IRC Archive for channel #xwiki
Last modified by Vincent Massol on 2012/10/18 19:14
00:00 <lucaa> has joined #xwiki
00:55 <lucaa> has quit
01:09 <mflorea> has quit
01:52 <DrLou> has quit
03:07 <abusenius> has quit
03:44 <venkatesh> has joined #xwiki
03:44 <abusenius> has joined #xwiki
03:45 <venkatesh> has quit
03:45 <venkatesh> has joined #xwiki
03:51 <abusenius> has quit
06:04 <Denis> has quit
06:09 <Denis> has joined #xwiki
06:55 <vmassol> has joined #xwiki
07:00 <sdumitriu> has quit
08:04 <mflorea> has joined #xwiki
08:33 <sburjan`> has quit
08:33 <sburjan``> has joined #xwiki
08:33 <sburjan``> is now known as <sburjan`>
08:33 <sburjan`> has left #xwiki
08:33 <sburjan_> has joined #xwiki
08:33 <sburjan_> has quit
08:33 <sburjan`> has joined #xwiki
08:35 <sburjan``> has joined #xwiki
08:43 <sburjan``> has left #xwiki
08:44 <jstoldt> has joined #xwiki
08:44 <jstoldt> good morning everyone
08:44 <@cjdelisle> good morning
09:15 <lucaa> has joined #xwiki
09:16 <rstavro> has joined #xwiki
09:17 <jstoldt> mflorea: will XWIKI-5897 also be in 2.7.1?
09:17 <sburjan> has joined #xwiki
09:17 <+mflorea> I can merge. let me do it
09:18 <jstoldt> great!
09:21 <+mflorea> done
09:23 <jstoldt> awesome! :)
09:23 <+mflorea> :)
09:26 <jstoldt> since you've already been around encoding trouble here, have you had a chance to check out if XWIKI-5866 is core related or a configuration error on my end? because i double checked and the encoding settings look just fine in Tomcat here.
09:37 <@cjdelisle> I thought of an interesting question
09:37 <@cjdelisle> When will there be a standardization of wiki syntax?
09:37 <+mflorea> jstoldt: In my case the export goes well when I'm connected to the OO server. The unicode characters are correctly displayed. I can't open though the rtf file exported when I'm not connected to OO server. But this is XWIKI-5251
09:39 <arkub> has joined #xwiki
09:51 <+sburjan> vmassol: FYI : I will work for few hours on a patch for selenium 2b1 migration
09:57 <jstoldt> mflorea: well, i really don't know why but by now i created three tickets about improper encoding of non-ascii chars. one is the RTF export, one concerns annotation display and the last one concerns auto complete. all of them i marked as related to XWIKI-5913
09:57 <jstoldt> like i said, i do run an OOo server and it should work fine for all i can see but it doesn't
09:58 <jstoldt> i changed OOo server encoding settings to UTF-8 following another ticket i found but it doesn't make a difference
10:07 <venkatesh> has quit
10:10 <maggo> has joined #xwiki
10:12 <maggo> Hi @ all, is someone awake who can tell me why i can not use passwords with more than 8 letters with LDAP Active Directory authentication ?
10:12 <vmassol> cjdelisle: re standardization of wiki syntax: never
10:12 <vmassol> sburjan: ok
10:12 <vmassol> cjdelisle: the last try was the creole syntax
10:12 <@cjdelisle> never? everything is standardised eventually.
10:13 <vmassol> atlassian thinks the standard will be HTML
10:13 <vmassol> ie that wiki syntax is not necessary
10:13 <vmassol> so the standardization = no wiki syntax, only wysiwyg editors
10:14 <@cjdelisle> Mediawiki had defined a defacto standard since it has such high usage.
10:14 <@cjdelisle> But it's still just defacto.
10:14 <vmassol> standard? no way
10:14 <vmassol> it's not even defacto at all
10:14 <vmassol> no other wiki uses it
10:14 <vmassol> I mean just check wikimatrix
10:14 <vmassol> you'll see they all have different syntaxes
10:15 <vmassol> some borrowing stuff from others
10:15 <vmassol> but they're still all different
10:15 <vmassol> xwiki is the most standard of all
10:15 <@cjdelisle> Yes but if you say "wiki" to most people they think "pedia". MW is itself a defacto standard.
10:15 <vmassol> depends
10:15 <vmassol> if you're talking about enterprise or general public
10:15 <vmassol> for enterprise wiki I don't think so
10:15 <+sburjan> xwiki is the most standard of all .. with what/who you make this comparation ?
10:16 <@cjdelisle> I agree there.
10:16 <vmassol> sburjan: because it supports all syntaxes :)
10:16 <+sburjan> oh
10:16 <vmassol> including creole which is the standard effort
10:16 <vmassol> (not all but quite a lot: about 6-7)
10:17 <@cjdelisle> IMO a standard is doomed unless the Wikipedia people are on board and that means it's some extension of MW.
10:17 <@cjdelisle> But these things always happen, it's just (IMO) a question of when the time is right.
10:18 <vmassol> implementations of creole: http://www.wikicreole.org/wiki/Engines
10:18 <+sburjan> from what I see these days, Wikipedia has financial issues .. I'm sure standardizing a syntax is not a top priority for them
10:19 <lucaa> has quit
10:19 <vmassol> updating http://wiki.wikicreole.org/Engines
10:19 <@cjdelisle> Hmm no mention of MW. They are the 800lb gorilla. Without them a standard isn't likely to go far.
10:19 <vmassol> actually it's already up to date, someone added xwiki cool :)
10:19 <venkatesh> has joined #xwiki
10:21 <vmassol> cjdelisle: there's another way: merge wikimodel/xwiki rendering and Eclipse's wikitext
10:21 <vmassol> (jason van zyl proposed to put us in contact to do this merge - I'm waiting for thomasM to be back to start a discussion about it)
10:22 <@cjdelisle> That's interesting.
10:23 <lucaa> has joined #xwiki
10:23 <jstoldt> question is if a standard is even necessary for wiki syntax because it's basically just a dialect of HTML, since all the markup will be changed to HTML at some point
10:24 <@cjdelisle> If you know MW syntax, that doesn't translate to knowing XWiki syntax or FosWiki.
10:25 <@cjdelisle> And everyone supports a set of converters of one type or another so pages can be imported.
10:25 <jstoldt> true. nevertheless do i think if it's just a dialect it needs to be flexible to fit the developers needs for change at all times
10:26 <jstoldt> so no standards and every software sticks to what they think is best to have a nicely working piece of software
10:28 <jstoldt> especially for corporate software...
10:28 <@cjdelisle> There is a long history of software developers locking their customers in to a given standard.
10:29 <jstoldt> if you just check out CAD software you'll notice that the many attempts to create file types supported across platforms eventually failed to be up to the job
10:29 <vmassol> cjdelisle: well xwiki is not doing that I think
10:29 <jstoldt> hold on
10:29 <@cjdelisle> No, I don't think there is any mallice in the wiki market. Just inefficency.
10:30 <+sburjan> vmassol: best way to to all ui-tests and see which fail due to Selenium2b1 migrations would be to run them from console, not eclipse, right ?
10:30 <vmassol> cjdelisle: pb is that for ex xwiki's syntax goes way beyond most other wiki syntaxes
10:30 <vmassol> that's why XWiki Syntax 2.0 is not exactly Creole BTW
10:30 <vmassol> it's Creole + more
10:31 <@cjdelisle> Ahh I see.
10:31 <@cjdelisle> So there was a standard that was followed. It just was not a fruitful standard.
10:31 <vmassol> sburjan: you can start with eclipse but console is the definitive answer indeed
10:31 <vmassol> cjdelisle: you can say that but it's tricky
10:32 <vmassol> since several syntaes voluntarily don't go too far to not make the syntax too complex
10:32 <vmassol> and deciding up to where to push it depends on one's judgment....
10:32 <vmassol> some do wiki syntax + HTML for more complex stuff
10:32 <vmassol> (I think mediawiki does that to some extent)
10:33 <@cjdelisle> Yes. that and "pseudo html"
10:33 <vmassol> http://www.wikimatrix.org/compare/MediaWiki+XWiki
10:33 <+sburjan> i get a tons of failures in eclipse when running the whole project, but then I select packages individually and run them, they pass .. I get something like : org.openqa.selenium.WebDriverException: org.apache.http.conn.HttpHostConnectException: Connection to http://127.0.0.1:7055 refused
10:34 <+sburjan> and I've never seen this before. So something happens when I run the whole project
10:37 <@cjdelisle> MW is a bit like Windows. It has the power to write defacto standards simply by the sheer number of installations and it (apparently) has no interest in working with anyone else.
