IRC Archive for channel #xwiki
Last modified by Vincent Massol on 2012/10/18 19:11
arkub joined #xwiki at 00:25
arkub left at 00:25 (Read error: Connection reset by peer
jvdrean joined #xwiki at 00:56
lucaa joined #xwiki at 00:57
sdumitriu left at 02:48 (Quit: Leaving.
SvenDowideit left at 03:33 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 03:35
jvdrean left at 05:02 (Quit: Leaving.
Denis left at 07:01 (Read error: Connection reset by peer
mflorea joined #xwiki at 07:02
mflorea left at 07:03 (Client Quit
Denis joined #xwiki at 07:08
mflorea joined #xwiki at 07:13
lucaa left at 07:15 (Quit: Leaving.
lucaa joined #xwiki at 07:39
mflorea left at 08:20 (Quit: Leaving.
lucaa left at 08:21 (Quit: Leaving.
plunden1 joined #xwiki at 08:26
annacondan joined #xwiki at 08:32
mariusbutuc joined #xwiki at 08:35
npm left at 08:41 (Quit: Leaving.
npm joined #xwiki at 08:50
lucaa joined #xwiki at 08:57
silviar joined #xwiki at 09:37
mariusbutuc left #xwiki at 09:43
jvdrean joined #xwiki at 09:47
SvenDowideit left at 09:49 (Ping timeout: 246 seconds
SvenDowideit_ joined #xwiki at 09:49
SvenDowideit_ is now known as SvenDowideit ([email protected]
tmortagne joined #xwiki at 09:53
lucaa - (10:04): thanks silviar
SvenDowideit left at 10:08 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:08
Enygma` joined #xwiki at 10:11
glerouge joined #xwiki at 10:14
silviar - (10:14): lucaa: np :)
evalica joined #xwiki at 10:21
nickless joined #xwiki at 10:23
kibahop joined #xwiki at 10:27
SvenDowideit left at 10:28 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:28
sdumitriu joined #xwiki at 10:35
tmortagne - (10:35): CalebJamesDeLisl: looks like you fixed XAADMINISTRATION-135 by working on the other issue :)
SvenDowideit left at 10:36 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:37
oanat joined #xwiki at 10:40
tmortagne - (10:49): CalebJamesDeLisl: not fully actually
SvenDowideit left at 10:53 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:53
lucaa - (10:54): ji guys
lucaa - (10:54): hi
lucaa - (10:55): I just discovered that tags with
lucaa - (10:55): <div /> (compact version of closing tags)
lucaa - (10:55): written in {{html}} macros are not actually closed
lucaa - (10:56): marta setup a test page for this http://incubator.myxwiki.org/xwiki/bin/view/Sandbox/HTMLDivTest
lucaa - (10:56): is this normal? are we doing something wrong?
tmortagne - (10:59): Probably because of HtmlCleaner
tmortagne - (11:01): it's parsing the html content and then rewrite a xhtml content and i guess for some reason it's expending the div element
jvdrean - (11:02): I also have an issue with the rendering, I've reported it there http://incubator.myxwiki.org/xwiki/bin/view/Sandbox/HeadingsTest
tmortagne - (11:02): lucaa: pretty sure we have a jira issue talking about that
lucaa - (11:03): ok, I'll go vote for it
tmortagne - (11:03): jvdrean: that's normal, header is generated by the parser and at this point all the parser have is the macro
lucaa - (11:03): it can cause a lot of pain, especially when for a <div class="clearfloats" />...
tmortagne - (11:05): lucaa: http://jira.xwiki.org/jira/browse/XWIKI-4005 but it seems closed actually
tmortagne - (11:05): it's not exactly the same thing
SvenDowideit left at 11:05 (Ping timeout: 246 seconds
lucaa - (11:06): no, it's not at all the same thing
SvenDowideit joined #xwiki at 11:06
lucaa - (11:06): I don't want it expanded, I just want it closed
lucaa - (11:06): somehow
tmortagne - (11:06): but it's the same way to handle it
lucaa - (11:06): so that all the rest of the content doesn't go in
sdumitriu - (11:06): tmortagne: No, it's not at all related
lucaa - (11:06): handle at code level or what, i don't understand?
sdumitriu - (11:06): It's not that the div is closed
tmortagne - (11:07): sdumitriu: it is because htmlcleaner has no way to remember if the div was compacted or no when its wrinting the xhtml
tmortagne - (11:07): so it has to choose
tmortagne - (11:07): between always compacted or not
sdumitriu - (11:08): No, that's not the issue
jvdrean - (11:08): tmortagne: so I guess I have to use custom params to set correct ids
lucaa - (11:08): the issue is that it doesn't read it as closed, it seems
sdumitriu - (11:08): The div is not closed at all, it's not a problem between closing it short or long
lucaa - (11:08): so it doesn't serialize it as closed (compact or not, I don't case)
lucaa - (11:08): care
tmortagne - (11:08): jvdrean: actually i'm not sure you can do that, the custom param will probably be overwritten
jvdrean - (11:09): I've just tried and it works
tmortagne - (11:09): sdumitriu: wayit i missread something then, i don't see how it could not be closed indeed
tmortagne - (11:10): looks like a htmlcleaner important bug
tmortagne - (11:11): jvdrean: ok
jvdrean - (11:11): tmortagne: but should'nt we generate random ids when we miss title content ?
tmortagne - (11:13): jvdrean: the way it's done it's just take macros as no content and that it miss content, also the goal of header ids to to taget then, it's it's random it's pretty useless
tmortagne - (11:14): only thing i see is to generate an id in the xhtml renderer with the final content when the header id from the XDOM is empty
tmortagne - (11:14): that would do it i guess
tmortagne - (11:14): it's juts unusable by macros
tmortagne - (11:14): (which is why ids are generated by the parser)
SvenDowideit left at 11:15 (Ping timeout: 246 seconds
jvdrean - (11:15): so:
jvdrean - (11:15): 1/ generate random ids, not human readable but usable by macros like toc
jvdrean - (11:15): 2/ use final content in ids, human readable but not usable by toc
jvdrean - (11:16): do I get it right ?
jvdrean - (11:16): if so 1/ sounds better to me, I'm think tocs are way more used that manual anchors\
tmortagne - (11:17): lucaa: i just tested on http://tuska.myxwiki.org/xwiki/bin/view/viewTest/Div and i can't reproduce it, <div/> is transformed into <div></div> so it's more complex that that it seems
lucaa - (11:17): wait a sec
tmortagne - (11:18): jvdrean: it's not that much about human readable or not, it's about stable id
tmortagne - (11:18): when you have [[Space.Page#Hsomeheader]] link in another page
lucaa - (11:18): try html clean false
lucaa - (11:19): ( tmortagne )
tmortagne - (11:19): lucaa: clean=false is doing nothing
lucaa - (11:19): what?
lucaa - (11:19): you mean no difference, but not that it's unused attribute, right?
