MoinMoin: MoinMoinChat/Logs/moin-dev/2006-01-31

2006-01-31T00:05:16  <xorAxAx> [PERIODIC ANNOUNCEMENT] Logs can be found on http://moinmoin.wikiwikiweb.de/MoinMoinChat/Logs/moin-dev
2006-01-31T19:00:51  <xorAxAx> 19:00:39 < xorAxAx> if we put all wikimacro/wikiaction code into files, we would allow people to override them
2006-01-31T19:00:56  <xorAxAx> any opinion on that?
2006-01-31T19:27:49  <ThomasWaldmann> xorAxAx: wikiutil.py is a file
2006-01-31T19:28:26  <xorAxAx> ThomasWaldmann: ?
2006-01-31T19:28:44  <xorAxAx> ThomasWaldmann: i mean separating the inlined plugins into separate files. wikiutil is not related at all here
2006-01-31T19:28:44  <ThomasWaldmann> do you mean plugins?
2006-01-31T19:29:21  <ThomasWaldmann> yes, we could do that. cgi performance would get lower, though.
2006-01-31T19:29:49  <ThomasWaldmann> so the first question is the one for the practical application :)
2006-01-31T19:30:25  <xorAxAx> it wont get much lower i guess because moin doesnt load that many plugins ...
2006-01-31T19:39:30  <ThomasWaldmann> now: import wikimacro, then: import br, import fullsearch, import ...
2006-01-31T19:39:55  <ThomasWaldmann> per file: 10ms+
2006-01-31T19:40:42  <xorAxAx> i think we should profile the mostly used ones and separate the rather infrequently used ones
2006-01-31T19:41:15  <ThomasWaldmann> and get some use cases first :)
2006-01-31T19:41:38  <ThomasWaldmann> we have lots of other stuff to invest time...
2006-01-31T19:41:38  <xorAxAx> we had one in #moin a few minutes ago
2006-01-31T19:50:28  <dreimark> Hi Thomas,
2006-01-31T19:50:40  <dreimark> MoinMoinBugs/WikiactionGuiActionCancelToLate
2006-01-31T19:52:47  <dreimark> I think if one cancels it does not need to call
2006-01-31T19:52:48  <dreimark> savetext = convert(request, pagename, savetext)
2006-01-31T19:53:30  <dreimark> and then trash it
2006-01-31T19:55:19  <Fabi> dreimark, we have to
2006-01-31T19:55:34  <xorAxAx> Fabi: why?
2006-01-31T19:55:35  <Fabi> it is saved in User/EditorBackup
2006-01-31T19:55:39  <xorAxAx> ah
2006-01-31T19:55:55  <xorAxAx> then put a try, except around it and reraise the exception later
2006-01-31T19:55:56  <Fabi> I changed this to get the HTML out of the backup
2006-01-31T19:57:10  <Fabi> are there any opinions about shipping a gallery parser+action
2006-01-31T19:57:57  <xorAxAx> Fabi: if it works ...
2006-01-31T19:58:04  <Fabi> not yet
2006-01-31T19:58:12  <Fabi> working on it
2006-01-31T19:58:15  <xorAxAx> the market ones are very dirty, at least the last time when i looked at it
2006-01-31T19:58:19  <Fabi> based on Gallery2
2006-01-31T19:58:28  <xorAxAx> i dont know gallery2, though
2006-01-31T19:58:33  <Fabi> but most parts get rewritten right now
2006-01-31T19:58:46  <Fabi> it is/was dirty
2006-01-31T19:59:14  <xorAxAx> :-)
2006-01-31T19:59:29  <Fabi> still relys on java script
2006-01-31T19:59:52  <Fabi> I consider a fall back
2006-01-31T20:01:43  <xorAxAx> good idea
2006-01-31T20:02:27  <dreimark> Fabi:what actions did you plan for the new Galelry parser
2006-01-31T20:03:30  <Fabi> see http://moinmoin.wikiwikiweb.de/ParserMarket/Gallery2
2006-01-31T20:03:36  <Fabi> search for slideshow
2006-01-31T20:04:10  <Fabi> I use other buttons though
2006-01-31T20:05:09  <dreimark> ok
2006-01-31T20:05:31  <dreimark> do you use description and exif data as well
2006-01-31T20:06:31  <Fabi> yup
2006-01-31T20:06:47  <Fabi> but it will be configurable via parser param
2006-01-31T20:07:43  <dreimark> what is with the other actions: rotate, full screen, remove
2006-01-31T20:08:17  <Fabi> I already dropped the sequence feature
2006-01-31T20:08:47  <Fabi> rotate - if left in - will depend on the Python Image Library
2006-01-31T20:09:33  <Fabi> als long as we don't have a rename action for attachments remove should get deactivated
2006-01-31T20:10:26  <dreimark> ok
2006-01-31T20:10:53  <Fabi> but rotate is very buggy right now...