10:38 <jstoldt> true but like i said, i don't there really is a need for a hard standard.
10:39 <vmassol> jstoldt: I don't think there's a need either and i don't think there'll ever be one
10:39 <vmassol> btw cjdelisle look at programming languages
10:39 <vmassol> there isn't a standard
10:39 <vmassol> so it doesn't always end up with a standard
10:39 <@cjdelisle> Hah nope. There are a lot of them.
10:40 <vmassol> my only question is whether wiki syntax is going to survive in the end or not
10:40 <@cjdelisle> But at least Sun java == IBM Java and GNU C ~= Borland C
10:40 <vmassol> atlassian thinks it won't and several other wikis are using HTML as their syntax too
10:40 <vmassol> well all wiki syntaxes are close too ;)
10:40 <jstoldt> vmassol: there is. at least for C and i think C++
10:41 <vmassol> jstoldt: and cobol and java and python and ruby and php and all others....
10:41 <jstoldt> difference is that the interpreters/compilers usually don't all do the same
10:41 <jstoldt> so there are few compilers that stick to the standard
10:42 <@cjdelisle> But to the credit of the programming language people, they do write books with the standard and there end up being multiple implementations.
10:42 <jstoldt> which is why basically i don't think a standard for wiki syntax is necessary... the interpreter will always do it a little different anyway
10:42 <vmassol> the best way to have a standard impl is by making it open source :)
10:42 <vmassol> so people reuse it instead of rewriting it
10:44 <jstoldt> vmassol: not even then, IMO.
10:44 <jstoldt> it's all about concepts, ideas and needs
10:44 <jstoldt> if they want to achieve something special they'll write it a little different
10:44 <jstoldt> or make a fork or something
10:44 <@cjdelisle> The question over whether wiki syntax will live on is interesting.
10:44 <jstoldt> whatever syntax is featured in a software that lives on
10:44 <jstoldt> i think that is about the best answer there is
10:44 <vmassol> yeah and I'm curious about the answer since we've invested a lot in it
10:44 <vmassol> (wiki syntax)
10:44 <vmassol> we = xwiki dev team
10:44 <jstoldt> and a great job you did
10:44 <@cjdelisle> I tend to think yes because it is old. It was bb code and blog syntax.
10:45 <jstoldt> well, even bb code lives on, doesn't it?
10:45 <@cjdelisle> Usually the goal of a syntax is to prevent people saying <script>
10:45 <jstoldt> despite it being more painful to use than wiki syntax
10:46 <vmassol> brb
10:46 <jstoldt> ^^
10:46 <@cjdelisle> Writing plain html has the advantage that it is indeed a standardized syntax. Not easy to write though.
10:47 <jstoldt> exactly
10:47 <+sburjan> vmassol: which way is the bet way to run tests from maven ? mvn surefire:tests ?
10:47 <jstoldt> but even with the standardized HTML you see all kinds of browsers screwing up... even the OpenSource ones
10:47 <+sburjan> because I am not able to run them, don't know why
10:48 <jstoldt> btw, this is why i prefere Opera... because they actually stick to the standard
10:48 <@cjdelisle> yes, implementations of html (and worse javascript) wander from the standard.
10:49 <@cjdelisle> Opera lol. It's a good js validator.
10:49 <jstoldt> ?
10:49 <@cjdelisle> Broken js will "work" in FF but in Opera it will fail.
10:49 <jstoldt> see! ;)
10:50 <@cjdelisle> But if the browser people rejected all invalid html then most of the web would be inaccessable.
10:51 <jstoldt> but that would also keep a lot of shitty websites from being available... most of them coded by 14yr old /b/ members for the sake of trolling, insulting and generally being a nuisance
10:54 <vmassol> sburjan: mvn test
10:54 <vmassol> there are some help tipcs here too: http://dev.xwiki.org/xwiki/bin/view/Community/Testing
10:55 <evalica> has joined #xwiki
10:57 <Enygma`> has joined #xwiki
10:59 <@cjdelisle> jstoldt: thought you might get a kick out of this being an engineer: http://www.youtube.com/watch?v=qg7hxaPAjnw
10:59 <@cjdelisle> implementation: http://www.youtube.com/watch?v=IX27qvPlHsI
11:01 <jstoldt> ah, one of those vertical wind mills...
11:01 <jstoldt> yeah, those are really neat from a technical POV.... where i come from we got shit load of those regular ones
11:02 <@cjdelisle> By rotating the vanes, the power extraction increases, friction decreases and you can use simple wind load calculations to figure power output.
11:02 <jstoldt> yeah. it's a very nice concept
11:02 <jstoldt> just like tide turbines
11:03 <@cjdelisle> Also you have a means to kill it by knocking off the timing or by allowing one vaine to rotate 90 degrees out of allignment you can let all of the air strait through.
11:04 <jstoldt> true... might be benefitial on high wind speed... the regular ones need to be shutdown when too much wind blows...
11:04 <@cjdelisle> Yup. Lotta problems.
11:05 <@cjdelisle> *and it can run slow and still extract a lot of power. Basic wind load calculations tell just how much.
11:05 <jstoldt> well, for me personally the best concept for renewable energy is desertec
11:07 <@cjdelisle> How exactly does it work?
11:08 <@cjdelisle> Photo voltaic?
11:08 <jstoldt> not at all
11:08 <jstoldt> hold on a minute
11:08 <jstoldt> that's the beauty
11:11 <jstoldt> the basic idea is to have parabolic mirrors that mirror the light onto a pipe which contains some sort of oil
11:12 <jstoldt> the oil will be heated and transfer its energy onto water which will boil in order to produce steam
11:12 <jstoldt> the steam will then be treated like in any other given turbine power station
11:12 <jstoldt> and there you go with nice, clean and efficient energy
11:12 <@cjdelisle> Sounds feasable (more so than PV). Trouble is you need to carry that power a long way.
11:12 <jstoldt> much unlike photo voltaic which has an efficiency degree of "don't bother"
11:13 <@cjdelisle> heh
11:13 <jstoldt> well, thing about desertec is that they don't aim at the european market in the first place
11:13 <jstoldt> first priority is to provide the local northern american and spanish markerts
11:13 <jstoldt> *markets
11:14 <@cjdelisle> If you're in the big electricity biz, it sounds sensable enough.
11:14 <@cjdelisle> Windmills can be used by everyone.
11:14 <jstoldt> anyway, for all i read the capabilities of transfering power via DC got improved a lot so it's only a matter of money to build up the grid
11:15 <jstoldt> well, energy is way overprized these days
11:15 <@cjdelisle> Inverter technology is a pain though, constantly exploding when there's a lightning storm.
11:15 <jstoldt> prolly gonna be the next bubble to burst... but i should doubt it, though
11:16 <jstoldt> that's why you got great minds becoming engineers... messing with nature to the point even a thunderstorm is no more than a moskito bite to the power grid ;)
11:17 <@cjdelisle> I studied the feasability of parallel generation. It can be done, need to get specific permission from the power company though.
11:17 <jstoldt> what do you mean by that?
11:18 <@cjdelisle> parallel? Just hooking up a generator, synchronizing it and throwing the switch.
11:19 <@cjdelisle> IIRC less than 1 megawatt and they'll let you use the grid to synchronize the generator head.
11:20 <jstoldt> ah, i see
11:21 <jstoldt> yeah, my technical english is still far from being good enough... improving on it, though
11:21 <jstoldt> nothing trains you like translating design guidelines ;)
11:21 <@cjdelisle> Hmm
11:33 <mflorea> has quit
11:33 <+sburjan> vmassol: I have a problem. I am unable to run ui-tests from console if I put the se2b1 in pom.xml. If I put the old version, the tests start. do you have any idea ?
11:34 <vmassol> what error do you get?
11:34 <+sburjan> non. build successful, but the server doesn't start.