tmortagne - (11:20): lucaa: the gaol of clean=false is to make sure the content of the html macro is not touched and written as it is in the final xhtml content
lucaa - (11:20): well, we just discovered that it is actually closed in the content sent
lucaa - (11:20): but
lucaa - (11:20): compact closed
lucaa - (11:21): and browsers mess it up, most probably because we send it as text/html
tmortagne - (11:21): lucaa: hos do you test your content exactly ?
lucaa - (11:21): (which has no compact form)
lucaa - (11:21): just a second
tmortagne - (11:21): are you using "View selection source" ?*
lucaa - (11:21): no, I am looking at what's displayed
lucaa - (11:22): I have a <div class="clearfloats"
lucaa - (11:22): which if not closed it will hide anything inside
lucaa - (11:22): since it has fixed height of 1 px et all
tmortagne - (11:22): ok but that's "browsers mess" as you said it's not related to rendering
lucaa - (11:22): yes, we just discovered it now
tmortagne - (11:22): it's probably why htmlcleaner expednds div element i guess
lucaa - (11:22): but it doesn't
tmortagne - (11:22): because of some browser bugs
lucaa - (11:22): that's the potin
lucaa - (11:22): point
lucaa - (11:22): lemme test
lucaa - (11:22): more
tmortagne - (11:22): lucaa: it does
tmortagne - (11:23): again when you user clean=false you don't use htmlcleaner
lucaa - (11:23): no it doesn't anymore
lucaa - (11:23): I just modified your page
tmortagne - (11:23): it doe sin http://tuska.myxwiki.org/xwiki/bin/view/viewTest/Div ...
tmortagne - (11:25): lucaa: <div/> is transformed into <div></div> in http://tuska.myxwiki.org/xwiki/bin/view/viewTest/Div how is this not expended ?
lucaa - (11:25): tmortagne: if html clean=false _and_ a space between tagname and / it doesn't
lucaa - (11:25): I usually put the space when I put attributes because of browser behaviour
lucaa - (11:25): as in
lucaa - (11:26): i modified the page now to _not_ close it
tmortagne - (11:26): lucaa: why are you talking about clean=false ?
lucaa - (11:26): check it out
tmortagne - (11:26): again clean=false do nothing and that' sthe point of it
lucaa - (11:26): ok
lucaa - (11:26): mmh
lucaa - (11:26): I see what you mean
tmortagne - (11:27): if you put wrong html in it that's your choice we will not clean the html with clean=false...
lucaa - (11:27): it's not wrong :)
lucaa - (11:27): as per the xwiki doctype ;)
lucaa - (11:28): or then, we should have an xhtml macro
lucaa - (11:28): if html macro means I cannot write xhtml in it
tmortagne - (11:29): lucaa: why can't you write xhtml in it ?
tmortagne - (11:29): you can write whatever you want with clean=false
lucaa - (11:29): because it's sent as text/html to browser and browsers will mess it up, as in not understand compact closing
lucaa - (11:29): it's long discussion, I need to finish something else now...
lucaa - (11:29): conclusion is that it's not a rendering bug
tmortagne - (11:30): i don't see what html macro has to do with what XWIKI is sending
tmortagne - (11:30): it's two totally different things
tmortagne - (11:30): have a xhtml macro will not change anything
lucaa - (11:32): it could transform xhtml that I write in html that xwiki sends but that's wjhat html does too, except that clean="false" inhibits this
lucaa - (11:33): it's messy, as I see it
lucaa - (11:33): we should not send html
lucaa - (11:33): or browsers should look at the doctype
tmortagne - (11:34): i agree that we should not send html but AFAIK we are doing that because of IE6
tmortagne - (11:34): sdumitriu would know better
sdumitriu - (11:34): Yep, ie6
sdumitriu - (11:34): Not even 7 understands xhtml
lucaa - (11:34): I'll go back to my code...
tmortagne - (11:41): CalebJamesDeLisl: i don't know if it's due to your change to support custom UI but there is a "</p>" without any "<p>" before it at the begining of the admin section content
tmortagne - (11:43): we really should find some way to execute XHTML and WCAG validation on admin sections
tmortagne - (11:43): since we currently have invalid pages and don't know it
tmortagne - (11:55): jvdrean: WDYT about adding support for explicit list of URL to validate in the AbstractValidationTest ? That would cover admin use case and any other view we never validate (like testing some editors view for example)
tmortagne - (11:56): (we could also validate applications that showing different things depending on the parameters)
jvdrean - (11:57): tmortagne: I also had this in mind but it'll generate a _lot_ of work once activated
jvdrean - (11:57): (at least for WCAG)
jvdrean - (11:57): xhtml validation would be a good start
tmortagne - (11:57): that' measn we really need it then :)
jvdrean - (11:58): sure but it also means we need someone to take care of it
tmortagne - (11:58): not showing them doe snot make them not exist
tmortagne - (12:00): do you have code already for this ?
annacondan left at 12:00 (Quit: annacondan
jvdrean - (12:00): hum no, all the code has been commited
tmortagne - (12:00): anyway i'm talking about a explicit list for now we can put whatever we want in this list, having the support for it doe snot add anything
tmortagne - (12:01): jvdrean: are you sure, AFAICS AbstractValidationTest only parse xar files
tmortagne - (12:02): it does not look at any other property than pathToXWikiXar (and documentsToTest to filter it)
jvdrean - (12:02): I've been thinking about more testing but haven't coded what is needed for it
tmortagne - (12:02): ok
jvdrean - (12:03): (actually what I was thinking at first was calling the edit/ and inline/ actions on all the pages from the XAR)
tmortagne - (12:04): jvdrean: i don't think we really need to call edit on all pages since it would be more to test theses view UI validity i think
jvdrean - (12:05): yes I agree calling edit on a single page is enough, not true for inline though
anamarias joined #xwiki at 12:06
tmortagne - (12:06): yes inline is different since it's like calleing a rendering with different parameters
jvdrean left at 12:19 (Quit: Leaving.
lucaa left at 12:33 (Quit: Leaving.
lucaa joined #xwiki at 12:33
jvdrean joined #xwiki at 12:46
sdumitriu left at 12:46 (Ping timeout: 276 seconds
lucaa left at 12:46 (Read error: Connection reset by peer
lucaa joined #xwiki at 12:47
annacondan joined #xwiki at 13:03
jvdrean left at 13:06 (Quit: Leaving.
flaviusolaru joined #xwiki at 13:12
sdumitriu joined #xwiki at 13:21
CalebJamesDeLisl - (13:41): tmortagne: There are a few instances of {{html}}</p> .......... <p>{{html}} because with the old renderer that was the only way I could figure out to make valid html.
tmortagne - (13:42): CalebJamesDeLisl: i seen the clean=false html macros but still there is a html macro starting with a </p> without any <p> in a previous html macro
tmortagne - (13:43): here is a result of velocity macro with wiki=false: http://pastebin.com/nkrk3Du5
tmortagne - (13:44): CalebJamesDeLisl: unless there is a <p> outside of the admin section content ?
tmortagne - (13:44): i looked because html validator doe snot like that page
tmortagne - (13:45): (i'm talking about the velocity macro in ConfigurableClass page)
CalebJamesDeLisl - (13:47): Do you know where the offending </p> is? I think maybe the best thing to do is get rid of all {{html}}</p> and <p>{{/html}} since the new renderer doesn't need them.
tmortagne - (13:48): CalebJamesDeLisl: it's somewhere in ConfigurableClass code but i did not tried anything
jvdrean joined #xwiki at 13:49
CalebJamesDeLisl - (14:02): I removed the <p> and </p> tags and now registration section is valid.
