2006-06-09T08:26:23  <ThomasWaldman2> moin
2006-06-09T11:18:38  <ThomasWaldmann> nwp: reading?
2006-06-09T11:42:22  <Kepplar> moin
2006-06-09T11:42:26  <mvirkkil> ThomasWaldmann: hi
2006-06-09T11:42:38  <mvirkkil> and hi to Kepplar too
2006-06-09T11:42:43  <Kepplar> hi mvirkkil
2006-06-09T11:43:50  * ThomasWaldmann looked over nwp's auth changes. looking nice :)
2006-06-09T11:44:00  <Kepplar> what changes have been made?
2006-06-09T11:44:30  <ThomasWaldmann> splitting the cookie stuff into moin cookie auth and moin session cookie
2006-06-09T11:44:45  <Kepplar> hm cool
2006-06-09T11:44:54  <ThomasWaldmann> adding a check against mysql group db
2006-06-09T11:45:09  <Kepplar> isnt that storage?
2006-06-09T11:49:28  <ThomasWaldmann> no, it is an auth module currently
2006-06-09T11:50:10  <ThomasWaldmann> nwp: are your changes tested? what exactly is tested (and what's not)?
2006-06-09T12:25:29  <mvirkkil> ThomasWaldmann: You mentioned that maybe the functionality of the Include macro should be a basic part of the wiki-language. I agree wholehartedly :)
2006-06-09T12:26:31  <mvirkkil> ThomasWaldmann: This would make it trivial to expose "edit only this part" since the "this part" would be a separate page.
2006-06-09T12:26:57  <mvirkkil> A little like what is possible with mediawiki right now.
2006-06-09T12:29:55  <mvirkkil> but all this could naturally be done with macros too.
2006-06-09T12:30:18  <ThomasWaldmann> i think our include macro can also show an edit link
2006-06-09T12:31:39  <ThomasWaldmann> my reasoning was rather along the lines that the most fundamental stuff should be markup
2006-06-09T12:31:59  <ThomasWaldmann> and if linking is fundamental, include is also
2006-06-09T12:32:46  <ThomasWaldmann> and we have less code if we remove image links magically being kind of includes
2006-06-09T12:36:33  <ThomasWaldmann> but the more "special effects" the linking stuff has to support, the less likely it will get markup
2006-06-09T12:36:43  <ThomasWaldmann> s/linking/including/
2006-06-09T12:37:22  <mvirkkil> And this was what I was agreeing with. I mentioned macros could support the "edit-section" part..
2006-06-09T12:37:37  <mvirkkil> But yeah, including should be a fundamental part of the language.
2006-06-09T12:38:42  <mvirkkil> About the syntax: something like [->PageToInclude] where the arrow part would signify that we are oing to include the rest from there.
2006-06-09T12:39:10  <ThomasWaldmann> that rather looks like a link :)
2006-06-09T12:39:24  <mvirkkil> ThomasWaldmann: Well, that was what I was aming for :)
2006-06-09T12:39:41  <mvirkkil> But do you think it look _too_ much like a link?
2006-06-09T12:40:07  <ThomasWaldmann> -> is a pointer, a pointer is a link, reference, whatever you call that
2006-06-09T12:40:17  <mvirkkil> hmm.. ok.
2006-06-09T12:40:25  <ThomasWaldmann> so that would be rather confusing
2006-06-09T12:40:35  <ThomasWaldmann> <<<ItemToInclude>>>
2006-06-09T12:40:40  <mvirkkil> [+PageToInclude]
2006-06-09T12:41:00  <mvirkkil> your version is more stylish :)
2006-06-09T12:41:53  <ThomasWaldmann> the question is rather if we can have that so simple or if we really need a ton of params
2006-06-09T12:42:01  <mvirkkil> <<<ItemToInclude, style="width: 3">>>
2006-06-09T12:42:10  <mvirkkil> ahh, just thinking about how to add params :)
2006-06-09T12:42:41  <mvirkkil> ThomasWaldmann: I like parameters, as long as they aren't necessary for common cases.