11:36 <vmassol> I can try it
11:37 <+sburjan> ok. modify your pom.xml from ui-tests and put 2.0b1 instead of 2.0a7 ,and try running a mvn clean install
11:40 <rrodriguez> has quit
11:41 <+sburjan> so basically I can't see which tests are failing due to migration :(
11:47 <+sburjan> I'm only interested how to get them started.. so if you find a way. just let me know
11:48 <vmassol> sburjan: I can reproduce. I'm debugging now
12:53 <rrodriguez> has joined #xwiki
12:56 <rrodriguez> has quit
12:57 <vmassol> sburjan: I know what's the pb. Testing something
12:58 <rrodriguez> has joined #xwiki
13:03 <rrodriguez> has quit
13:06 <vmassol> sburjan: ok fixed. Please svn up the tests-ui project
13:09 <rrodriguez> has joined #xwiki
13:09 <mbryant> has joined #xwiki
13:11 <mbryant> Good day all. My Company wants me to add a few javascript widgets to the current wysiwyg enviorment, is the current wysiwyg extendable, or will it accept plug ins? so on so forth. Basicly I would like to know if you guys know of a "best" way to approach this so i dont get waist deep and find out its not the best approach
13:15 <vmassol> mbryant: don't know if there's the information you're looking for there but you can check http://extensions.xwiki.org/xwiki/bin/view/Extension/WYSIWYG+Editor+Module
13:15 <vmassol> it's supposed to contain the architecture of the wysiwyg editor
13:16 <vmassol> some other info: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/WysiwygEditor
13:18 <evalica> has quit
13:21 <mbryant> vmassol: thanks I think that is a good starting point, I was not sure if the best way to do it would be to create an app/maco and then add a button to that, or what. The big dream is to allow something like the mindmap where it is playable, and live content (i think the mind map is that way, I have not looked in a while)
13:21 <rrodriguez> has quit
13:22 <Enygma`> has quit
13:22 <DrLou> has joined #xwiki
13:23 <+sburjan> vmassol: what was the problem ?
13:23 <vmassol> sburjan: check the commit
13:24 <+sburjan> vmassol: I would had never guessed that. I see you changed the pom.xml to Se2b1 .. so you made my job easier, now I can see directly from hudson which tests will fail due to migration ;)
13:27 <+sburjan> the more interesting part is how you managed to figure this out :)
13:28 <vmassol> here it goes:
13:28 <vmassol> 1) I check mvn 3 vs mvn 2.2.1
13:28 <maggo> Hi @ all, is someone here who can tell me why i can not use passwords with more than 8 letters with LDAP Active Directory authentication ?
13:28 <vmassol> same result
13:29 <vmassol> 2) I saw the problem was that tests were not found so I googled about issues with cpsuite and maven
13:29 <mflorea> has joined #xwiki
13:29 <vmassol> found this one: http://jira.codehaus.org/browse/MNG-4861
13:29 <vmassol> created by me a while ago ;)
13:30 <+sburjan> lol :)
13:30 <vmassol> so that lead me to 3) check the dependency since you said it was working with 2.0a7 but not 2.0b1
13:30 <vmassol> so I ran mvn dependency:tree
13:30 <vmassol> and saw the testng dep
13:30 <vmassol> I ran mvn dependency:tree on 2.0a7 and compared
13:30 <+sburjan> so it seems there is a big diff between a7 and b1, not just API methods, also dependencies
13:30 <vmassol> saw that there no dep on junit or testng in 2.0a7
13:31 <vmassol> so I exlucded it assuming it could lead surefire to pick up tetsng automatically
13:31 <vmassol> see http://maven.apache.org/plugins/maven-surefire-plugin/testng.html
13:31 <+sburjan> very clever :)
13:31 <+sburjan> I would never thought of that
13:31 <vmassol> (having tetsng in your cp makes surefire runs testng)
13:31 <vmassol> IMO it's an error that they packaged testng in 2.0b1
13:32 <vmassol> i haven't reported it to them but it should be reported to them
13:32 <+sburjan> still don't know what is cp
13:33 <vmassol> http://tinyurl.com/4kjpm4z
13:34 <+sburjan> I'll try to report this to selenium guys, but i'm not sure I can give them all the needed details
13:37 <+sburjan> it seems the migration didn't affect anything
13:37 <+sburjan> same flickering tests as before
13:43 <maggo> Hi all i don't know if someone has read my question here for passwords with more than 8 letters. I figured out that only if the user has a password with more than 8 letter on the first login on the xwiki i get the authentication error. If i set the password to less than 8 letters an log in it works. changing the passwort after that to more than 8 letters works too. delete the user and login...
13:43 <maggo> ...in again works too. So i think only on the very first of all logins xwiki has a problem with authentication. Any idea?
13:52 <@cjdelisle> Sorry about not replying, what is the error message you get in the log?
13:54 <@cjdelisle> I seem to recall something about 8 character passwords having to do with the fact that Sun thought that a good way to stop governments from preventing importation of java was to hobble it's encryption facilities.
13:58 <@cjdelisle> Kind of silly since their limitation of 128 bits is plenty for "unbreakable" encryption if implemented right but ruinously insecure with some implementations IE: PKCS#12.
14:00 <@cjdelisle> Anyway the solution was to install something called an "Unlimited Strength Jurisdiction Policy File"
14:05 <venkatesh> has quit
14:08 <+sburjan> vmassol: is there any way to run tests that are in a certain class ? For example I want to run all methods from UeserProfileTest
14:09 <@cjdelisle> -Dtest=UserProfileTest in the command line
14:11 <+sburjan> thanks
14:12 <@cjdelisle> no pb
14:21 <venkatesh> has joined #xwiki
14:30 <venkatesh> has quit
14:32 <evalica> has joined #xwiki
14:34 <sdumitriu> has joined #xwiki
14:43 <jvelociter> has quit
14:43 <jvelociter> has joined #xwiki
14:47 <mflorea1> has joined #xwiki
14:49 <jstoldt> hmmm i thought having an internee while being an internee myself would be a lot more fun...
14:49 <sburjan> has quit
14:49 <mflorea> has quit
14:54 <maggo> cjdelisle: Sorry for my delay ;-) my error log you can see here http://pastebin.com/0uBXTkDb
15:02 <sburjan> has joined #xwiki
15:16 <jstoldt> noticed a nice little bug
15:16 <jstoldt> or prolly not a bug but poor design
15:16 <jstoldt> when i add a TOC within a box and choose the scope to be "LOCAL" it will not show anything at all
15:18 <jstoldt> so heaving a TOC floatbox on the right only for the contents of heading level 1 (so level 2...6) will inevitably fail.
15:23 <@cjdelisle> http://forums.opensuse.org/get-help-here/network-internet/432183-ldap-password-truncated-8-a.html
15:23 <rrodriguez> has joined #xwiki
15:24 <@cjdelisle> ``One very bare install latter and I figured it out. Under User Management, Authentication, LDAP, Advanced Configuration, Administration Settings, Configure User Management Settings, User Configuration. I changed the Max Password Length from 8 to higher and the Password Hash to MD5. ''
15:26 <@cjdelisle> maggo: ^ looks like your problem might be on the ldap end.
15:28 <rrodriguez> has quit
15:35 <newbie|3> has joined #xwiki
15:38 <mbryant> has quit
15:43 <mflorea1> sdumitriu: I have this problem XWIKI-5895 it's quite serious and easy to reproduce. Do you know why XWikiHibernateStore.saveXWikiDoc would consider an existing document (that has an attachment) as new?
15:44 <mflorea1> line 516
15:45 <jstoldt> yeah, i did kind of notice that
15:45 <mflorea1> looks like "select xwikidoc.id from XWikiDocument as xwikidoc where xwikidoc.id = :id".uniqueResult() returns null
15:45 <@sdumitriu> mflorea1: I'll check
15:45 <mflorea1> even though the document exists, and has an attachment
15:45 <@sdumitriu> Any browser, right?
15:46 <mflorea1> shouldn't matter, I checked with FF, but the issue is on the server side IMO
15:47 <mflorea1> what's interesting is that the document is fine after a server restart
15:50 <@sdumitriu> mflorea1: "Insert image" means upload a new image to the document, right?
15:50 <jstoldt> +1 for fixing this in 2.7.1 ;)
15:50 <mflorea1> sdumitriu: yep
15:50 <@sdumitriu> Insert + save worked well, now trying the exact steps (insert, preview, back, insert)
15:50 <+sburjan> vmassol: sorry for disturbing. What is the migration plan of selenium-tests -> ui-tets ?
15:51 <+sburjan> create a single patch / ported class ?
