by /*<![CDATA[*/  automatically in 6.1-SNAP, but not in 6.0.1. No idea where this comes from (or if I am hunting ghosts).
13:57 <ClemensR> I have admit I'll leave it at that
14:01 <Slashman_> has joined #xwiki
14:04 <Slashman> has quit
14:09 <cjd> has joined #xwiki
14:21 <vmassol> has quit
14:26 <gdelhumeau> ClemensR: indeed, removing CDATA and put {{html clean="false"}} fix the problem
14:27 <ClemensR> oh, odd ... no idea why it worked in 6.0.1 then ..
14:28 <GuillaumeF> has joined #xwiki
14:29 <gdelhumeau> /<![CDATA[ is in macros.vm for many years
14:29 <gdelhumeau> but maybe the problem comes from the HTML Cleaner
14:30 <gdelhumeau> which replaces it by  /*<![CDATA[*/
14:30 <lucaa> has joined #xwiki
14:32 <tmortagne> gdelhumeau: so that would be a regression caused by a htmlcleaner upgrade
14:33 <gdelhumeau> actually it does not replace it
14:33 <gdelhumeau> but web inspector played with me
14:33 <tmortagne> hmm we did not upgraded htmlcleaner in 6.1 it seems
14:35 <tmortagne> was in 6.0M2...
14:35 <tmortagne> so 6.0.1 and 6.1-SNAPSHOT should have the same htmlcleaner version
14:36 <gdelhumeau> {{html clean=false}} fixes the issue. I'm gonna compare the generated code with and without the cleane
14:37 <ClemensR> yes, 6.1-SNAPSHOT and 6.0.1 have the same htmlcleaner-2.8.jar   (at least over here ;)
14:37 <gdelhumeau> clean=false AND removing <![CDATA fixes the issue
14:39 <ClemensR> but if I  use wget to fetch i.e. WikiManager.WebHome (no js / ajax stuff or login needed to get that page) I get //<![CDATA[   in the livetable script tag with 6.0.1, and a /* <![CDATA[ */  with 6.1-SNAPSHOT ... maybe it is possible to configure this somewhere ?
14:41 <gdelhumeau> yes it replaces it by /*<![CDATA[*/
14:41 <gdelhumeau> so there is definitively something there
14:43 <tmortagne> gdelhumeau, ClemensR: you end up with /*<![CDATA[*/ even with clean =false ?
14:43 <ClemensR> no
14:43 <gdelhumeau> no
14:43 <tmortagne> so why "clean=false AND removing <![CDATA" is required ?
14:43 <ClemensR> I am only still trying to figure out what introduced the difference ...
14:44 <tmortagne> clean=false is not enough to have the same than 6.0.1 ?
14:45 <tmortagne> anyway it's weird than the exact same version of htmlcleaner can have different behavior
14:47 <tmortagne> gdelhumeau: https://github.com/xwiki/xwiki-commons/commit/4908f99341673af2776ff5866c05a7db82d8b489 ?
14:47 <gdelhumeau> yep
14:47 <tmortagne> did not check all the details but seems to be adding /* */
14:47 <gdelhumeau> it looks like this!
14:47 <tmortagne> and it's recent
14:48 <gdelhumeau> yes it must be this issue
14:48 <tmortagne> not sure how this change can be related to css
14:48 <gdelhumeau> because in CSS the comments are /* */, not //
14:48 <ClemensR> ah, ys, that is correct for inline CSS, but seemingly not for script (and IE9)...
14:49 <tmortagne> very nice, so a change only valid for css in the default htmlcleaner ?
14:50 <ClemensR> yes, I would say so
14:52 <gdelhumeau> hmm I have a cache issue with IE so I have troubles when checking what is wrong
14:52 <gdelhumeau> but logically, only the commit introduced by Sergiu could be the origin of this issue
14:53 <gdelhumeau> tmortagne: it's not normal that IE have problems with /* CDATA */ instead of // CDATA
14:53 <gdelhumeau> again it's an IE stupid issue
14:54 <tmortagne> the comment say all that is for IE6, does we need it at all now ?