CalebJamesDeLisl - (14:12): XAADMINISTRATION-138
arkub joined #xwiki at 14:16
tmortagne - (14:24): CalebJamesDeLisl: ok thanks, i just finished to add support for explicit URL in XE validation tool, i will add some admin related pages
kibahop left at 14:25 (Read error: Operation timed out
plunden1 left at 14:29 (Ping timeout: 240 seconds
bbc581 left at 14:42 (Remote host closed the connection
lucaa left #xwiki at 14:46
plunden1 joined #xwiki at 14:53
kibahop joined #xwiki at 14:54
bbc581 joined #xwiki at 14:55
annacondan left at 14:55 (Quit: annacondan
plunden1 left #xwiki at 15:09
evalica left at 15:22 (Quit: Leaving.
annacondan joined #xwiki at 15:23
plunden left at 15:49 (Ping timeout: 252 seconds
kibahop left #xwiki at 15:52
plunden joined #xwiki at 15:55
jvdrean left at 16:08 (Quit: Leaving.
tmortagne - (16:14): CalebJamesDeLisl: see http://hudson.xwiki.org/job/xwiki-product-enterprise-tests/com.xpn.xwiki.products$xwiki-enterprise-test-webstandards/764/testReport/com.xpn.xwiki.it/XHTMLValidationTest/Validating_XHTML_validity_for___xwiki_bin_admin_XWiki_XWikiPreferences_editor_globaladmin_section_Registration/
CalebJamesDeLisl - (16:16): Yes fixing now. That's great being able to test arbitrary urls!
tmortagne - (16:16): CalebJamesDeLisl: yep, i listed several but feel free to add more if you think of some important one
CalebJamesDeLisl - (16:17): I'll add register/XWiki/Register if it's not there.
tmortagne - (16:17): not there indeed
tmortagne - (16:18): except for admin i just added one of each editor wysiwyg/wiki/object/class
CalebJamesDeLisl - (16:23): Hmm, must be lots of violations...
tmortagne - (16:24): not that much for XHTML validation actually
CalebJamesDeLisl - (16:24): Should I just create a big issue for all Administration app violations?
tmortagne - (16:24): but a lot for Dutch Web Guidelines Validity
tmortagne - (16:24): CalebJamesDeLisl: yep shoud be ok
tmortagne - (16:25): if all is fixed before 2.3.1/2.4M1
jvdrean joined #xwiki at 16:28
CalebJamesDeLisl - (16:31): XAADMINISTRATION-139 XAADMINISTRATION-140
CalebJamesDeLisl - (16:32): XAADMINISTRATION-140
silviar left at 16:42 (Quit: Leaving.
Enygma` left at 16:43 (Quit: Leaving.
CalebJamesDeLisl - (17:00): The wcag tests run logged in as admin right?
oanat left at 17:01 (Quit: Leaving.
tmortagne - (17:02): CalebJamesDeLisl: yes
tmortagne - (17:02): same for xhtml tests (both are using the same code)
tmortagne - (17:02): (unless i broke something)
CalebJamesDeLisl - (17:02): Hmm, how to test login page? Maybe http://127.0.0.1 because cookies are different?
tmortagne - (17:03): CalebJamesDeLisl: there is no way to not be admin currently
tmortagne - (17:03): that would need some refactoring
CalebJamesDeLisl - (17:04): There is a selenium test where I load the url http://127.0.0.1:8080/xwiki/.... to drop admin status for one page load.
CalebJamesDeLisl - (17:04): (to check if a document is locked.)
tmortagne - (17:04): i mean there is no way in the validation framework
tmortagne - (17:04): you can do what you want in a custom test
CalebJamesDeLisl - (17:05): Do you know if it'd take a full url ?
tmortagne - (17:06): what ? urlsToTest ?
CalebJamesDeLisl - (17:07): Yup.
tmortagne - (17:07): CalebJamesDeLisl: for urlsToTest property: yes it supports absolute url
CalebJamesDeLisl - (17:07): You see any problem with using such a hack?
anamarias left at 17:08 (Quit: anamarias
tmortagne - (17:08): CalebJamesDeLisl: i don't understand this hack actually :)
CalebJamesDeLisl - (17:08): Login at localhost:8080/xwiki/.... then go to 127.0.0.1:8080/.... and your cookies are not sent.
CalebJamesDeLisl - (17:09): Because of cookie domain.
CalebJamesDeLisl - (17:09): Another way would be to switch to localhost.localdomain:8080/
tmortagne - (17:10): CalebJamesDeLisl: test framework is using basic authentication to make sure the user is admin
tmortagne - (17:10): whatever the URL you will be admin
tmortagne - (17:11): best i can propose now id to have a urlsToTestAsGuest property
tmortagne - (17:11): with a list of pages to test as guest instead of admin
tmortagne - (17:12): should not be difficult to do
CalebJamesDeLisl - (17:12): Yea, that looks like the only way.
tmortagne - (17:13): or probably rename urlsToTest as urlsToTestAsAdmin to be more clear
tmortagne - (17:14): doing this now
Denis left at 17:15 (Quit: Leaving.
tmortagne - (17:20): CalebJamesDeLisl: committed, urlsToTestAsAdmin and urlsToTestAsGuest
CalebJamesDeLisl - (17:21): Hey you added the guest part, cool. I was just looking at the code trying to make sense of it.
sdumitriu left at 17:23 (Ping timeout: 276 seconds
glerouge left at 17:57 (Quit: Leaving.
flaviusolaru left at 18:08 (Read error: Connection reset by peer
tmortagne left at 18:08 (Quit: Leaving.
annacondan left at 18:25 (Quit: annacondan
arkub left at 18:33 (Quit: Leaving
nickless left at 19:05 (Ping timeout: 264 seconds
jvdrean left at 19:26 (Quit: Leaving.
gvallarelli joined #xwiki at 19:54
gvallarelli - (19:54): good evening
CalebJamesDeLisl - (19:56): Good afternoon ;)
gvallarelli - (19:58): CalebJamesDeLisl: but you're american it's unfair :P
CalebJamesDeLisl - (19:58): Unfair? Because you get the early hours and we get the late ones?
CalebJamesDeLisl - (20:03): I can see how it would be tough having all morning and no afternoon ;)
gvallarelli - (20:06): I guess you should feel less tired
gvallarelli - (20:06): with a continous morning
CalebJamesDeLisl - (20:09): What have you been up to lately?
gvallarelli - (20:11): CalebJamesDeLisl: well nothing special except read read and read reaseach articles about agents
gvallarelli - (20:11): *research
CalebJamesDeLisl - (20:13): Heading in the AI direction?
gvallarelli - (20:14): I think that AI stuff are the more interesting topics on my CS department
CalebJamesDeLisl - (20:15): I was thinking if you had a program with a lot of settings (eg: bittorrent) you could have an AI program change the settings a little at a time and try for the best throughput.