15:51 <mflorea1> yep, save worked well for me too, and cancel too, so there's something with preview, but to get there I debugged what happens when the second image is attached, and saw that the document is considered new (but maybe that's normal)
15:53 <@sdumitriu> mflorea1: OK, I got the error
16:04 <mflorea1> hmm, it doesn't seem related. I debugged again and now the document is not new but the image attachment is still not saved
16:04 <vmassol> sburjan: plan is http://dev.xwiki.org/xwiki/bin/view/Community/Testing#HGUItests
16:05 <vmassol> create a jira issue and attach patches to it.
16:06 <+sburjan> maybe you could create some jira subproject for this migration
16:06 <+sburjan> so all issues go there
16:06 <vmassol> we don't know if they're going to be implemented
16:06 <vmassol> better create when you do the work
16:06 <mflorea1> sdumitriu: strange because there's no exception thrown. FYI the WYSIWYG editor makes two requests: (1) submit to upload action (no exception) and then (2) get attachment info (image attachment is missing)
16:06 <+sburjan> so for each class I port I should create a new JIRA issue ?
16:06 <vmassol> no
16:07 <+sburjan> or I can aggregate more classes in one issue/patch
16:07 <vmassol> you can but you can also group several in one
16:07 <+sburjan> ok
16:07 <+sburjan> thanks
16:07 <+sburjan> ATM I finished with UserProfileTest
16:07 <vmassol> don't wait to have 10 classes ported though
16:07 <vmassol> the smallest the patch the easiest it is to apply
16:07 <abusenius> has joined #xwiki
16:07 <+sburjan> ok
16:07 <camaritei> has joined #xwiki
16:10 <vmassol> latest test results: 3 tests failing: http://hudson.xwiki.org/job/xwiki-product-enterprise-tests/319/testReport/?
16:12 <+sburjan> yes, I get those too on local
16:12 <+sburjan> org.xwiki.test.ui.RegisterTest.testRegisterDifferentPasswords is almost sure it's a flicker
16:12 <+sburjan> and the other two I think they are because of migration.. I will check those and try a fix
16:13 <+sburjan> I will investigate tomorrow, I'm in a meeting ATM
16:13 <+sburjan> basically this migration will break some things when working with inline and WYSIWYG
16:13 <vmassol> flicker need to be fixed too
16:13 <vmassol> I'll have a look at the flicker one
16:18 <@sdumitriu> mflorea1: If I flushCache(), the first image disappears as well
16:18 <Enygma`> has joined #xwiki
16:20 <rstavro> has quit
16:21 <+sburjan> vmassol: odd thing is that locally I have more tests failing than on hudson
16:21 <+sburjan> for example : testRegisterDifferentPasswords(org.xwiki.test.ui.RegisterTest)
16:21 <+sburjan> testRegisterEmptyUserName(org.xwiki.test.ui.RegisterTest)
16:21 <+sburjan> testRegisterEmptyPassword(org.xwiki.test.ui.administration.RegisterFromAdministrationTest)
16:21 <+sburjan> testRegisterEmptyUserName(org.xwiki.test.ui.administration.RegisterFromAdministrationTest)
16:21 <+sburjan> testRegisterInvalidEmail(org.xwiki.test.ui.administration.RegisterFromAdministrationTest)
16:22 <mflorea1> sdumitriu: if you flushCache after preview, right? otherwise the image remains attached.
16:22 <@sdumitriu> Yes
16:35 <@sdumitriu> mflorea1: Does the GWT service do anything else when uploading images?
16:35 <@sdumitriu> Does the syntax conversion filter do anything else other than changing the content?
16:37 <mflorea1> sdumitriu: the editor doesn't use a service to upload the images. It uses a HTML form submitted to /xwiki/bin/upload/Space/EditedPage . After the form is submitted it uses a GWT service to retrieve information about the uploaded attachment, but this service only reads information
16:37 <@sdumitriu> I see two requests
16:38 <@sdumitriu> http://localhost:8080/xwiki/resources/js/xwiki/wysiwyg/xwe/WikiService.gwtrpc before the upload
16:38 <@sdumitriu> Which returns //OK[1,["/xwiki/bin/upload/Sandbox/a14"],0,6]
16:38 <@sdumitriu> And indeed one after the normal upload
16:38 <@sdumitriu> Which returns //OK[3,2,1,["org.xwiki.gwt.wysiwyg.client.wiki.EntityConfig/166478544","false|-|attach|-|content.png","/xwiki/bin/download/Sandbox/a14/content.png"],0,6]
16:39 <@sdumitriu> I'm trying to reproduce the problem in the wiki editor
16:39 <mflorea1> that request asks for the upload URL
16:39 <mflorea1> and it made only once
16:40 <mflorea1> only for the first image
16:42 <@sdumitriu> OK, reproduced in the wiki editor + attach form
16:42 <@sdumitriu> So it's not related to the WYSIWYG at all
16:42 <@sdumitriu> Either preview, or back to edit
16:42 <mflorea1> conversion filter changes only the content, and only if the request contains some parameters, which isn't the case when the image is uploaded because a different form is used, not the form the includes the editor
16:59 <newbie|3> has quit
17:08 <@sdumitriu> mflorea1: I think I know what the problem is
17:09 <@sdumitriu> PreviewAction#render
17:09 <maggo> has quit
17:09 <@sdumitriu> There's a big if..else branch in there
17:09 <@sdumitriu> The code goes on the else branch
17:09 <@sdumitriu> Where tdoc is changed before it is cloned
17:10 <@sdumitriu> So this puts tdoc in the cache
17:10 <@sdumitriu> Since translations are not enabled, tdoc is actually doc
17:11 <@sdumitriu> The one from the cache
17:11 <@sdumitriu> And since it's marked as a translation, it has a different ID
17:11 <@sdumitriu> So attachments are added to the wrong document
17:16 <@sdumitriu> mflorea1: Indeed, works
17:16 <@sdumitriu> I'll commit the fix
17:16 <@sdumitriu> Can you write a test?
17:16 <mflorea1> you moved setLanguage and setTranslation after the clone call?
17:16 <@sdumitriu> Yes, to tdoc2
17:16 <mflorea1> a selenium test?
17:17 <@sdumitriu> Yes
17:17 <mflorea1> sure
17:17 <mflorea1> let me do it
17:18 <mflorea1> hmm, how did you reproduced the problem with the wiki editor? you used two browser tabs?
17:20 <mflorea1> sdumitriu: it might be easier to write the test for the wiki editor + upload form because the page objects are created.
17:21 <@sdumitriu> Yes
17:22 <@sdumitriu> But it's enough to:
17:22 <@sdumitriu> - upload something (will need to save the document first, since the upload form doesn't appear unless you use xpage=xpart&vm=attachments.vm
17:23 <@sdumitriu> - preview the document (just call /preview/, doesn't have to have any particular data)
17:23 <@sdumitriu> - upload something else
17:23 <@sdumitriu> The second attachment should be there only after the fi
17:23 <@sdumitriu> x
17:23 <mflorea1> let me try
17:24 <@sdumitriu> Should I also commit for 2.5 and 2.6?
17:25 <lucaa> has quit
17:26 <lucaa> has joined #xwiki
17:26 <mflorea1> I'm not sure why tdoc was modified before being cloned. There was a comment before that line saying something about the need to save the language. Can there be any regressions?
17:30 <@sdumitriu> No regressions
17:30 <@sdumitriu> Preview should never make cached changes
17:31 <lucaa> has quit
17:33 <@sdumitriu> Updated XWIKI-5895, please close it after/if you write the test
17:33 <@sdumitriu> Thanks
17:37 <mflorea1> sdumitriu: Thank you ;) I'll write the test
17:48 <abusenius> has quit
18:00 <camaritei> has left #xwiki
18:03 <sburjan> has quit
18:04 <@sdumitriu> Discovered a new smiley: (.Y.)
18:05 <jstoldt> has quit
18:09 <+jvelociter> it's not a frowning smiley, right ?