2006-06-09T12:43:13  <mvirkkil> <<<ItemToInclude, class="screenshot">>>
2006-06-09T12:47:21  <ThomasWaldmann> i could have a page called ItemToInclude, class="screenshot"
2006-06-09T12:48:33  <ThomasWaldmann> anyway, gtg & bbl
2006-06-09T13:01:25  <xorAxAx> umm, introducing include markup doesnt solve any problems :)
2006-06-09T13:01:38  <xorAxAx> that were discussed in the last days
2006-06-09T13:14:43  <mvirkkil> xorAxAx: No, it sloves no problems, but I like the idea of having it embedded in the language.
2006-06-09T13:15:54  <xorAxAx> yeah, there are long discussions about it in the wiki
2006-06-09T13:16:00  <mvirkkil> xorAxAx: You basically solved the include problem in a very simple and elegant way with the filterparser
2006-06-09T13:16:28  <mvirkkil> xorAxAx: and the items that aren't parseable could be handled in the macro
2006-06-09T13:16:44  <xorAxAx> sure, but the include macro can output tocs as well
2006-06-09T13:17:25  <mvirkkil> I just like the idea of having include in the language, instead of keeping it a macro.
2006-06-09T13:18:11  <mvirkkil> xorAxAx: One reason could be that the include stuff should be done before parsing and macros... but that's also just philosophy.
2006-06-09T13:18:50  <mvirkkil> I have no real reasons, I just like the idea and the architecture of handling include first, and then do the rest on the combined document.
2006-06-09T13:19:31  <mvirkkil> actually, that idea, making the inclusions first is kind of against how the wikiparser works right now.
2006-06-09T13:20:35  <mvirkkil> In either case, the Include macro and the TOC macro should get fixed. But then again, I don't think anyone disagrees with that part.
2006-06-09T13:21:37  <mvirkkil> xorAxAx: I might have a go at writing the filter-parser and include macro.
2006-06-09T13:22:02  <mvirkkil> xorAxAx: But I don't feel I'm going to be able to tackle the TOC counting formatter.
2006-06-09T13:22:46  <xorAxAx> hehe :)
2006-06-09T13:23:00  <mvirkkil> Or the counting part is easy, I'm just a little worried about the generation of IDs, and how to pass them to the real formatter.
2006-06-09T13:40:02  <mvirkkil> I'd need a feature for my own code that does the following: if formatter is docbook: Include(pagename) else: formatter.text(pagename)
2006-06-09T13:40:26  <mvirkkil> So I'm going to be using the Include macro only to include whole pages anyway..
2006-06-09T13:40:39  <Kepplar> moin
2006-06-09T13:41:05  <mvirkkil> The macro I wrote is like 4 lines long, and uses the include macro.
2006-06-09T13:41:29  <mvirkkil> It's called DocBookChapter
2006-06-09T13:42:39  <mvirkkil> but it would work a lot faster if it would use the "new Include-macro", since now the Include macro returns a whole freaking docbook document as a string, and I need to re-parse it to include it in the parent page.
2006-06-09T13:43:09  <mvirkkil> parent page == the docbook that the chapter should be included in.
2006-06-09T13:43:56  <Kepplar> anyone know a mobile phone with gps and digital compass accessible via J2ME?
2006-06-09T13:45:09  <mvirkkil> Kepplar: I don't even know of a mobile phone with gps.
2006-06-09T13:45:22  <mvirkkil> Kepplar: Usually the gps stuff is a bluetooth addon.
2006-06-09T13:45:39  <mvirkkil> Kepplar: So the issue is whether J2ME can handle bluetooth.
2006-06-09T13:45:49  <Kepplar> i know of a few
2006-06-09T13:45:54  <Kepplar> well the digital compass is most important
2006-06-09T13:45:58  <Kepplar> as gps is easy
2006-06-09T14:04:27  <nwp> ThomasWaldmann: well, I'm using all the patches I posted, but all together. I haven't used them in combination with others.