CalebJamesDeLisl - (20:17): Would only work with a program where success is easily measured (like network throughput) and would only work when settings all were similar (change one unit of setting x has similar magnitude of effect as one unit of setting y)
gvallarelli - (20:19): CalebJamesDeLisl: actually I'm trying to figure out the mearning of my thesis topic
gvallarelli - (20:19): *meaning
CalebJamesDeLisl - (20:20): The topic was given to you? Is there a short description?
gvallarelli - (20:20): unfortunately no :D
gvallarelli - (20:21): anyway about multiagent coalition formation
gvallarelli - (20:21): I will implement an algorithm to create a coalition given a set of agents in C
gvallarelli - (20:21): It's a question of time and things will be more clear
gvallarelli - (20:21): I'm at the beginning
CalebJamesDeLisl - (20:24): C seems a bit obsolete, at least it's not C++
gvallarelli - (20:25): CalebJamesDeLisl: yep, C is more simple and straightforward than C++
gvallarelli - (20:25): also books on C++ are about 1000 pages :D it's a shame
gvallarelli - (20:26): not for the books but for the language so complex
CalebJamesDeLisl - (20:26): I remember hearing the language spec had grown to some ridiculous number of pages.
CalebJamesDeLisl - (20:27): My understanding of C++ is it was a failed attempt to apply object orientation to C.
CalebJamesDeLisl - (20:29): I don't really like C either because bad code can just construct a pointer and call any function it wants.
CalebJamesDeLisl - (20:30): And the runtime errors: " Crash at 0xABCDEF Message: I hate you. "
gvallarelli - (20:32): today I was studying about pointers and pointer arithmetic and fascinates me
gvallarelli - (20:32): I understand that could be a source of many errors
gvallarelli - (20:33): but one thing that I like is that there are so little and tiny abstractions, for the first time no one made the decisions for me on which things to abstract.
CalebJamesDeLisl - (20:34): Also you can easily port your code to run on nvidia GPUs which are fast. Like 500 cores.
gvallarelli - (20:35): eheh cool!
CalebJamesDeLisl - (20:38): If you're bored, there's a great video about making software secure by controlling what it can get references to http://www.youtube.com/watch?v=eL5o4PFuxTY
CalebJamesDeLisl - (20:39): If you can only have what you are given and you can load classes, you are pretty well boxed in.
CalebJamesDeLisl - (20:39): * and you can't load classes
CalebJamesDeLisl - (20:40): Constructing arbitrary pointers breaks that concept.
nickless joined #xwiki at 20:40
gvallarelli - (20:42): CalebJamesDeLisl: yep
gvallarelli - (20:43): I love subtitles
CalebJamesDeLisl - (20:43): A paper I read a while back was about cryptanalysis on the nvidia gpu. http://eecm.cr.yp.to/pc109-20090901.pdf
gvallarelli - (20:43): CalebJamesDeLisl: you study and read a lot of things, I hope you study CS :)
CalebJamesDeLisl - (20:44): I guess I do.
CalebJamesDeLisl - (20:45): Anyway what I found interesting about the paper was that you should have more threads than you have cores so when one blocks, the other can pick up.
gvallarelli - (20:45): CalebJamesDeLisl: cool video thank you!
CalebJamesDeLisl - (20:46): It's really neat stuff, there's a whole community around the concept of "object capability security"
gvallarelli - (20:50): CalebJamesDeLisl: interesting I'm watching it right now
anamarias joined #xwiki at 20:55
sdumitriu joined #xwiki at 20:56
gvallarelli - (21:35): CalebJamesDeLisl: I was thinking that using immutable data structures improve a lot security because it's impossible to corrupt data, so functional languages are more secure than side effect slave programming languages
CalebJamesDeLisl - (21:37): Looking up side effects, still a bit ambiguous, can you give me your definition?
gvallarelli - (21:40): side effects happens when you accomplish a task with a function for example and alter object state
gvallarelli - (21:40): substitute function with method
CalebJamesDeLisl - (21:43): Yes. I don't think it's a problem when a method alters the state of it's object, but it shouldn't be mutating parameters.
CalebJamesDeLisl - (21:43): Also IMO java should never have gotten the static keyword, there is no good use for it.
gvallarelli - (21:46): if something strange happens and for example a method don't complete his execution you could have a corrupted object with a very bad state
gvallarelli - (21:46): wait somewhere I bookmarked an interesting article about it
lucaa joined #xwiki at 21:46
gvallarelli - (21:48): CalebJamesDeLisl: is a little related to problems about altering state http://damienkatz.net/2006/04/error_code_vs_e.html
CalebJamesDeLisl - (21:48): Then where does the data go?
gvallarelli - (21:49): look at the article it's better explained than I could do :S
CalebJamesDeLisl - (21:50): Haha firefox. Javascript is safer, but I can still get it to segfault :P
gvallarelli - (21:51): :P
CalebJamesDeLisl - (22:02): "In every popular language, the updating of a variable means the previous state of that variable is destroyed, vanished, gone and you can't get it back"
CalebJamesDeLisl - (22:03): Not necessarily.
CalebJamesDeLisl - (22:08): "(photo of actual atom)" LOL
CalebJamesDeLisl - (22:11): Interesting article. He didn't mention copy-on-write!
CalebJamesDeLisl - (22:13): Our code is littered with document.clone() invocations so that document saving can't break the document in the cache.
CalebJamesDeLisl - (22:14): Prototype (javascript) Object.extend() returns the same object except changes to the new object are separate from the old one. AFAIK
CalebJamesDeLisl - (22:15): So it's like clone but it doesn't require so much processor horsepower and RAM.
gvallarelli - (22:24): interesting :)
gvallarelli - (22:25): got to go
gvallarelli - (22:25): good evening/night
CalebJamesDeLisl - (22:25): See ya. One last thought, security and error handling are the same.
gvallarelli - (22:25): goodbye CalebJamesDeLisl
gvallarelli - (22:26): thank you for the conversation
CalebJamesDeLisl - (22:26): Thank you.
gvallarelli left #xwiki at 22:26
CalebJamesDeLisl - (23:32): 'getAllMatchedUsers' in class com.xpn.xwiki.plugin.rightsmanager.RightsManagerUsersApi threw exception com.xpn.xwiki.XWikiException: Error number 3223 in 3: Exception while searching documents with SQL [, BaseObject as obj where doc.fullName=obj.name and doc.fullName<>? and obj.className=? order by doc.name asc]
CalebJamesDeLisl - (23:32): JDBCExceptionReporter - Operation requires a scrollable ResultSet, but this ResultSet is FORWARD_ONLY.