18:09 <@sdumitriu> No
18:10 <+jvelociter> :)
18:30 <rstavro> has joined #xwiki
18:59 <evalica> has quit
19:20 <abusenius> has joined #xwiki
19:41 <mflorea1> has quit
20:02 <arkub> has quit
20:11 <rstavro> has left #xwiki
21:54 <sburjan`> has quit
22:02 <sburjan`> has joined #xwiki
22:21 <sburjan`> has quit
22:46 <lucaa> has joined #xwiki
22:52 <sdumitriu> has quit
22:56 <rrodriguez> has joined #xwiki
23:53 <vmassol> has quit
00:55 <lucaa> has quit
01:09 <mflorea> has quit
01:52 <DrLou> has quit
03:07 <abusenius> has quit
03:44 <venkatesh> has joined #xwiki
03:44 <abusenius> has joined #xwiki
03:45 <venkatesh> has quit
03:45 <venkatesh> has joined #xwiki
03:51 <abusenius> has quit
06:04 <Denis> has quit
06:09 <Denis> has joined #xwiki
06:55 <vmassol> has joined #xwiki
07:00 <sdumitriu> has quit
08:04 <mflorea> has joined #xwiki
08:33 <sburjan`> has quit
08:33 <sburjan``> has joined #xwiki
08:33 <sburjan``> is now known as <sburjan`>
08:33 <sburjan`> has left #xwiki
08:33 <sburjan_> has joined #xwiki
08:33 <sburjan_> has quit
08:33 <sburjan`> has joined #xwiki
08:35 <sburjan``> has joined #xwiki
08:43 <sburjan``> has left #xwiki
08:44 <jstoldt> has joined #xwiki
08:44 <jstoldt> good morning everyone
08:44 <@cjdelisle> good morning
09:15 <lucaa> has joined #xwiki
09:16 <rstavro> has joined #xwiki
09:17 <jstoldt> mflorea: will XWIKI-5897 also be in 2.7.1?
09:17 <sburjan> has joined #xwiki
09:17 <+mflorea> I can merge. let me do it
09:18 <jstoldt> great!
09:21 <+mflorea> done
09:23 <jstoldt> awesome! :)
09:23 <+mflorea> :)
09:26 <jstoldt> since you've already been around encoding trouble here, have you had a chance to check out if XWIKI-5866 is core related or a configuration error on my end? because i double checked and the encoding settings look just fine in Tomcat here.
09:37 <@cjdelisle> I thought of an interesting question
09:37 <@cjdelisle> When will there be a standardization of wiki syntax?
09:37 <+mflorea> jstoldt: In my case the export goes well when I'm connected to the OO server. The unicode characters are correctly displayed. I can't open though the rtf file exported when I'm not connected to OO server. But this is XWIKI-5251
09:39 <arkub> has joined #xwiki
09:51 <+sburjan> vmassol: FYI : I will work for few hours on a patch for selenium 2b1 migration
09:57 <jstoldt> mflorea: well, i really don't know why but by now i created three tickets about improper encoding of non-ascii chars. one is the RTF export, one concerns annotation display and the last one concerns auto complete. all of them i marked as related to XWIKI-5913
09:57 <jstoldt> like i said, i do run an OOo server and it should work fine for all i can see but it doesn't
09:58 <jstoldt> i changed OOo server encoding settings to UTF-8 following another ticket i found but it doesn't make a difference
10:07 <venkatesh> has quit
10:10 <maggo> has joined #xwiki
10:12 <maggo> Hi @ all, is someone awake who can tell me why i can not use passwords with more than 8 letters with LDAP Active Directory authentication ?
10:12 <vmassol> cjdelisle: re standardization of wiki syntax: never
10:12 <vmassol> sburjan: ok
10:12 <vmassol> cjdelisle: the last try was the creole syntax
10:12 <@cjdelisle> never? everything is standardised eventually.
10:13 <vmassol> atlassian thinks the standard will be HTML
10:13 <vmassol> ie that wiki syntax is not necessary
10:13 <vmassol> so the standardization = no wiki syntax, only wysiwyg editors
10:14 <@cjdelisle> Mediawiki had defined a defacto standard since it has such high usage.
10:14 <@cjdelisle> But it's still just defacto.
10:14 <vmassol> standard? no way
10:14 <vmassol> it's not even defacto at all
10:14 <vmassol> no other wiki uses it
10:14 <vmassol> I mean just check wikimatrix
10:14 <vmassol> you'll see they all have different syntaxes
10:15 <vmassol> some borrowing stuff from others
10:15 <vmassol> but they're still all different
10:15 <vmassol> xwiki is the most standard of all
10:15 <@cjdelisle> Yes but if you say "wiki" to most people they think "pedia". MW is itself a defacto standard.
10:15 <vmassol> depends
10:15 <vmassol> if you're talking about enterprise or general public
10:15 <vmassol> for enterprise wiki I don't think so
10:15 <+sburjan> xwiki is the most standard of all .. with what/who you make this comparation ?
10:16 <@cjdelisle> I agree there.
10:16 <vmassol> sburjan: because it supports all syntaxes :)
10:16 <+sburjan> oh
10:16 <vmassol> including creole which is the standard effort
10:16 <vmassol> (not all but quite a lot: about 6-7)
10:17 <@cjdelisle> IMO a standard is doomed unless the Wikipedia people are on board and that means it's some extension of MW.
10:17 <@cjdelisle> But these things always happen, it's just (IMO) a question of when the time is right.
10:18 <vmassol> implementations of creole: http://www.wikicreole.org/wiki/Engines
10:18 <+sburjan> from what I see these days, Wikipedia has financial issues .. I'm sure standardizing a syntax is not a top priority for them
10:19 <lucaa> has quit
10:19 <vmassol> updating http://wiki.wikicreole.org/Engines
10:19 <@cjdelisle> Hmm no mention of MW. They are the 800lb gorilla. Without them a standard isn't likely to go far.
10:19 <vmassol> actually it's already up to date, someone added xwiki cool :)
10:19 <venkatesh> has joined #xwiki
10:21 <vmassol> cjdelisle: there's another way: merge wikimodel/xwiki rendering and Eclipse's wikitext
10:21 <vmassol> (jason van zyl proposed to put us in contact to do this merge - I'm waiting for thomasM to be back to start a discussion about it)
10:22 <@cjdelisle> That's interesting.
10:23 <lucaa> has joined #xwiki
10:23 <jstoldt> question is if a standard is even necessary for wiki syntax because it's basically just a dialect of HTML, since all the markup will be changed to HTML at some point
10:24 <@cjdelisle> If you know MW syntax, that doesn't translate to knowing XWiki syntax or FosWiki.
10:25 <@cjdelisle> And everyone supports a set of converters of one type or another so pages can be imported.
10:25 <jstoldt> true. nevertheless do i think if it's just a dialect it needs to be flexible to fit the developers needs for change at all times
10:26 <jstoldt> so no standards and every software sticks to what they think is best to have a nicely working piece of software
10:28 <jstoldt> especially for corporate software...
10:28 <@cjdelisle> There is a long history of software developers locking their customers in to a given standard.
10:29 <jstoldt> if you just check out CAD software you'll notice that the many attempts to create file types supported across platforms eventually failed to be up to the job
10:29 <vmassol> cjdelisle: well xwiki is not doing that I think
10:29 <jstoldt> hold on
10:29 <@cjdelisle> No, I don't think there is any mallice in the wiki market. Just inefficency.
10:30 <+sburjan> vmassol: best way to to all ui-tests and see which fail due to Selenium2b1 migrations would be to run them from console, not eclipse, right ?
10:30 <vmassol> cjdelisle: pb is that for ex xwiki's syntax goes way beyond most other wiki syntaxes
10:30 <vmassol> that's why XWiki Syntax 2.0 is not exactly Creole BTW
10:30 <vmassol> it's Creole + more
10:31 <@cjdelisle> Ahh I see.
10:31 <@cjdelisle> So there was a standard that was followed. It just was not a fruitful standard.
10:31 <vmassol> sburjan: you can start with eclipse but console is the definitive answer indeed
10:31 <vmassol> cjdelisle: you can say that but it's tricky
10:32 <vmassol> since several syntaes voluntarily don't go too far to not make the syntax too complex
10:32 <vmassol> and deciding up to where to push it depends on one's judgment....
10:32 <vmassol> some do wiki syntax + HTML for more complex stuff
10:32 <vmassol> (I think mediawiki does that to some extent)
10:33 <@cjdelisle> Yes. that and "pseudo html"
10:33 <vmassol> http://www.wikimatrix.org/compare/MediaWiki+XWiki
10:33 <+sburjan> i get a tons of failures in eclipse when running the whole project, but then I select packages individually and run them, they pass .. I get something like : org.openqa.selenium.WebDriverException: org.apache.http.conn.HttpHostConnectException: Connection to http://127.0.0.1:7055 refused
10:34 <+sburjan> and I've never seen this before. So something happens when I run the whole project
10:37 <@cjdelisle> MW is a bit like Windows. It has the power to write defacto standards simply by the sheer number of installations and it (apparently) has no interest in working with anyone else.