14:54 <gdelhumeau> so they only way to fix it is to have a new parameters for the HTML cleaner...
14:54 <tmortagne> (since we don't support IE6 anymore)
14:54 <tmortagne> shouldn't we simply stop commenting CDATA ?
14:56 <tmortagne> if it's really only for IE6 there is no reason to keep doing it
14:56 <gdelhumeau> http://stackoverflow.com/questions/66837/when-is-a-cdata-section-necessary-within-a-script-tag
14:56 <ClemensR> ftr: here just using clean="false" does not work, too ... the commented CDATA seens to be the prpbnlem, no mater if // or */ ... */
14:57 <ClemensR> (modulo typos ...)
14:57 <gdelhumeau> ClemensR: try to clean your browser cache
14:57 <gdelhumeau> I don't know how to do it with IE
14:57 <gdelhumeau> but when I go to an other page that have a livetable, it works with //CDATA
14:57 <ClemensR> gdelhumeau:  I checked with dev-tools that the HTML has a //<![CDATA[  with clean="false"
15:01 <gdelhumeau> http://dorward.me.uk/www/comments-cdata/#trumpet
15:01 <tmortagne> gdelhumeau, ClemensR: any comment on my lets remove all that suggestion ?
15:01 <gdelhumeau> I don't understand
15:01 <tmortagne> by the way according to the that has been applied to 5.4 branch too so it's probably in 5.4.5
15:01 <ClemensR> aty least in this case it seems to be the change that fixed the bug, the CDATA scrapping, I mean
15:02 <ClemensR> I am hesitating as I have no idea what else might break
15:02 <tmortagne> gdelhumeau: the code say that CDATA is commented for IE6 only
15:02 <Enygma`> has quit
15:02 <tmortagne> either the code is not true or we don't need to do that anymore
15:03 <gdelhumeau> or we improve XWikiDOMSerializer#flushContent()
15:04 <gdelhumeau> trivial to improve actually
15:04 <ClemensR> I am 95% confident we do not need this anymore, but the remaining 5% bother me - such a change is a good way to blow up a release in my experience (please dont ask for details ...)
15:04 <tmortagne> I don't really see the point in supporting something useless
15:04 <tmortagne> it's making the code more complex and so more bug prone for nothing
15:04 <ClemensR> for me it is that there is 5% chance it is notuseless
15:04 <Enygma`> has joined #xwiki
15:05 <gdelhumeau> "The comment syntax is helpful only to user agents that you are extremely unlikely to encounter and, as such, are not worth worrying about."
15:05 <gdelhumeau> he doesn't say explictly which ones
15:06 <tmortagne> "user agents that you are extremely unlikely to encounter" is enough for me
15:06 <tmortagne> we have an explicit list of supported browsers so it should not be too hard to simple test them at worst
15:06 <tmortagne> s/simple/simply/
15:07 <gdelhumeau> Ok it looks the smartest choice then
15:07 <gdelhumeau> XWIki does not support old stupid browsers!
15:08 <ClemensR> "extremely unlikely" even back in 2009 actually :)
15:09 <gdelhumeau> should I create a new issue in XCOMMONS?
15:16 <evalica> sorry for the spam :) but I still have work to do
15:16 <tmortagne> gdelhumeau: you can move XWIKI-10497 I guess since there is nothing to do at platform level, right ?
15:17 <gdelhumeau> tmortagne: nothing to do but it is good to keep the description of the bug to avoid having it reported twice?