CalebJamesDeLisl - (23:32): getusers.vm
sdumitriu left at 23:33 (Ping timeout: 246 seconds
arkub left at 00:25 (Read error: Connection reset by peer
jvdrean joined #xwiki at 00:56
lucaa joined #xwiki at 00:57
sdumitriu left at 02:48 (Quit: Leaving.
SvenDowideit left at 03:33 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 03:35
jvdrean left at 05:02 (Quit: Leaving.
Denis left at 07:01 (Read error: Connection reset by peer
mflorea joined #xwiki at 07:02
mflorea left at 07:03 (Client Quit
Denis joined #xwiki at 07:08
mflorea joined #xwiki at 07:13
lucaa left at 07:15 (Quit: Leaving.
lucaa joined #xwiki at 07:39
mflorea left at 08:20 (Quit: Leaving.
lucaa left at 08:21 (Quit: Leaving.
plunden1 joined #xwiki at 08:26
annacondan joined #xwiki at 08:32
mariusbutuc joined #xwiki at 08:35
npm left at 08:41 (Quit: Leaving.
npm joined #xwiki at 08:50
lucaa joined #xwiki at 08:57
silviar joined #xwiki at 09:37
mariusbutuc left #xwiki at 09:43
jvdrean joined #xwiki at 09:47
SvenDowideit left at 09:49 (Ping timeout: 246 seconds
SvenDowideit_ joined #xwiki at 09:49
SvenDowideit_ is now known as SvenDowideit ([email protected]
tmortagne joined #xwiki at 09:53
lucaa - (10:04): thanks silviar
SvenDowideit left at 10:08 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:08
Enygma` joined #xwiki at 10:11
glerouge joined #xwiki at 10:14
silviar - (10:14): lucaa: np :)
evalica joined #xwiki at 10:21
nickless joined #xwiki at 10:23
kibahop joined #xwiki at 10:27
SvenDowideit left at 10:28 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:28
sdumitriu joined #xwiki at 10:35
tmortagne - (10:35): CalebJamesDeLisl: looks like you fixed XAADMINISTRATION-135 by working on the other issue :)
SvenDowideit left at 10:36 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:37
oanat joined #xwiki at 10:40
tmortagne - (10:49): CalebJamesDeLisl: not fully actually
SvenDowideit left at 10:53 (Ping timeout: 246 seconds
SvenDowideit joined #xwiki at 10:53
lucaa - (10:54): ji guys
lucaa - (10:54): hi
lucaa - (10:55): I just discovered that tags with
lucaa - (10:55): <div /> (compact version of closing tags)
lucaa - (10:55): written in {{html}} macros are not actually closed
lucaa - (10:56): marta setup a test page for this http://incubator.myxwiki.org/xwiki/bin/view/Sandbox/HTMLDivTest
lucaa - (10:56): is this normal? are we doing something wrong?
tmortagne - (10:59): Probably because of HtmlCleaner
tmortagne - (11:01): it's parsing the html content and then rewrite a xhtml content and i guess for some reason it's expending the div element
jvdrean - (11:02): I also have an issue with the rendering, I've reported it there http://incubator.myxwiki.org/xwiki/bin/view/Sandbox/HeadingsTest
tmortagne - (11:02): lucaa: pretty sure we have a jira issue talking about that
lucaa - (11:03): ok, I'll go vote for it
tmortagne - (11:03): jvdrean: that's normal, header is generated by the parser and at this point all the parser have is the macro
lucaa - (11:03): it can cause a lot of pain, especially when for a <div class="clearfloats" />...
tmortagne - (11:05): lucaa: http://jira.xwiki.org/jira/browse/XWIKI-4005 but it seems closed actually
tmortagne - (11:05): it's not exactly the same thing
SvenDowideit left at 11:05 (Ping timeout: 246 seconds
lucaa - (11:06): no, it's not at all the same thing
SvenDowideit joined #xwiki at 11:06
lucaa - (11:06): I don't want it expanded, I just want it closed
lucaa - (11:06): somehow
tmortagne - (11:06): but it's the same way to handle it
lucaa - (11:06): so that all the rest of the content doesn't go in
sdumitriu - (11:06): tmortagne: No, it's not at all related
lucaa - (11:06): handle at code level or what, i don't understand?
sdumitriu - (11:06): It's not that the div is closed
tmortagne - (11:07): sdumitriu: it is because htmlcleaner has no way to remember if the div was compacted or no when its wrinting the xhtml
tmortagne - (11:07): so it has to choose
tmortagne - (11:07): between always compacted or not
sdumitriu - (11:08): No, that's not the issue
jvdrean - (11:08): tmortagne: so I guess I have to use custom params to set correct ids
lucaa - (11:08): the issue is that it doesn't read it as closed, it seems
sdumitriu - (11:08): The div is not closed at all, it's not a problem between closing it short or long
lucaa - (11:08): so it doesn't serialize it as closed (compact or not, I don't case)
lucaa - (11:08): care
tmortagne - (11:08): jvdrean: actually i'm not sure you can do that, the custom param will probably be overwritten
jvdrean - (11:09): I've just tried and it works
tmortagne - (11:09): sdumitriu: wayit i missread something then, i don't see how it could not be closed indeed
tmortagne - (11:10): looks like a htmlcleaner important bug
tmortagne - (11:11): jvdrean: ok
jvdrean - (11:11): tmortagne: but should'nt we generate random ids when we miss title content ?
tmortagne - (11:13): jvdrean: the way it's done it's just take macros as no content and that it miss content, also the goal of header ids to to taget then, it's it's random it's pretty useless
tmortagne - (11:14): only thing i see is to generate an id in the xhtml renderer with the final content when the header id from the XDOM is empty
tmortagne - (11:14): that would do it i guess
tmortagne - (11:14): it's juts unusable by macros
tmortagne - (11:14): (which is why ids are generated by the parser)
SvenDowideit left at 11:15 (Ping timeout: 246 seconds
jvdrean - (11:15): so:
jvdrean - (11:15): 1/ generate random ids, not human readable but usable by macros like toc
jvdrean - (11:15): 2/ use final content in ids, human readable but not usable by toc
jvdrean - (11:16): do I get it right ?
jvdrean - (11:16): if so 1/ sounds better to me, I'm think tocs are way more used that manual anchors\
tmortagne - (11:17): lucaa: i just tested on http://tuska.myxwiki.org/xwiki/bin/view/viewTest/Div and i can't reproduce it, <div/> is transformed into <div></div> so it's more complex that that it seems
lucaa - (11:17): wait a sec
tmortagne - (11:18): jvdrean: it's not that much about human readable or not, it's about stable id
tmortagne - (11:18): when you have [[Space.Page#Hsomeheader]] link in another page
lucaa - (11:18): try html clean false
lucaa - (11:19): ( tmortagne )
tmortagne - (11:19): lucaa: clean=false is doing nothing
lucaa - (11:19): what?
lucaa - (11:19): you mean no difference, but not that it's unused attribute, right?