2006-01-31T20:11:28  <dreimark> which timetable do you have planned for this
2006-01-31T20:11:53  <xorAxAx> im sure he will release it earlier than a lupy fix :)
2006-01-31T20:12:03  <Fabi> hope to publish a first snapshot next WE
2006-01-31T20:12:09  <dreimark> lupy is dead or isn't it
2006-01-31T20:12:33  <dreimark> what is the problem with rotate
2006-01-31T20:12:58  <xorAxAx> dreimark: moin is dead if you enable lupy as well :)
2006-01-31T20:13:02  <Fabi> it doesn't get the image, webnail, thumbnail right
2006-01-31T20:13:11  <Fabi> it rotates only some
2006-01-31T20:13:18  <xorAxAx> integrating pylucene patch and fixing the remaining bugs would push this further ...
2006-01-31T20:13:44  <Fabi> problem is that there are some major bug in lupy
2006-01-31T20:14:31  <dreimark> remove thumbnail, webnail, rotate the image, recreate thumbnail, webnai
2006-01-31T20:14:39  <dreimark> l
2006-01-31T20:16:22  <dreimark> normally it does it always right the browser is the problem
2006-01-31T20:16:48  <dreimark> konqueror remembers the old image it does not depend if cache is off or on
2006-01-31T20:17:11  <dreimark> firefox shows it always right
2006-01-31T20:17:53  <Fabi> IIRC it sometimes beleaves that the webnail is the real image
2006-01-31T20:17:56  <xorAxAx> Fabi: i said pylucene
2006-01-31T20:18:07  <xorAxAx> Fabi: i dont see any problems using pylucene
2006-01-31T20:18:23  <Fabi> me neither
2006-01-31T20:18:24  <xorAxAx> if somebody seriously needs it, then he can install a binary as well
2006-01-31T20:18:51  <Fabi> yes, but we can't use it as infrastructure
2006-01-31T20:19:31  <xorAxAx> i didnt agree with that anyway
2006-01-31T20:19:58  <ThomasWaldmann> re
2006-01-31T20:21:23  <dreimark> Fabi:does this happen with Gallery2 too?
2006-01-31T20:24:47  <ThomasWaldmann> btw, i intend to make another bugfix release with the formatter / win32 fixes soon
2006-01-31T20:25:13  <ThomasWaldmann> so if someone wants to look at the code before i merge it to main, do it soon :)
2006-01-31T20:25:43  <xorAxAx> i am just interested in it when it got into main :)
2006-01-31T20:27:40  <dreimark> is then the user.py.patch added too
2006-01-31T20:27:59  <dreimark> from MoinMoinTodo/Release_1.5.1
2006-01-31T20:29:37  <ThomasWaldmann> dreimark: the branch is only parser/formatter stuff
2006-01-31T20:29:56  <ThomasWaldmann> but I will look at it
2006-01-31T20:30:09  <dreimark> that's right I thought for bugfix release
2006-01-31T20:32:26  <dreimark> Fabi:ParserMarket/Gallery2 page is at PageHits list at position 113
2006-01-31T20:32:45  <dreimark> so it seems interesting many users
2006-01-31T20:33:13  <ThomasWaldmann> xorAxAx: why do you reopen old bugs?
2006-01-31T20:33:43  <ThomasWaldmann> the one where you wrote "duplicate" was fixed by user by a modpy patch
2006-01-31T20:34:14  <xorAxAx> ThomasWaldmann: because the old one concluded that its a bug that should be worked around in moin ;-)
2006-01-31T20:36:57  <dreimark> xorAxAx:same like the copy.py bug CategoryMoinMoinWorkaround or something else?
2006-01-31T20:38:17  <xorAxAx> dreimark: that was a bug in moin that was worked around by avoiding edge cases in the std lib :)
2006-01-31T20:38:22  <xorAxAx> in this case, its a real modpy bug
2006-01-31T20:38:35  <ThomasWaldmann> xorAxAx: so how do you want to work around?