10:38 <jstoldt> true but like i said, i don't there really is a need for a hard standard.
10:39 <vmassol> jstoldt: I don't think there's a need either and i don't think there'll ever be one
10:39 <vmassol> btw cjdelisle look at programming languages
10:39 <vmassol> there isn't a standard
10:39 <vmassol> so it doesn't always end up with a standard
10:39 <@cjdelisle> Hah nope. There are a lot of them.
10:40 <vmassol> my only question is whether wiki syntax is going to survive in the end or not
10:40 <@cjdelisle> But at least Sun java == IBM Java and GNU C ~= Borland C
10:40 <vmassol> atlassian thinks it won't and several other wikis are using HTML as their syntax too
10:40 <vmassol> well all wiki syntaxes are close too ;)
10:40 <jstoldt> vmassol: there is. at least for C and i think C++
10:41 <vmassol> jstoldt: and cobol and java and python and ruby and php and all others....
10:41 <jstoldt> difference is that the interpreters/compilers usually don't all do the same
10:41 <jstoldt> so there are few compilers that stick to the standard
10:42 <@cjdelisle> But to the credit of the programming language people, they do write books with the standard and there end up being multiple implementations.
10:42 <jstoldt> which is why basically i don't think a standard for wiki syntax is necessary... the interpreter will always do it a little different anyway
10:42 <vmassol> the best way to have a standard impl is by making it open source :)
10:42 <vmassol> so people reuse it instead of rewriting it
10:44 <jstoldt> vmassol: not even then, IMO.
10:44 <jstoldt> it's all about concepts, ideas and needs
10:44 <jstoldt> if they want to achieve something special they'll write it a little different
10:44 <jstoldt> or make a fork or something
10:44 <@cjdelisle> The question over whether wiki syntax will live on is interesting.
10:44 <jstoldt> whatever syntax is featured in a software that lives on
10:44 <jstoldt> i think that is about the best answer there is
10:44 <vmassol> yeah and I'm curious about the answer since we've invested a lot in it
10:44 <vmassol> (wiki syntax)
10:44 <vmassol> we = xwiki dev team
10:44 <jstoldt> and a great job you did
10:44 <@cjdelisle> I tend to think yes because it is old. It was bb code and blog syntax.
10:45 <jstoldt> well, even bb code lives on, doesn't it?
10:45 <@cjdelisle> Usually the goal of a syntax is to prevent people saying <script>
10:45 <jstoldt> despite it being more painful to use than wiki syntax
10:46 <vmassol> brb
10:46 <jstoldt> ^^
10:46 <@cjdelisle> Writing plain html has the advantage that it is indeed a standardized syntax. Not easy to write though.
10:47 <jstoldt> exactly
10:47 <+sburjan> vmassol: which way is the bet way to run tests from maven ? mvn surefire:tests ?
10:47 <jstoldt> but even with the standardized HTML you see all kinds of browsers screwing up... even the OpenSource ones
10:47 <+sburjan> because I am not able to run them, don't know why
10:48 <jstoldt> btw, this is why i prefere Opera... because they actually stick to the standard
10:48 <@cjdelisle> yes, implementations of html (and worse javascript) wander from the standard.
10:49 <@cjdelisle> Opera lol. It's a good js validator.
10:49 <jstoldt> ?
10:49 <@cjdelisle> Broken js will "work" in FF but in Opera it will fail.
10:49 <jstoldt> see! ;)
10:50 <@cjdelisle> But if the browser people rejected all invalid html then most of the web would be inaccessable.
10:51 <jstoldt> but that would also keep a lot of shitty websites from being available... most of them coded by 14yr old /b/ members for the sake of trolling, insulting and generally being a nuisance
10:54 <vmassol> sburjan: mvn test
10:54 <vmassol> there are some help tipcs here too: http://dev.xwiki.org/xwiki/bin/view/Community/Testing
10:55 <evalica> has joined #xwiki
10:57 <Enygma`> has joined #xwiki
10:59 <@cjdelisle> jstoldt: thought you might get a kick out of this being an engineer: http://www.youtube.com/watch?v=qg7hxaPAjnw
10:59 <@cjdelisle> implementation: http://www.youtube.com/watch?v=IX27qvPlHsI
11:01 <jstoldt> ah, one of those vertical wind mills...
11:01 <jstoldt> yeah, those are really neat from a technical POV.... where i come from we got shit load of those regular ones
11:02 <@cjdelisle> By rotating the vanes, the power extraction increases, friction decreases and you can use simple wind load calculations to figure power output.
11:02 <jstoldt> yeah. it's a very nice concept
11:02 <jstoldt> just like tide turbines
11:03 <@cjdelisle> Also you have a means to kill it by knocking off the timing or by allowing one vaine to rotate 90 degrees out of allignment you can let all of the air strait through.
11:04 <jstoldt> true... might be benefitial on high wind speed... the regular ones need to be shutdown when too much wind blows...
11:04 <@cjdelisle> Yup. Lotta problems.
11:05 <@cjdelisle> *and it can run slow and still extract a lot of power. Basic wind load calculations tell just how much.
11:05 <jstoldt> well, for me personally the best concept for renewable energy is desertec
11:07 <@cjdelisle> How exactly does it work?
11:08 <@cjdelisle> Photo voltaic?
11:08 <jstoldt> not at all
11:08 <jstoldt> hold on a minute
11:08 <jstoldt> that's the beauty
11:11 <jstoldt> the basic idea is to have parabolic mirrors that mirror the light onto a pipe which contains some sort of oil
11:12 <jstoldt> the oil will be heated and transfer its energy onto water which will boil in order to produce steam
11:12 <jstoldt> the steam will then be treated like in any other given turbine power station
11:12 <jstoldt> and there you go with nice, clean and efficient energy
11:12 <@cjdelisle> Sounds feasable (more so than PV). Trouble is you need to carry that power a long way.
11:12 <jstoldt> much unlike photo voltaic which has an efficiency degree of "don't bother"
11:13 <@cjdelisle> heh
11:13 <jstoldt> well, thing about desertec is that they don't aim at the european market in the first place
11:13 <jstoldt> first priority is to provide the local northern american and spanish markerts
11:13 <jstoldt> *markets
11:14 <@cjdelisle> If you're in the big electricity biz, it sounds sensable enough.
11:14 <@cjdelisle> Windmills can be used by everyone.
11:14 <jstoldt> anyway, for all i read the capabilities of transfering power via DC got improved a lot so it's only a matter of money to build up the grid
11:15 <jstoldt> well, energy is way overprized these days
11:15 <@cjdelisle> Inverter technology is a pain though, constantly exploding when there's a lightning storm.
11:15 <jstoldt> prolly gonna be the next bubble to burst... but i should doubt it, though
11:16 <jstoldt> that's why you got great minds becoming engineers... messing with nature to the point even a thunderstorm is no more than a moskito bite to the power grid ;)
11:17 <@cjdelisle> I studied the feasability of parallel generation. It can be done, need to get specific permission from the power company though.
11:17 <jstoldt> what do you mean by that?
11:18 <@cjdelisle> parallel? Just hooking up a generator, synchronizing it and throwing the switch.
11:19 <@cjdelisle> IIRC less than 1 megawatt and they'll let you use the grid to synchronize the generator head.
11:20 <jstoldt> ah, i see
11:21 <jstoldt> yeah, my technical english is still far from being good enough... improving on it, though
11:21 <jstoldt> nothing trains you like translating design guidelines ;)
11:21 <@cjdelisle> Hmm
11:33 <mflorea> has quit
11:33 <+sburjan> vmassol: I have a problem. I am unable to run ui-tests from console if I put the se2b1 in pom.xml. If I put the old version, the tests start. do you have any idea ?
11:34 <vmassol> what error do you get?
11:34 <+sburjan> non. build successful, but the server doesn't start.