tmortagne - (11:20): lucaa: the gaol of clean=false is to make sure the content of the html macro is not touched and written as it is in the final xhtml content
lucaa - (11:20): well, we just discovered that it is actually closed in the content sent
lucaa - (11:20): but
lucaa - (11:20): compact closed
lucaa - (11:21): and browsers mess it up, most probably because we send it as text/html
tmortagne - (11:21): lucaa: hos do you test your content exactly ?
lucaa - (11:21): (which has no compact form)
lucaa - (11:21): just a second
tmortagne - (11:21): are you using "View selection source" ?*
lucaa - (11:21): no, I am looking at what's displayed
lucaa - (11:22): I have a <div class="clearfloats"
lucaa - (11:22): which if not closed it will hide anything inside
lucaa - (11:22): since it has fixed height of 1 px et all
tmortagne - (11:22): ok but that's "browsers mess" as you said it's not related to rendering
lucaa - (11:22): yes, we just discovered it now
tmortagne - (11:22): it's probably why htmlcleaner expednds div element i guess
lucaa - (11:22): but it doesn't
tmortagne - (11:22): because of some browser bugs
lucaa - (11:22): that's the potin
lucaa - (11:22): point
lucaa - (11:22): lemme test
lucaa - (11:22): more
tmortagne - (11:22): lucaa: it does
tmortagne - (11:23): again when you user clean=false you don't use htmlcleaner
lucaa - (11:23): no it doesn't anymore
lucaa - (11:23): I just modified your page
tmortagne - (11:23): it doe sin http://tuska.myxwiki.org/xwiki/bin/view/viewTest/Div ...
tmortagne - (11:25): lucaa: <div/> is transformed into <div></div> in http://tuska.myxwiki.org/xwiki/bin/view/viewTest/Div how is this not expended ?
lucaa - (11:25): tmortagne: if html clean=false _and_ a space between tagname and / it doesn't
lucaa - (11:25): I usually put the space when I put attributes because of browser behaviour
lucaa - (11:25): as in
lucaa - (11:26): i modified the page now to _not_ close it
tmortagne - (11:26): lucaa: why are you talking about clean=false ?
lucaa - (11:26): check it out
tmortagne - (11:26): again clean=false do nothing and that' sthe point of it
lucaa - (11:26): ok
lucaa - (11:26): mmh
lucaa - (11:26): I see what you mean
tmortagne - (11:27): if you put wrong html in it that's your choice we will not clean the html with clean=false...
lucaa - (11:27): it's not wrong :)
lucaa - (11:27): as per the xwiki doctype ;)
lucaa - (11:28): or then, we should have an xhtml macro
lucaa - (11:28): if html macro means I cannot write xhtml in it
tmortagne - (11:29): lucaa: why can't you write xhtml in it ?
tmortagne - (11:29): you can write whatever you want with clean=false
lucaa - (11:29): because it's sent as text/html to browser and browsers will mess it up, as in not understand compact closing
lucaa - (11:29): it's long discussion, I need to finish something else now...
lucaa - (11:29): conclusion is that it's not a rendering bug
tmortagne - (11:30): i don't see what html macro has to do with what XWIKI is sending
tmortagne - (11:30): it's two totally different things
tmortagne - (11:30): have a xhtml macro will not change anything
lucaa - (11:32): it could transform xhtml that I write in html that xwiki sends but that's wjhat html does too, except that clean="false" inhibits this
lucaa - (11:33): it's messy, as I see it
lucaa - (11:33): we should not send html
lucaa - (11:33): or browsers should look at the doctype
tmortagne - (11:34): i agree that we should not send html but AFAIK we are doing that because of IE6
tmortagne - (11:34): sdumitriu would know better
sdumitriu - (11:34): Yep, ie6
sdumitriu - (11:34): Not even 7 understands xhtml
lucaa - (11:34): I'll go back to my code...
tmortagne - (11:41): CalebJamesDeLisl: i don't know if it's due to your change to support custom UI but there is a "</p>" without any "<p>" before it at the begining of the admin section content
tmortagne - (11:43): we really should find some way to execute XHTML and WCAG validation on admin sections
tmortagne - (11:43): since we currently have invalid pages and don't know it
tmortagne - (11:55): jvdrean: WDYT about adding support for explicit list of URL to validate in the AbstractValidationTest ? That would cover admin use case and any other view we never validate (like testing some editors view for example)
tmortagne - (11:56): (we could also validate applications that showing different things depending on the parameters)
jvdrean - (11:57): tmortagne: I also had this in mind but it'll generate a _lot_ of work once activated
jvdrean - (11:57): (at least for WCAG)
jvdrean - (11:57): xhtml validation would be a good start
tmortagne - (11:57): that' measn we really need it then :)
jvdrean - (11:58): sure but it also means we need someone to take care of it
tmortagne - (11:58): not showing them doe snot make them not exist
tmortagne - (12:00): do you have code already for this ?
annacondan left at 12:00 (Quit: annacondan
jvdrean - (12:00): hum no, all the code has been commited
tmortagne - (12:00): anyway i'm talking about a explicit list for now we can put whatever we want in this list, having the support for it doe snot add anything
tmortagne - (12:01): jvdrean: are you sure, AFAICS AbstractValidationTest only parse xar files
tmortagne - (12:02): it does not look at any other property than pathToXWikiXar (and documentsToTest to filter it)
jvdrean - (12:02): I've been thinking about more testing but haven't coded what is needed for it
tmortagne - (12:02): ok
jvdrean - (12:03): (actually what I was thinking at first was calling the edit/ and inline/ actions on all the pages from the XAR)
tmortagne - (12:04): jvdrean: i don't think we really need to call edit on all pages since it would be more to test theses view UI validity i think
jvdrean - (12:05): yes I agree calling edit on a single page is enough, not true for inline though
anamarias joined #xwiki at 12:06
tmortagne - (12:06): yes inline is different since it's like calleing a rendering with different parameters
jvdrean left at 12:19 (Quit: Leaving.
lucaa left at 12:33 (Quit: Leaving.
lucaa joined #xwiki at 12:33
jvdrean joined #xwiki at 12:46
sdumitriu left at 12:46 (Ping timeout: 276 seconds
lucaa left at 12:46 (Read error: Connection reset by peer
lucaa joined #xwiki at 12:47
annacondan joined #xwiki at 13:03
jvdrean left at 13:06 (Quit: Leaving.
flaviusolaru joined #xwiki at 13:12
sdumitriu joined #xwiki at 13:21
CalebJamesDeLisl - (13:41): tmortagne: There are a few instances of {{html}}</p> .......... <p>{{html}} because with the old renderer that was the only way I could figure out to make valid html.
tmortagne - (13:42): CalebJamesDeLisl: i seen the clean=false html macros but still there is a html macro starting with a </p> without any <p> in a previous html macro
tmortagne - (13:43): here is a result of velocity macro with wiki=false: http://pastebin.com/nkrk3Du5
tmortagne - (13:44): CalebJamesDeLisl: unless there is a <p> outside of the admin section content ?
tmortagne - (13:44): i looked because html validator doe snot like that page
tmortagne - (13:45): (i'm talking about the velocity macro in ConfigurableClass page)
CalebJamesDeLisl - (13:47): Do you know where the offending </p> is? I think maybe the best thing to do is get rid of all {{html}}</p> and <p>{{/html}} since the new renderer doesn't need them.
tmortagne - (13:48): CalebJamesDeLisl: it's somewhere in ConfigurableClass code but i did not tried anything
jvdrean joined #xwiki at 13:49
CalebJamesDeLisl - (14:02): I removed the <p> and </p> tags and now registration section is valid.
