MoinMoin: MoinMoinChat/Logs/moin-dev/2006-01-22

2006-01-22T00:05:16  <xorAxAx> [PERIODIC ANNOUNCEMENT] Logs can be found on http://moinmoin.wikiwikiweb.de/MoinMoinChat/Logs/moin-dev
2006-01-22T00:44:03  <dreimark> 1.5.1 looks fine
2006-01-22T00:45:48  <ThomasWaldmann> you installed it from the archive?
2006-01-22T00:46:23  <dreimark> first and second from the new tgz
2006-01-22T00:46:59  <dreimark> I do used my old instance
2006-01-22T00:47:39  <ThomasWaldmann> first and second?
2006-01-22T00:48:37  <dreimark> first I used it from tla update
2006-01-22T00:49:13  <ThomasWaldmann> it isnt everything in archive yet
2006-01-22T00:49:37  <dreimark> then I removed the installed site-packages/MoinMoin
2006-01-22T00:49:50  <dreimark> and installed the one fromn the tgz.
2006-01-22T00:50:26  <dreimark> it seems to work with all of my testcases I have in my local wiki
2006-01-22T00:54:01  <ThomasWaldmann> if no new problems are reported with it, i'll upload it to sf later today
2006-01-22T12:54:36  * xorAxAx distributed all release announcements
2006-01-22T12:56:49  <ThomasWaldmann> btw, sf.net has sent 106 notify emails out
2006-01-22T12:57:25  <ThomasWaldmann> maybe we should just point to that possibility
2006-01-22T12:57:37  <xorAxAx> yeah, i am subscribed there as well
2006-01-22T12:58:01  <xorAxAx> and i still think that my mail announcements are read
2006-01-22T12:58:52  <ThomasWaldmann> sure, I read them, too X)
2006-01-22T13:00:53  <xorAxAx> why? :)
2006-01-22T13:01:56  <ThomasWaldmann> i read all mailing list mails. some only for 100ms, though.
2006-01-22T13:02:39  <ThomasWaldmann> (so that they get state "read" :)
2006-01-22T13:03:09  <xorAxAx> yeah, at least on our list i do it like that as well
2006-01-22T13:03:19  <xorAxAx> but for most other lists, i prefer ignore
2006-01-22T13:11:39  <starshine> congrats on release
2006-01-22T13:12:24  <starshine> I do know people who favor their announcements "from the real source" so I think xorAxAx is right to maintain it.
2006-01-22T17:17:06  <ThomasWaldmann> moin Fabi
2006-01-22T17:17:13  <Fabi> moin
2006-01-22T17:21:00  <ThomasWaldmann> did you see/follow the recent formatter discussions on the wiki?
2006-01-22T17:21:56  <Fabi> not yet
2006-01-22T17:24:21  <Fabi> anything else than DeronMeranda/DiscussPythonFormatter ?
2006-01-22T17:24:26  <ThomasWaldmann> search for DeronMeranda, he already put quite some work into it
2006-01-22T17:25:23  <ThomasWaldmann> i delayed this stuff to after 1.5.1 (== now :) because it does quite some changes
2006-01-22T17:25:55  <xorAxAx> ThomasWaldmann: no incompatible ones as far as i could see
2006-01-22T17:27:58  <Fabi> this *kw stuff looks sensible
2006-01-22T17:28:09  <Fabi> and shouldn't break anything
2006-01-22T17:28:29  <ThomasWaldmann> he seems to have problems with the assymetric kind of the python formatter
2006-01-22T17:29:16  <ThomasWaldmann> so the question is whether we want to give that up and make the tag-end stuff dynamic, too, if the begin stuff is
2006-01-22T17:30:05  <ThomasWaldmann> or just make some dummy begin call
2006-01-22T17:30:21  <ThomasWaldmann> when producing begin source code
2006-01-22T17:31:33  <xorAxAx> ThomasWaldmann: not sensible because text_python is written to work with text_html and doesnt support any non-html formatters as  i wrote already
2006-01-22T17:31:34  <ThomasWaldmann> (might work as long as the end stuff is really static)
2006-01-22T17:31:54  <Fabi> mompl
2006-01-22T17:33:44  <Fabi> xorAxAx: why is text_python for html only?
2006-01-22T17:33:59  <xorAxAx> Fabi: you wrote it, you should know that :-)
2006-01-22T17:34:14  <xorAxAx> Fabi: <<<<>>>> should refresh your mind :)
2006-01-22T17:34:43  <xorAxAx> i mean, it obviously collides with data in other formats as it is used in-place
2006-01-22T17:36:20  <Fabi> you mean <<<<>>>> is not unique enough?
2006-01-22T17:36:40  <Fabi> if this is the only problem...