11:36 <vmassol> I can try it
11:37 <+sburjan> ok. modify your pom.xml from ui-tests and put 2.0b1 instead of 2.0a7 ,and try running a mvn clean install
11:40 <rrodriguez> has quit
11:41 <+sburjan> so basically I can't see which tests are failing due to migration :(
11:47 <+sburjan> I'm only interested how to get them started.. so if you find a way. just let me know
11:48 <vmassol> sburjan: I can reproduce. I'm debugging now
12:53 <rrodriguez> has joined #xwiki
12:56 <rrodriguez> has quit
12:57 <vmassol> sburjan: I know what's the pb. Testing something
12:58 <rrodriguez> has joined #xwiki
13:03 <rrodriguez> has quit
13:06 <vmassol> sburjan: ok fixed. Please svn up the tests-ui project
13:09 <rrodriguez> has joined #xwiki
13:09 <mbryant> has joined #xwiki
13:11 <mbryant> Good day all. My Company wants me to add a few javascript widgets to the current wysiwyg enviorment, is the current wysiwyg extendable, or will it accept plug ins? so on so forth. Basicly I would like to know if you guys know of a "best" way to approach this so i dont get waist deep and find out its not the best approach
13:15 <vmassol> mbryant: don't know if there's the information you're looking for there but you can check http://extensions.xwiki.org/xwiki/bin/view/Extension/WYSIWYG+Editor+Module
13:15 <vmassol> it's supposed to contain the architecture of the wysiwyg editor
13:16 <vmassol> some other info: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/WysiwygEditor
13:18 <evalica> has quit
13:21 <mbryant> vmassol: thanks I think that is a good starting point, I was not sure if the best way to do it would be to create an app/maco and then add a button to that, or what. The big dream is to allow something like the mindmap where it is playable, and live content (i think the mind map is that way, I have not looked in a while)
13:21 <rrodriguez> has quit
13:22 <Enygma`> has quit
13:22 <DrLou> has joined #xwiki
13:23 <+sburjan> vmassol: what was the problem ?
13:23 <vmassol> sburjan: check the commit
13:24 <+sburjan> vmassol: I would had never guessed that. I see you changed the pom.xml to Se2b1 .. so you made my job easier, now I can see directly from hudson which tests will fail due to migration ;)
13:27 <+sburjan> the more interesting part is how you managed to figure this out :)
13:28 <vmassol> here it goes:
13:28 <vmassol> 1) I check mvn 3 vs mvn 2.2.1
13:28 <maggo> Hi @ all, is someone here who can tell me why i can not use passwords with more than 8 letters with LDAP Active Directory authentication ?
13:28 <vmassol> same result
13:29 <vmassol> 2) I saw the problem was that tests were not found so I googled about issues with cpsuite and maven
13:29 <mflorea> has joined #xwiki
13:29 <vmassol> found this one: http://jira.codehaus.org/browse/MNG-4861
13:29 <vmassol> created by me a while ago ;)
13:30 <+sburjan> lol :)
13:30 <vmassol> so that lead me to 3) check the dependency since you said it was working with 2.0a7 but not 2.0b1
13:30 <vmassol> so I ran mvn dependency:tree
13:30 <vmassol> and saw the testng dep
13:30 <vmassol> I ran mvn dependency:tree on 2.0a7 and compared
13:30 <+sburjan> so it seems there is a big diff between a7 and b1, not just API methods, also dependencies
13:30 <vmassol> saw that there no dep on junit or testng in 2.0a7
13:31 <vmassol> so I exlucded it assuming it could lead surefire to pick up tetsng automatically
13:31 <vmassol> see http://maven.apache.org/plugins/maven-surefire-plugin/testng.html
13:31 <+sburjan> very clever :)
13:31 <+sburjan> I would never thought of that
13:31 <vmassol> (having tetsng in your cp makes surefire runs testng)
13:31 <vmassol> IMO it's an error that they packaged testng in 2.0b1
13:32 <vmassol> i haven't reported it to them but it should be reported to them
13:32 <+sburjan> still don't know what is cp
13:33 <vmassol> http://tinyurl.com/4kjpm4z
13:34 <+sburjan> I'll try to report this to selenium guys, but i'm not sure I can give them all the needed details
13:37 <+sburjan> it seems the migration didn't affect anything
13:37 <+sburjan> same flickering tests as before
13:43 <maggo> Hi all i don't know if someone has read my question here for passwords with more than 8 letters. I figured out that only if the user has a password with more than 8 letter on the first login on the xwiki i get the authentication error. If i set the password to less than 8 letters an log in it works. changing the passwort after that to more than 8 letters works too. delete the user and login...
13:43 <maggo> ...in again works too. So i think only on the very first of all logins xwiki has a problem with authentication. Any idea?
13:52 <@cjdelisle> Sorry about not replying, what is the error message you get in the log?
13:54 <@cjdelisle> I seem to recall something about 8 character passwords having to do with the fact that Sun thought that a good way to stop governments from preventing importation of java was to hobble it's encryption facilities.
13:58 <@cjdelisle> Kind of silly since their limitation of 128 bits is plenty for "unbreakable" encryption if implemented right but ruinously insecure with some implementations IE: PKCS#12.
14:00 <@cjdelisle> Anyway the solution was to install something called an "Unlimited Strength Jurisdiction Policy File"
14:05 <venkatesh> has quit
14:08 <+sburjan> vmassol: is there any way to run tests that are in a certain class ? For example I want to run all methods from UeserProfileTest
14:09 <@cjdelisle> -Dtest=UserProfileTest in the command line
14:11 <+sburjan> thanks
14:12 <@cjdelisle> no pb
14:21 <venkatesh> has joined #xwiki
14:30 <venkatesh> has quit
14:32 <evalica> has joined #xwiki
14:34 <sdumitriu> has joined #xwiki
14:43 <jvelociter> has quit
14:43 <jvelociter> has joined #xwiki
14:47 <mflorea1> has joined #xwiki
14:49 <jstoldt> hmmm i thought having an internee while being an internee myself would be a lot more fun...
14:49 <sburjan> has quit
14:49 <mflorea> has quit
14:54 <maggo> cjdelisle: Sorry for my delay ;-) my error log you can see here http://pastebin.com/0uBXTkDb
15:02 <sburjan> has joined #xwiki
15:16 <jstoldt> noticed a nice little bug
15:16 <jstoldt> or prolly not a bug but poor design
15:16 <jstoldt> when i add a TOC within a box and choose the scope to be "LOCAL" it will not show anything at all
15:18 <jstoldt> so heaving a TOC floatbox on the right only for the contents of heading level 1 (so level 2...6) will inevitably fail.
15:23 <@cjdelisle> http://forums.opensuse.org/get-help-here/network-internet/432183-ldap-password-truncated-8-a.html
15:23 <rrodriguez> has joined #xwiki
15:24 <@cjdelisle> ``One very bare install latter and I figured it out. Under User Management, Authentication, LDAP, Advanced Configuration, Administration Settings, Configure User Management Settings, User Configuration. I changed the Max Password Length from 8 to higher and the Password Hash to MD5. ''
15:26 <@cjdelisle> maggo: ^ looks like your problem might be on the ldap end.
15:28 <rrodriguez> has quit
15:35 <newbie|3> has joined #xwiki
15:38 <mbryant> has quit
15:43 <mflorea1> sdumitriu: I have this problem XWIKI-5895 it's quite serious and easy to reproduce. Do you know why XWikiHibernateStore.saveXWikiDoc would consider an existing document (that has an attachment) as new?
15:44 <mflorea1> line 516
15:45 <jstoldt> yeah, i did kind of notice that
15:45 <mflorea1> looks like "select xwikidoc.id from XWikiDocument as xwikidoc where xwikidoc.id = :id".uniqueResult() returns null
15:45 <@sdumitriu> mflorea1: I'll check
15:45 <mflorea1> even though the document exists, and has an attachment
15:45 <@sdumitriu> Any browser, right?
15:46 <mflorea1> shouldn't matter, I checked with FF, but the issue is on the server side IMO
15:47 <mflorea1> what's interesting is that the document is fine after a server restart
15:50 <@sdumitriu> mflorea1: "Insert image" means upload a new image to the document, right?
15:50 <jstoldt> +1 for fixing this in 2.7.1 ;)
15:50 <mflorea1> sdumitriu: yep
15:50 <@sdumitriu> Insert + save worked well, now trying the exact steps (insert, preview, back, insert)
15:50 <+sburjan> vmassol: sorry for disturbing. What is the migration plan of selenium-tests -> ui-tets ?
15:51 <+sburjan> create a single patch / ported class ?