2006-01-31T20:38:47  <ThomasWaldmann> the patches are for modpy, not for moin
2006-01-31T20:38:57  <xorAxAx> ThomasWaldmann: lets ask the modpy guys :)
2006-01-31T20:39:21  <xorAxAx> i think thats the way to go
2006-01-31T20:39:41  <xorAxAx> they should explain us why they have been shipping b0rked code for ages
2006-01-31T20:40:03  <ThomasWaldmann> maybe you should add "no moin bug" to status
2006-01-31T20:45:17  * ThomasWaldmann tries to fix that macro p bug now
2006-01-31T20:48:04  <ThomasWaldmann>         #self.inhibit_p = 1 # fixes UserPreferences, but makes new trouble!
2006-01-31T20:50:19  <Fabi> hehe
2006-01-31T20:52:30  <ThomasWaldmann> hmm, reverted that change, UP still looks ok
2006-01-31T20:54:44  <ThomasWaldmann> feel free to test in http://refactor.wikiwikiweb.de/ and find stuff that breaks now. Clear the cache on any page you try.
2006-01-31T20:55:57  <dreimark> is it possible to clear the cache automaticly before a page is opened in gui?
2006-01-31T20:58:31  <dreimark> GUI_IndentMoving2 isn't fixed or ?
2006-01-31T20:58:52  <ThomasWaldmann> it should clear when you save anyway
2006-01-31T20:59:28  <dreimark> but does it get data formatted from a wrong cache?
2006-01-31T21:00:20  <ThomasWaldmann> it gets data from text_gedit formatter and that is not cached
2006-01-31T21:00:37  <dreimark> ok
2006-01-31T21:06:22  <dreimark> soemtimes I wish to have HelpContents herarchical
2006-01-31T21:06:52  <dreimark> HelpContents/HelpForBeginners and so on
2006-01-31T21:13:13  <dreimark>  http://refactor.wikiwikiweb.de/HelpOnTables
2006-01-31T21:13:43  <dreimark>  some missing lines in firefox
2006-01-31T21:14:40  <dreimark> Alignment example does not show all horizontal lines
2006-01-31T21:15:09  <xorAxAx> you are red, dreimark :)
2006-01-31T21:15:41  <dreimark>  if it works ;-)
2006-01-31T21:16:54  <xorAxAx> thats red again^Whey, i am have problems with red-green-differentiation according to the doctor of the KWEA :)
2006-01-31T21:17:01  <xorAxAx> s/i am/i/
2006-01-31T21:17:10  <xorAxAx> but i havent noticed them anywhere besides there
2006-01-31T21:18:55  <dreimark> konqueoror does show them
2006-01-31T21:21:08  <ThomasWaldmann> dreimark: what exactly is missing?
2006-01-31T21:22:15  <dreimark> NEW STYLE: General table layout and HTML like options
2006-01-31T21:22:24  <dreimark> below the horizontal line was missing
2006-01-31T21:22:42  <dreimark> I have resized the browser and reloaded
2006-01-31T21:22:57  <dreimark> now it's there
2006-01-31T21:23:20  <dreimark> at Alignment
2006-01-31T21:23:49  <dreimark> it was not missinf the horizontal line but the vertical between top and bottom
2006-01-31T21:24:55  <dreimark> so false alarm  I can't reproduce it now
2006-01-31T21:28:10  <ThomasWaldmann> maybe a rendering problem in browser (not wrong html)?
2006-01-31T21:29:31  <dreimark> while it goes well in konqueror and by now in firefox too I think it was a rendering problem in firefox
2006-01-31T21:29:42  <dreimark> if it happens again I save the source
2006-01-31T21:32:59  <ThomasWaldmann> but vertical alignment is broken
2006-01-31T21:36:03  <dreimark> I see top on top
2006-01-31T21:36:10  <dreimark> in the table
2006-01-31T21:36:56  <dreimark> same line as left in first column
2006-01-31T21:38:27  <dreimark> in SyntaxReference
2006-01-31T21:39:13  <dreimark> and in HelpOnTables it looks broken but probably syntax error?