CalebJamesDeLisl - (14:12): XAADMINISTRATION-138
arkub joined #xwiki at 14:16
tmortagne - (14:24): CalebJamesDeLisl: ok thanks, i just finished to add support for explicit URL in XE validation tool, i will add some admin related pages
kibahop left at 14:25 (Read error: Operation timed out
plunden1 left at 14:29 (Ping timeout: 240 seconds
bbc581 left at 14:42 (Remote host closed the connection
lucaa left #xwiki at 14:46
plunden1 joined #xwiki at 14:53
kibahop joined #xwiki at 14:54
bbc581 joined #xwiki at 14:55
annacondan left at 14:55 (Quit: annacondan
plunden1 left #xwiki at 15:09
evalica left at 15:22 (Quit: Leaving.
annacondan joined #xwiki at 15:23
plunden left at 15:49 (Ping timeout: 252 seconds
kibahop left #xwiki at 15:52
plunden joined #xwiki at 15:55
jvdrean left at 16:08 (Quit: Leaving.
tmortagne - (16:14): CalebJamesDeLisl: see http://hudson.xwiki.org/job/xwiki-product-enterprise-tests/com.xpn.xwiki.products$xwiki-enterprise-test-webstandards/764/testReport/com.xpn.xwiki.it/XHTMLValidationTest/Validating_XHTML_validity_for___xwiki_bin_admin_XWiki_XWikiPreferences_editor_globaladmin_section_Registration/
CalebJamesDeLisl - (16:16): Yes fixing now. That's great being able to test arbitrary urls!
tmortagne - (16:16): CalebJamesDeLisl: yep, i listed several but feel free to add more if you think of some important one
CalebJamesDeLisl - (16:17): I'll add register/XWiki/Register if it's not there.
tmortagne - (16:17): not there indeed
tmortagne - (16:18): except for admin i just added one of each editor wysiwyg/wiki/object/class
CalebJamesDeLisl - (16:23): Hmm, must be lots of violations...
tmortagne - (16:24): not that much for XHTML validation actually
CalebJamesDeLisl - (16:24): Should I just create a big issue for all Administration app violations?
tmortagne - (16:24): but a lot for Dutch Web Guidelines Validity
tmortagne - (16:24): CalebJamesDeLisl: yep shoud be ok
tmortagne - (16:25): if all is fixed before 2.3.1/2.4M1
jvdrean joined #xwiki at 16:28
CalebJamesDeLisl - (16:31): XAADMINISTRATION-139 XAADMINISTRATION-140
CalebJamesDeLisl - (16:32): XAADMINISTRATION-140
silviar left at 16:42 (Quit: Leaving.
Enygma` left at 16:43 (Quit: Leaving.
CalebJamesDeLisl - (17:00): The wcag tests run logged in as admin right?
oanat left at 17:01 (Quit: Leaving.
tmortagne - (17:02): CalebJamesDeLisl: yes
tmortagne - (17:02): same for xhtml tests (both are using the same code)
tmortagne - (17:02): (unless i broke something)
CalebJamesDeLisl - (17:02): Hmm, how to test login page? Maybe http://127.0.0.1 because cookies are different?
tmortagne - (17:03): CalebJamesDeLisl: there is no way to not be admin currently
tmortagne - (17:03): that would need some refactoring
CalebJamesDeLisl - (17:04): There is a selenium test where I load the url http://127.0.0.1:8080/xwiki/.... to drop admin status for one page load.
CalebJamesDeLisl - (17:04): (to check if a document is locked.)
tmortagne - (17:04): i mean there is no way in the validation framework
tmortagne - (17:04): you can do what you want in a custom test
CalebJamesDeLisl - (17:05): Do you know if it'd take a full url ?
tmortagne - (17:06): what ? urlsToTest ?
CalebJamesDeLisl - (17:07): Yup.
tmortagne - (17:07): CalebJamesDeLisl: for urlsToTest property: yes it supports absolute url
CalebJamesDeLisl - (17:07): You see any problem with using such a hack?
anamarias left at 17:08 (Quit: anamarias
tmortagne - (17:08): CalebJamesDeLisl: i don't understand this hack actually :)
CalebJamesDeLisl - (17:08): Login at localhost:8080/xwiki/.... then go to 127.0.0.1:8080/.... and your cookies are not sent.
CalebJamesDeLisl - (17:09): Because of cookie domain.
CalebJamesDeLisl - (17:09): Another way would be to switch to localhost.localdomain:8080/
tmortagne - (17:10): CalebJamesDeLisl: test framework is using basic authentication to make sure the user is admin
tmortagne - (17:10): whatever the URL you will be admin
tmortagne - (17:11): best i can propose now id to have a urlsToTestAsGuest property
tmortagne - (17:11): with a list of pages to test as guest instead of admin
tmortagne - (17:12): should not be difficult to do
CalebJamesDeLisl - (17:12): Yea, that looks like the only way.
tmortagne - (17:13): or probably rename urlsToTest as urlsToTestAsAdmin to be more clear
tmortagne - (17:14): doing this now
Denis left at 17:15 (Quit: Leaving.
tmortagne - (17:20): CalebJamesDeLisl: committed, urlsToTestAsAdmin and urlsToTestAsGuest
CalebJamesDeLisl - (17:21): Hey you added the guest part, cool. I was just looking at the code trying to make sense of it.
sdumitriu left at 17:23 (Ping timeout: 276 seconds
glerouge left at 17:57 (Quit: Leaving.
flaviusolaru left at 18:08 (Read error: Connection reset by peer
tmortagne left at 18:08 (Quit: Leaving.
annacondan left at 18:25 (Quit: annacondan
arkub left at 18:33 (Quit: Leaving
nickless left at 19:05 (Ping timeout: 264 seconds
jvdrean left at 19:26 (Quit: Leaving.
gvallarelli joined #xwiki at 19:54
gvallarelli - (19:54): good evening
CalebJamesDeLisl - (19:56): Good afternoon ;)
gvallarelli - (19:58): CalebJamesDeLisl: but you're american it's unfair :P
CalebJamesDeLisl - (19:58): Unfair? Because you get the early hours and we get the late ones?
CalebJamesDeLisl - (20:03): I can see how it would be tough having all morning and no afternoon ;)
gvallarelli - (20:06): I guess you should feel less tired
gvallarelli - (20:06): with a continous morning
CalebJamesDeLisl - (20:09): What have you been up to lately?
gvallarelli - (20:11): CalebJamesDeLisl: well nothing special except read read and read reaseach articles about agents
gvallarelli - (20:11): *research
CalebJamesDeLisl - (20:13): Heading in the AI direction?
gvallarelli - (20:14): I think that AI stuff are the more interesting topics on my CS department
CalebJamesDeLisl - (20:15): I was thinking if you had a program with a lot of settings (eg: bittorrent) you could have an AI program change the settings a little at a time and try for the best throughput.