2006-01-22T17:36:49  <xorAxAx> yeah, and it just works for html because you cannot yield it by calling the formatter methods
2006-01-22T17:37:16  <xorAxAx> then the next problem is that the caching code paths do not support other formats
2006-01-22T17:37:17  <Fabi> should be easy to fix and not a "real" problem
2006-01-22T17:37:34  <xorAxAx> (can be fixed as well of course)
2006-01-22T17:37:39  <xorAxAx> and then there is the api problem
2006-01-22T17:37:55  <xorAxAx> of the formatter api. about calling order, symmetry etc.
2006-01-22T17:37:56  <Fabi> caching code path?
2006-01-22T17:38:02  <xorAxAx> Fabi: send_page in Page
2006-01-22T17:38:06  <xorAxAx> that does the caching
2006-01-22T17:38:29  <xorAxAx> but thats a minor issue
2006-01-22T17:39:15  <Fabi> why doesn't that work?
2006-01-22T17:39:21  <ThomasWaldmann> for me one of the main questions is if that is for 1.5.2 or 1.6 or 2.0
2006-01-22T17:39:24  <Fabi> just set the config var right?
2006-01-22T17:39:50  <xorAxAx> Fabi: its not tested :)
2006-01-22T17:40:07  <Fabi> it is at a "should work" state
2006-01-22T17:40:15  <Fabi> if it doesn't file a bug
2006-01-22T17:40:23  <xorAxAx> except for the weird escaping system
2006-01-22T17:40:38  <xorAxAx> but that can be fixed
2006-01-22T17:40:55  <xorAxAx> ThomasWaldmann: at least this can be solved without breaking anything
2006-01-22T17:41:01  <xorAxAx> ThomasWaldmann: and its not a big issue
2006-01-22T17:41:24  <xorAxAx> i think one can easily start with a few test cases that break the current system
2006-01-22T17:41:30  <Fabi> the only real problem I see are the formatters that are trying to keep state
2006-01-22T17:41:37  <ThomasWaldmann> well, i got the impression that this patch gets larger and still is expanding ... :)
2006-01-22T17:42:11  <Fabi> right now there is a evil hack to make some state information work for the wiki -> html case
2006-01-22T17:42:21  <ThomasWaldmann> otoh it might maybe fix some of our converter problems
2006-01-22T17:42:30  <xorAxAx> Fabi: most formatters do that
2006-01-22T17:42:54  <Fabi> may be we need some more elaborated code for other formatters
2006-01-22T17:43:01  <Fabi> xorAxAx,  to do what?
2006-01-22T17:43:24  <xorAxAx> Fabi: most formatters build stacks and other tree structures internally to "keep state"
2006-01-22T17:43:38  <xorAxAx> which is a bad thing because it shows the weakness of the api
2006-01-22T17:43:51  <xorAxAx> but we can/should not solve that IMHO
2006-01-22T17:44:07  <xorAxAx> the other formatter api things can be already tackled today
2006-01-22T17:44:16  <Fabi> if you need an internal stack it is simply impossible to cache part of the page
2006-01-22T17:44:25  <Fabi> simply don't use the cache
2006-01-22T17:44:33  <xorAxAx> yeah
2006-01-22T17:44:34  <Fabi> which is the default
2006-01-22T17:44:42  <Fabi> so where'S the problem?
2006-01-22T17:44:54  <xorAxAx> i dont see any
2006-01-22T17:45:18  <Fabi> may be this heading stuff is a bit evil
2006-01-22T17:45:31  <ThomasWaldmann> if we switch off cache completely, it will get 5x slower again
2006-01-22T17:45:40  <xorAxAx> ThomasWaldmann: ?
2006-01-22T17:45:49  <xorAxAx> ThomasWaldmann: nobody talks about switching it off
2006-01-22T17:45:54  <Fabi> I'd agree to change this if there is a real world example
2006-01-22T17:46:22  <xorAxAx> Fabi: change what?
2006-01-22T17:46:24  <xorAxAx> ?
2006-01-22T17:46:33  * xorAxAx would like to see his patches applied
2006-01-22T17:47:33  <ThomasWaldmann> lets talk about a time estimation
2006-01-22T17:48:02  <ThomasWaldmann> how much time do we need to get this into a "better than 1.5.1" state again?
2006-01-22T17:49:05  <ThomasWaldmann> (dont be overoptimistic as this usually proves wrong)
2006-01-22T17:49:26  <xorAxAx> depends on the goals ... completly rewriting the inner logic would solve this :)
2006-01-22T17:49:45  <xorAxAx> (i.e. tree representations, using graph transformations for things etc.)
2006-01-22T17:49:58  <ThomasWaldmann> the goal is to have less problems. e.g. with the converter.
2006-01-22T17:50:20  <ThomasWaldmann> a complete inner rewrite for sure isnt 1.5.x
2006-01-22T17:50:25  <xorAxAx> hmm, i dont know where the formatter sees the limit of the api here (havent checked it out)
2006-01-22T17:54:53  * Fabi must go
2006-01-22T17:56:07  <DasSheep> ever considered feeding the html to tidy in formatter just before displaying it?