2006-01-31T21:42:17  <dreimark> it's broken syntax is right
2006-01-31T21:45:48  <ThomasWaldmann> i just fixed the colouring error on moinmaster
2006-01-31T21:45:59  <ThomasWaldmann> maybe fix your error there, too
2006-01-31T21:48:03  <dreimark> NEW STYLE: alignment top is wrong
2006-01-31T21:49:38  <dreimark> It is a syntax error, mixed code old style new style
2006-01-31T21:50:06  <dreimark> ma be?
2006-01-31T21:50:08  <dreimark> may
2006-01-31T21:53:00  <dreimark> top and bottom don't work in ExampleAlignment
2006-01-31T21:59:32  <dreimark> see right syntax now at http://refactor.wikiwikiweb.de/ExampleAlignment
2006-01-31T22:02:17  <ThomasWaldmann> if help page is wrong, please fix help page
2006-01-31T22:11:36  <dreimark> gui has in table mode a lot of problems
2006-01-31T22:14:33  <dreimark>  * in text mode you could use indent level for the table you can't in gui mode
2006-01-31T22:17:57  <dreimark> http://refactor.wikiwikiweb.de/GUI_Table in gui does not work
2006-01-31T22:18:13  <dreimark> it's an other indent . problem
2006-01-31T22:19:20  <dreimark> fabi: gui changes the page if nothing is changed
2006-01-31T22:23:13  <dreimark> GUI_Table2 shows what happens if you choose TableProperties
2006-01-31T22:24:40  <ThomasWaldmann> gui editor table stuff will get fixed later
2006-01-31T22:25:40  <dreimark> ok then I stop testing this now
2006-01-31T22:26:17  <ThomasWaldmann> the last todo i found formatter/parser related is that entities stuff
2006-01-31T22:26:40  <ThomasWaldmann> if someone find more that is easily fixed, tell me :)
2006-01-31T22:34:25  <ThomasWaldmann> xorAxAx: do you see a problem with adding a amp=1 param to wikiutil.escape and call it from text_html._text with amp=0 so it doesnt replace the & in &entity; ?
2006-01-31T22:35:15  <xorAxAx> ThomasWaldmann: &s in text() args need to be replaced of course
2006-01-31T22:35:25  <xorAxAx> if that is ensured ...
2006-01-31T22:36:12  <ThomasWaldmann> _text is called by text...
2006-01-31T22:36:36  <ThomasWaldmann> why does it need to be replaced?
2006-01-31T22:37:15  <ThomasWaldmann> Hinz & Kunz ?
2006-01-31T22:38:14  <xorAxAx> needs to become &amp;
2006-01-31T22:38:29  <xorAxAx> otherwise its invalid html, you can do XSS etc.
2006-01-31T22:38:34  <ThomasWaldmann> yes
2006-01-31T22:38:41  <ThomasWaldmann> hmm, not that easy
2006-01-31T22:39:17  <xorAxAx> pretty easy ... where is the problem?
2006-01-31T22:39:51  <ThomasWaldmann> both cases can occur mixed
2006-01-31T22:39:58  <xorAxAx> ?
2006-01-31T22:40:13  <xorAxAx> the wiki parser did it correctly
2006-01-31T22:40:16  <ThomasWaldmann> Hinz & M&uuml;ller
2006-01-31T22:40:18  <xorAxAx> did you break that? %-)
2006-01-31T22:40:37  <ThomasWaldmann> it never worked for = a & b =
2006-01-31T22:40:42  <ThomasWaldmann> ehrm
2006-01-31T22:40:42  <xorAxAx> of course. because of that, you can just use numerical entities
2006-01-31T22:40:51  <ThomasWaldmann> it never worked for = M&uuml;ller =
2006-01-31T22:40:57  <xorAxAx> ?
2006-01-31T22:41:13  <ThomasWaldmann> http://moinmoin.wikiwikiweb.de/MoinMoinBugs/EntitiesInHeadingAndUrls
2006-01-31T22:41:38  <xorAxAx> &uuml; is not a numeric entity
2006-01-31T22:42:14  <ThomasWaldmann> i added symbolic entities in refactor
2006-01-31T22:42:26  <ThomasWaldmann> treated the same way as numeric
2006-01-31T22:42:47  <xorAxAx> as you can see, thats b0rked
2006-01-31T22:43:06  <xorAxAx> please think beforehand *sigh*
2006-01-31T22:43:38  <ThomasWaldmann> did i ever mention you need to be more concrete?