15:51 <mflorea1> yep, save worked well for me too, and cancel too, so there's something with preview, but to get there I debugged what happens when the second image is attached, and saw that the document is considered new (but maybe that's normal)
15:53 <@sdumitriu> mflorea1: OK, I got the error
16:04 <mflorea1> hmm, it doesn't seem related. I debugged again and now the document is not new but the image attachment is still not saved
16:04 <vmassol> sburjan: plan is http://dev.xwiki.org/xwiki/bin/view/Community/Testing#HGUItests
16:05 <vmassol> create a jira issue and attach patches to it.
16:06 <+sburjan> maybe you could create some jira subproject for this migration
16:06 <+sburjan> so all issues go there
16:06 <vmassol> we don't know if they're going to be implemented
16:06 <vmassol> better create when you do the work
16:06 <mflorea1> sdumitriu: strange because there's no exception thrown. FYI the WYSIWYG editor makes two requests: (1) submit to upload action (no exception) and then (2) get attachment info (image attachment is missing)
16:06 <+sburjan> so for each class I port I should create a new JIRA issue ?
16:06 <vmassol> no
16:07 <+sburjan> or I can aggregate more classes in one issue/patch
16:07 <vmassol> you can but you can also group several in one
16:07 <+sburjan> ok
16:07 <+sburjan> thanks
16:07 <+sburjan> ATM I finished with UserProfileTest
16:07 <vmassol> don't wait to have 10 classes ported though
16:07 <vmassol> the smallest the patch the easiest it is to apply
16:07 <abusenius> has joined #xwiki
16:07 <+sburjan> ok
16:07 <camaritei> has joined #xwiki
16:10 <vmassol> latest test results: 3 tests failing: http://hudson.xwiki.org/job/xwiki-product-enterprise-tests/319/testReport/?
16:12 <+sburjan> yes, I get those too on local
16:12 <+sburjan> org.xwiki.test.ui.RegisterTest.testRegisterDifferentPasswords is almost sure it's a flicker
16:12 <+sburjan> and the other two I think they are because of migration.. I will check those and try a fix
16:13 <+sburjan> I will investigate tomorrow, I'm in a meeting ATM
16:13 <+sburjan> basically this migration will break some things when working with inline and WYSIWYG
16:13 <vmassol> flicker need to be fixed too
16:13 <vmassol> I'll have a look at the flicker one
16:18 <@sdumitriu> mflorea1: If I flushCache(), the first image disappears as well
16:18 <Enygma`> has joined #xwiki
16:20 <rstavro> has quit
16:21 <+sburjan> vmassol: odd thing is that locally I have more tests failing than on hudson
16:21 <+sburjan> for example : testRegisterDifferentPasswords(org.xwiki.test.ui.RegisterTest)
16:21 <+sburjan> testRegisterEmptyUserName(org.xwiki.test.ui.RegisterTest)
16:21 <+sburjan> testRegisterEmptyPassword(org.xwiki.test.ui.administration.RegisterFromAdministrationTest)
16:21 <+sburjan> testRegisterEmptyUserName(org.xwiki.test.ui.administration.RegisterFromAdministrationTest)
16:21 <+sburjan> testRegisterInvalidEmail(org.xwiki.test.ui.administration.RegisterFromAdministrationTest)
16:22 <mflorea1> sdumitriu: if you flushCache after preview, right? otherwise the image remains attached.
16:22 <@sdumitriu> Yes
16:35 <@sdumitriu> mflorea1: Does the GWT service do anything else when uploading images?
16:35 <@sdumitriu> Does the syntax conversion filter do anything else other than changing the content?
16:37 <mflorea1> sdumitriu: the editor doesn't use a service to upload the images. It uses a HTML form submitted to /xwiki/bin/upload/Space/EditedPage . After the form is submitted it uses a GWT service to retrieve information about the uploaded attachment, but this service only reads information
16:37 <@sdumitriu> I see two requests
16:38 <@sdumitriu> http://localhost:8080/xwiki/resources/js/xwiki/wysiwyg/xwe/WikiService.gwtrpc before the upload
16:38 <@sdumitriu> Which returns //OK[1,["/xwiki/bin/upload/Sandbox/a14"],0,6]
16:38 <@sdumitriu> And indeed one after the normal upload
16:38 <@sdumitriu> Which returns //OK[3,2,1,["org.xwiki.gwt.wysiwyg.client.wiki.EntityConfig/166478544","false|-|attach|-|content.png","/xwiki/bin/download/Sandbox/a14/content.png"],0,6]
16:39 <@sdumitriu> I'm trying to reproduce the problem in the wiki editor
16:39 <mflorea1> that request asks for the upload URL
16:39 <mflorea1> and it made only once
16:40 <mflorea1> only for the first image
16:42 <@sdumitriu> OK, reproduced in the wiki editor + attach form
16:42 <@sdumitriu> So it's not related to the WYSIWYG at all
16:42 <@sdumitriu> Either preview, or back to edit
16:42 <mflorea1> conversion filter changes only the content, and only if the request contains some parameters, which isn't the case when the image is uploaded because a different form is used, not the form the includes the editor
16:59 <newbie|3> has quit
17:08 <@sdumitriu> mflorea1: I think I know what the problem is
17:09 <@sdumitriu> PreviewAction#render
17:09 <maggo> has quit
17:09 <@sdumitriu> There's a big if..else branch in there
17:09 <@sdumitriu> The code goes on the else branch
17:09 <@sdumitriu> Where tdoc is changed before it is cloned
17:10 <@sdumitriu> So this puts tdoc in the cache
17:10 <@sdumitriu> Since translations are not enabled, tdoc is actually doc
17:11 <@sdumitriu> The one from the cache
17:11 <@sdumitriu> And since it's marked as a translation, it has a different ID
17:11 <@sdumitriu> So attachments are added to the wrong document
17:16 <@sdumitriu> mflorea1: Indeed, works
17:16 <@sdumitriu> I'll commit the fix
17:16 <@sdumitriu> Can you write a test?
17:16 <mflorea1> you moved setLanguage and setTranslation after the clone call?
17:16 <@sdumitriu> Yes, to tdoc2
17:16 <mflorea1> a selenium test?
17:17 <@sdumitriu> Yes
17:17 <mflorea1> sure
17:17 <mflorea1> let me do it
17:18 <mflorea1> hmm, how did you reproduced the problem with the wiki editor? you used two browser tabs?
17:20 <mflorea1> sdumitriu: it might be easier to write the test for the wiki editor + upload form because the page objects are created.
17:21 <@sdumitriu> Yes
17:22 <@sdumitriu> But it's enough to:
17:22 <@sdumitriu> - upload something (will need to save the document first, since the upload form doesn't appear unless you use xpage=xpart&vm=attachments.vm
17:23 <@sdumitriu> - preview the document (just call /preview/, doesn't have to have any particular data)
17:23 <@sdumitriu> - upload something else
17:23 <@sdumitriu> The second attachment should be there only after the fi
17:23 <@sdumitriu> x
17:23 <mflorea1> let me try
17:24 <@sdumitriu> Should I also commit for 2.5 and 2.6?
17:25 <lucaa> has quit
17:26 <lucaa> has joined #xwiki
17:26 <mflorea1> I'm not sure why tdoc was modified before being cloned. There was a comment before that line saying something about the need to save the language. Can there be any regressions?
17:30 <@sdumitriu> No regressions
17:30 <@sdumitriu> Preview should never make cached changes
17:31 <lucaa> has quit
17:33 <@sdumitriu> Updated XWIKI-5895, please close it after/if you write the test
17:33 <@sdumitriu> Thanks
17:37 <mflorea1> sdumitriu: Thank you ;) I'll write the test
17:48 <abusenius> has quit
18:00 <camaritei> has left #xwiki
18:03 <sburjan> has quit
18:04 <@sdumitriu> Discovered a new smiley: (.Y.)
18:05 <jstoldt> has quit
18:09 <+jvelociter> it's not a frowning smiley, right ?
18:09 <@sdumitriu> No
18:10 <+jvelociter> :)
18:30 <rstavro> has joined #xwiki
18:59 <evalica> has quit
19:20 <abusenius> has joined #xwiki
19:41 <mflorea1> has quit
20:02 <arkub> has quit
20:11 <rstavro> has left #xwiki
21:54 <sburjan`> has quit
22:02 <sburjan`> has joined #xwiki
22:21 <sburjan`> has quit
22:46 <lucaa> has joined #xwiki
22:52 <sdumitriu> has quit
22:56 <rrodriguez> has joined #xwiki
23:53 <vmassol> has quit