15:17 <tmortagne> you can keep the same description IMO it's ok
15:18 <tmortagne> (you can create another issue if you really prefer :))
15:18 <gdelhumeau> yes :)
15:19 <Enygma`> has quit
15:20 <Enygma`> has joined #xwiki
15:20 <Enygma`> has quit
15:20 <Enygma`> has joined #xwiki
15:21 <Enygma`> has quit
15:21 <Enygma`> has joined #xwiki
15:24 <gdelhumeau> XCOMMONS-611
15:26 <Enygma`> has quit
15:27 <Enygma`> has joined #xwiki
15:30 <Enygma`> has quit
15:44 <lucaa> has quit
15:50 <tmortagne> gdelhumeau: seems to me you can also remove COMMENT_START and COMMENT_END
15:57 <msmeria> has quit
15:59 <lucaa> has joined #xwiki
16:10 <vfzevaze> has quit
16:12 <tmortagne> gdelhumeau: there is also a few tests to modify in rendering
16:16 <DarkKnightCZ> has quit
16:29 <gdelhumeau> ok, handling it
16:29 <lucaa> has quit
16:31 <evalica> has quit
16:32 <lucaa> has joined #xwiki
16:47 <gdelhumeau> shit
16:47 <gdelhumeau> now IE9 says syntax error about <!CDATA
16:48 <gdelhumeau> this is so stupid...
16:48 <gdelhumeau> actually, IE9 needs the // comment on the CDATA
16:48 <gdelhumeau> actually, IE9 needs the // comment on the CDATA tag
16:48 <tmortagne> in script ?
16:48 <gdelhumeau> yes -_-
16:49 <gdelhumeau> now I'm gonna revert my modifications and improve escapeXML
16:49 <tmortagne> so much for the "user agents that you are extremely unlikely to encounter"
16:50 <ClemensR> I guess the issue is that we are "in HTML mode", not "in XHTML mode"
16:50 <ClemensR> maybe it helps having spaces around the CDATA (just for IE9?)
16:51 <tmortagne> right that's the over hack about IE6, not really declaring xhtml
16:59 <tmortagne> looks like it's really not clear that milestones are beta versions...
17:02 <tmortagne> we should maybe use "beta" instead of "milestone"
17:22 <Lyes1> has quit
17:22 <DarkKnightCZ> has joined #xwiki
17:24 <gdelhumeau> this time I have a the good fix
17:26 <ClemensR> ah, ok, phew
17:29 <tmortagne> gdelhumeau: great :)
17:37 <ClemensR> has quit
17:39 <gdelhumeau> fixing rendering
17:46 <polx> has quit
17:53 <sdumitriu> has quit
17:53 <sdumitriu> has joined #xwiki
17:55 <Denis> has quit
17:56 <tmortagne> has quit
18:05 <gdelhumeau> has quit
18:05 <mflorea> has quit
18:13 <Slashman_> has quit
18:14 <Slashman> has joined #xwiki
18:20 <polx> has joined #xwiki
18:20 <KermitTheFragger> has quit
18:25 <polx> has quit
18:30 <cjd> has quit
18:32 <polx> has joined #xwiki
18:40 <woshilapin> has quit
18:42 <GuillaumeF> has quit
18:45 <abusenius> has joined #xwiki
18:54 <abusenius> has quit
19:05 <polx> has quit
19:08 <polx> has joined #xwiki
19:25 <Slashman> has quit
19:31 <polx> has quit
19:32 <Lyes> has joined #xwiki
19:36 <lucaa> has quit
19:41 <Lyes> has quit
19:52 <polx> has joined #xwiki
20:17 <abusenius> has joined #xwiki
20:27 <polx> has quit
20:27 <polx> has joined #xwiki
20:32 <Lyes> has joined #xwiki
20:35 <Lyes> has quit
20:37 <OSIMasson> has joined #xwiki
22:36 <vmassol> has joined #xwiki
22:36 <Denis> has joined #xwiki
22:49 <DarkKnightCZ> has quit
22:50 <vmassol> has quit
23:07 <Denis> has quit
23:21 <polx> has quit
Tags:
   

Get Connected