CalebJamesDeLisl - (20:17): Would only work with a program where success is easily measured (like network throughput) and would only work when settings all were similar (change one unit of setting x has similar magnitude of effect as one unit of setting y)
gvallarelli - (20:19): CalebJamesDeLisl: actually I'm trying to figure out the mearning of my thesis topic
gvallarelli - (20:19): *meaning
CalebJamesDeLisl - (20:20): The topic was given to you? Is there a short description?
gvallarelli - (20:20): unfortunately no :D
gvallarelli - (20:21): anyway about multiagent coalition formation
gvallarelli - (20:21): I will implement an algorithm to create a coalition given a set of agents in C
gvallarelli - (20:21): It's a question of time and things will be more clear
gvallarelli - (20:21): I'm at the beginning
CalebJamesDeLisl - (20:24): C seems a bit obsolete, at least it's not C++
gvallarelli - (20:25): CalebJamesDeLisl: yep, C is more simple and straightforward than C++
gvallarelli - (20:25): also books on C++ are about 1000 pages :D it's a shame
gvallarelli - (20:26): not for the books but for the language so complex
CalebJamesDeLisl - (20:26): I remember hearing the language spec had grown to some ridiculous number of pages.
CalebJamesDeLisl - (20:27): My understanding of C++ is it was a failed attempt to apply object orientation to C.
CalebJamesDeLisl - (20:29): I don't really like C either because bad code can just construct a pointer and call any function it wants.
CalebJamesDeLisl - (20:30): And the runtime errors: " Crash at 0xABCDEF Message: I hate you. "
gvallarelli - (20:32): today I was studying about pointers and pointer arithmetic and fascinates me
gvallarelli - (20:32): I understand that could be a source of many errors
gvallarelli - (20:33): but one thing that I like is that there are so little and tiny abstractions, for the first time no one made the decisions for me on which things to abstract.
CalebJamesDeLisl - (20:34): Also you can easily port your code to run on nvidia GPUs which are fast. Like 500 cores.
gvallarelli - (20:35): eheh cool!
CalebJamesDeLisl - (20:38): If you're bored, there's a great video about making software secure by controlling what it can get references to http://www.youtube.com/watch?v=eL5o4PFuxTY
CalebJamesDeLisl - (20:39): If you can only have what you are given and you can load classes, you are pretty well boxed in.
CalebJamesDeLisl - (20:39): * and you can't load classes
CalebJamesDeLisl - (20:40): Constructing arbitrary pointers breaks that concept.
nickless joined #xwiki at 20:40
gvallarelli - (20:42): CalebJamesDeLisl: yep
gvallarelli - (20:43): I love subtitles
CalebJamesDeLisl - (20:43): A paper I read a while back was about cryptanalysis on the nvidia gpu. http://eecm.cr.yp.to/pc109-20090901.pdf
gvallarelli - (20:43): CalebJamesDeLisl: you study and read a lot of things, I hope you study CS :)
CalebJamesDeLisl - (20:44): I guess I do.
CalebJamesDeLisl - (20:45): Anyway what I found interesting about the paper was that you should have more threads than you have cores so when one blocks, the other can pick up.
gvallarelli - (20:45): CalebJamesDeLisl: cool video thank you!
CalebJamesDeLisl - (20:46): It's really neat stuff, there's a whole community around the concept of "object capability security"
gvallarelli - (20:50): CalebJamesDeLisl: interesting I'm watching it right now
anamarias joined #xwiki at 20:55
sdumitriu joined #xwiki at 20:56
gvallarelli - (21:35): CalebJamesDeLisl: I was thinking that using immutable data structures improve a lot security because it's impossible to corrupt data, so functional languages are more secure than side effect slave programming languages
CalebJamesDeLisl - (21:37): Looking up side effects, still a bit ambiguous, can you give me your definition?
gvallarelli - (21:40): side effects happens when you accomplish a task with a function for example and alter object state
gvallarelli - (21:40): substitute function with method
CalebJamesDeLisl - (21:43): Yes. I don't think it's a problem when a method alters the state of it's object, but it shouldn't be mutating parameters.
CalebJamesDeLisl - (21:43): Also IMO java should never have gotten the static keyword, there is no good use for it.
gvallarelli - (21:46): if something strange happens and for example a method don't complete his execution you could have a corrupted object with a very bad state
gvallarelli - (21:46): wait somewhere I bookmarked an interesting article about it
lucaa joined #xwiki at 21:46
gvallarelli - (21:48): CalebJamesDeLisl: is a little related to problems about altering state http://damienkatz.net/2006/04/error_code_vs_e.html
CalebJamesDeLisl - (21:48): Then where does the data go?
gvallarelli - (21:49): look at the article it's better explained than I could do :S
CalebJamesDeLisl - (21:50): Haha firefox. Javascript is safer, but I can still get it to segfault :P
gvallarelli - (21:51): :P
CalebJamesDeLisl - (22:02): "In every popular language, the updating of a variable means the previous state of that variable is destroyed, vanished, gone and you can't get it back"
CalebJamesDeLisl - (22:03): Not necessarily.
CalebJamesDeLisl - (22:08): "(photo of actual atom)" LOL
CalebJamesDeLisl - (22:11): Interesting article. He didn't mention copy-on-write!
CalebJamesDeLisl - (22:13): Our code is littered with document.clone() invocations so that document saving can't break the document in the cache.
CalebJamesDeLisl - (22:14): Prototype (javascript) Object.extend() returns the same object except changes to the new object are separate from the old one. AFAIK
CalebJamesDeLisl - (22:15): So it's like clone but it doesn't require so much processor horsepower and RAM.
gvallarelli - (22:24): interesting :)
gvallarelli - (22:25): got to go
gvallarelli - (22:25): good evening/night
CalebJamesDeLisl - (22:25): See ya. One last thought, security and error handling are the same.
gvallarelli - (22:25): goodbye CalebJamesDeLisl
gvallarelli - (22:26): thank you for the conversation
CalebJamesDeLisl - (22:26): Thank you.
gvallarelli left #xwiki at 22:26
CalebJamesDeLisl - (23:32): 'getAllMatchedUsers' in class com.xpn.xwiki.plugin.rightsmanager.RightsManagerUsersApi threw exception com.xpn.xwiki.XWikiException: Error number 3223 in 3: Exception while searching documents with SQL [, BaseObject as obj where doc.fullName=obj.name and doc.fullName<>? and obj.className=? order by doc.name asc]
CalebJamesDeLisl - (23:32): JDBCExceptionReporter - Operation requires a scrollable ResultSet, but this ResultSet is FORWARD_ONLY.
CalebJamesDeLisl - (23:32): getusers.vm
sdumitriu left at 23:33 (Ping timeout: 246 seconds