2006-01-22T17:58:16  <ThomasWaldmann> that would be a platform dependant dependency
2006-01-22T17:58:44  <DasSheep> still could be an optional plugin
2006-01-22T17:59:08  * DasSheep wonders what would tidy do with all the span tags
2006-01-22T18:09:19  <ThomasWaldmann> http://www-128.ibm.com/developerworks/web/library/wa-ajaxintro2/?ca=dgr-lnxw07AJAX-Request
2006-01-22T18:11:14  <xorAxAx> ThomasWaldmann: so?
2006-01-22T18:12:24  <ThomasWaldmann> (for moin3 :)
2006-01-22T18:25:36  <ThomasWaldmann> 202.108.1. just made it on hosts_deny
2006-01-22T18:44:44  <ThomasWaldmann> ok, when nobody dares to estimate time frame (and helps staying within that), I think my feeling is right that it could take longer
2006-01-22T18:47:05  <xorAxAx> yeah
2006-01-22T18:47:23  <xorAxAx> i suggested already to do this on a separate branch ...
2006-01-22T18:50:24  <ThomasWaldmann> ok, so i make a moin--formatter--1.5 branch
2006-01-22T18:53:12  <xorAxAx> but i still dont know what you expect there
2006-01-22T18:53:31  <xorAxAx> making all formatters cachable? is that a reasonable goal?
2006-01-22T18:54:27  <ThomasWaldmann> the goal is making formatter output more clean, esp. the html formatter output
2006-01-22T18:54:54  <ThomasWaldmann> some gui edito converter crashes are due to invalid html
2006-01-22T18:55:38  <ThomasWaldmann> hmm, i think it will be moin--refactor--1.5
2006-01-22T18:57:07  <xorAxAx> so you propose to refactor on a branch?
2006-01-22T18:57:13  <xorAxAx> that is not really sensible IMHO
2006-01-22T18:57:27  <xorAxAx> i rather thought about small fixes on that branch
2006-01-22T18:57:46  <xorAxAx> if refactoring should take place, a maitenance branch might make sense
2006-01-22T18:57:53  <ThomasWaldmann> deron merandas stuff isnt small
2006-01-22T18:58:16  <xorAxAx> its not critical either ...
2006-01-22T18:58:44  <ThomasWaldmann> but it is refactoring
2006-01-22T18:58:56  <ThomasWaldmann> so you want a branch or not, decide
2006-01-22T18:59:14  <xorAxAx> depends on the goals. i wont do a branch for his patches
2006-01-22T18:59:31  <xorAxAx> the internal rewrite could be done in a branch, rather
2006-01-22T18:59:34  <ThomasWaldmann> and for the other stuff we need to complete it?
2006-01-22T18:59:42  <xorAxAx> complete?
2006-01-22T19:00:03  <ThomasWaldmann> see goal
2006-01-22T19:01:35  <xorAxAx> which goal?
2006-01-22T19:01:43  <ThomasWaldmann> (18:54) < ThomasWal> the goal is making formatter output more clean, esp. the html formatter output
2006-01-22T19:02:16  <xorAxAx> hmm, yeah
2006-01-22T19:02:26  <xorAxAx> but i am not sure if it is necessary
2006-01-22T19:02:35  <xorAxAx> as i said, i dont know the gui editor formatter very well
2006-01-22T19:02:54  <ThomasWaldmann> it is mostly the same as html formatter
2006-01-22T19:03:06  <ThomasWaldmann> only some stuff is specialcased
2006-01-22T19:03:21  <xorAxAx> and, we may forget that the biggest problem is the parser in this case
2006-01-22T19:03:27  <ThomasWaldmann> one of the main problems is that list nesting sucks
2006-01-22T19:03:31  <xorAxAx> it rather yields invalid sequences i guess
2006-01-22T19:05:43  <ThomasWaldmann> ok, so we might end up fixing the parser, too, to reach the goal
2006-01-22T19:07:50  <ThomasWaldmann> this really sounds like a refactor branch
2006-01-22T19:08:38  <ThomasWaldmann> and like stopping everything else except small bugfixes on the main branch
2006-01-22T19:09:38  <xorAxAx> in order to have a 1.6 then?
2006-01-22T19:10:31  <ThomasWaldmann> depends on how big that branch gets.
2006-01-22T19:12:40  <ThomasWaldmann> if we can do it with small to medium fixes, no user visible changes (except that it works better) and in a medium time frame, it could be 1.5.3 maybe
2006-01-22T19:19:55  <xorAxAx> ok
2006-01-22T19:49:35  <ThomasWaldmann> if we have to introduce moindom to reach the goal, it will rather get 1.6
2006-01-22T20:42:42  <ThomasWaldmann> moin--refactor--1.5 is open, please try to get it
2006-01-22T22:46:53  * ThomasWaldmann sets up a refactor branch test wiki