2006-01-31T22:44:09  <xorAxAx> umm, yeah
2006-01-31T22:44:22  <xorAxAx> you cannot decide if a & starts a numeric entity without knowing all
2006-01-31T22:44:27  <xorAxAx> s/numeric//
2006-01-31T22:44:47  <xorAxAx> which makes the parser really complex
2006-01-31T22:45:04  <ThomasWaldmann> it just uses some regex
2006-01-31T22:45:19  <xorAxAx> as we have already all entities in the code ...
2006-01-31T22:45:21  <ThomasWaldmann> &[a-z][A-Z]+;
2006-01-31T22:45:51  <xorAxAx> hmm, that doesnt match &Uuml; does it?
2006-01-31T22:46:01  <ThomasWaldmann> wait
2006-01-31T22:46:46  <ThomasWaldmann> (?P<ent_symbolic>&[a-zA-Z]+;)
2006-01-31T22:47:08  <ThomasWaldmann> of course that can make a wrong match &wtf;
2006-01-31T22:47:24  <xorAxAx> it can yield false matches, yeah
2006-01-31T22:47:30  <xorAxAx> leading to invalid html code
2006-01-31T22:47:51  <xorAxAx> if we want to go for xhtml output, that might be a problem
2006-01-31T22:48:05  <ThomasWaldmann> not as long as you have correct content :)
2006-01-31T22:48:05  <xorAxAx> if we wont, then we do not necessarily need to care :)
2006-01-31T22:49:08  <ThomasWaldmann> ordered lists maybe will never be strict, btw. not even strict html 4.01
2006-01-31T22:49:21  <xorAxAx> why?
2006-01-31T22:49:37  <ThomasWaldmann> because those idiots remove value= and start= attrs
2006-01-31T22:50:46  <ThomasWaldmann> list type can be done in css, start= can be done in css 2.1, value = ???
2006-01-31T22:51:07  <ThomasWaldmann> (and who implements css 2.1?)
2006-01-31T22:53:09  <ThomasWaldmann> so we cant do 1.#4 in strict *html
2006-01-31T22:55:06  <xorAxAx> we could use css
2006-01-31T22:55:10  <ThomasWaldmann> ok, I'll merge into main now
2006-01-31T22:55:33  <ThomasWaldmann> xorAxAx: if you know how, make a FeatureRequest for XhtmlStrictOl
2006-01-31T22:56:13  <xorAxAx> if i get to know the solution, i will do
2006-01-31T22:59:21  <ThomasWaldmann> i googled for it and the usual advice is to continue using value and start, ignore w3c validator and claim strict nevertheless
2006-01-31T23:00:56  <xorAxAx> :-)
2006-01-31T23:01:16  <dreimark>  a.#4 does work too :-)
2006-01-31T23:06:38  <ThomasWaldmann> dreimark: it should
2006-01-31T23:11:22  <ThomasWaldmann> duck!
2006-01-31T23:12:19  <ThomasWaldmann> ok, all wikis running on merged code now
2006-01-31T23:12:31  <ThomasWaldmann> (merged refactor branch)
2006-01-31T23:12:48  <dreimark> ok
2006-01-31T23:15:20  <dreimark> are you sure the cache is cleared?
2006-01-31T23:16:13  <dreimark> it would make it much easier to see if something does not go right
2006-01-31T23:18:32  <ThomasWaldmann> i touched the MoinMoin dir, that should invalidate the cache
2006-01-31T23:24:18  <ThomasWaldmann> we call url() with unescaped=x, but there is no code to check unescaped there...
2006-01-31T23:31:08  <dreimark> xorAxAx:where are the MoinChat/Logs/moin-dev if the last 7 days
2006-01-31T23:31:35  <xorAxAx> dreimark: checking
2006-01-31T23:32:25  <xorAxAx> hah, this time, its thomas fault :)
2006-01-31T23:32:43  <xorAxAx> ThomasWaldmann: please check the cronjob
2006-01-31T23:37:14  <dreimark> gn
2006-01-31T23:38:56  <ThomasWaldmann> xorAxAx: cmdline?
2006-01-31T23:39:22  <xorAxAx> ThomasWaldmann: that runs the irclogimporter
2006-01-31T23:55:35  <ThomasWaldmann> it filled up /tmp
2006-01-31T23:55:46  <xorAxAx> it doesnt use tmp at all, does it?
2006-01-31T23:57:25  <ThomasWaldmann> i used if for a logfile