2006-06-09T14:04:50  <nwp> so I haven't yet tried anything other than LDAP with the moin_session cookie, for example
2006-06-09T14:04:56  <nwp> but it should be simple to do
2006-06-09T14:05:25  <nwp> may need a few small changes to decouple other auth methods from the assumption that moin_cookie is used for session info though
2006-06-09T14:05:35  <nwp> similar to the changes to the LDAP in that patch
2006-06-09T20:37:28  <ThomasWaldmann> re
2006-06-09T20:37:53  <fpletz> moin ;)
2006-06-09T20:38:06  <fpletz> got the indexing to work properly
2006-06-09T20:38:32  <ThomasWaldmann> nwp: so you tested ldap, mysql and session
2006-06-09T20:39:31  <ThomasWaldmann> fpletz: didnt it work?
2006-06-09T20:40:28  <fpletz> ThomasWaldmann: yes, a few small code changes were required.. and --mode=add of course ;)
2006-06-09T20:41:04  <ThomasWaldmann> how's moin qp going?
2006-06-09T20:44:22  <fpletz> working on that now.. btw. I understood you correctly that I should just rewrite your xapian_term functions to return xapian.Query objects instead of strings to feed xapians qp?
2006-06-09T20:44:57  <ThomasWaldmann> yes
2006-06-09T20:45:07  <fpletz> great
2006-06-09T20:45:11  <ThomasWaldmann> maybe hack RawQuery from xapwrap first
2006-06-09T20:45:44  <fpletz> yep, will finish this probably this evening
2006-06-09T20:46:12  <ThomasWaldmann> so that it accepts and detects unicode objects
2006-06-09T20:56:58  <ThomasWaldmann> how about splitting auth stuff into single files per auth method (with some basic stuff in auth/__init__.py)?
2006-06-09T21:00:54  <ThomasWaldmann> nwp: using the session stuff, we could kill the normal MOIN_ID cookie as it was, right?
2006-06-09T21:02:10  <xorAxAx> ThomasWaldmann: if sessions work, yes
2006-06-09T21:03:07  <ThomasWaldmann> nwp: does it have some special reason you implemented moin_session_cookie_name?
2006-06-09T21:21:18  <fpletz> ThomasWaldmann: I'm pretty sure I just stumbled on a bug in mainline 1.6 in the wikiutil.MimeType class.. can I just commit my working fix to the 1.6 repo?
2006-06-09T21:22:03  <xorAxAx> fpletz: IMHO yes
2006-06-09T21:23:04  <ThomasWaldmann> fpletz: put a diff on my homepage
2006-06-09T21:24:23  <fpletz> ThomasWaldmann: one the moinmoin wiki?
2006-06-09T21:24:29  <ThomasWaldmann> yes
2006-06-09T21:24:32  <fpletz> ok
2006-06-09T21:27:48  <fpletz> hmm, how to do that properly? it's a onliner anyway ;)
2006-06-09T21:28:23  <fpletz> one ^^
2006-06-09T21:49:32  <ThomasWaldmann> you can also just tell me :)
2006-06-09T21:50:06  <fpletz> yeah ;)
2006-06-09T21:50:22  <fpletz> MoinMoin/wikiutil.py, line 790
2006-06-09T21:50:27  <fpletz> mtype, encoding = mimetypes.guess_type(filename)
2006-06-09T21:50:37  <fpletz> guess_type takes filename, not nothing ;)
2006-06-09T21:56:11  <ThomasWaldmann> fixed :)
2006-06-09T21:56:26  <xorAxAx>  This
2006-06-09T21:56:27  <xorAxAx> shipment is a small fun something for all of our SoCers.
2006-06-09T21:56:35  <xorAxAx> fpletz: have you got the present already?
2006-06-09T21:56:50  <ThomasWaldmann> T-Shirts?
2006-06-09T21:56:56  <fpletz> xorAxAx: nope, nothing yet
2006-06-09T21:56:59  <ThomasWaldmann> Living Pythons?
2006-06-09T21:57:03  <fpletz> I was wondering what this might be ;)
2006-06-09T21:57:14  <xorAxAx> a block of chocolate? :)
2006-06-09T21:57:22  <fpletz> well, the t-shirts aren't ready yet afaik
2006-06-09T21:57:41  <fpletz> yeah, google-chocolate :p
2006-06-09T22:00:35  <xorAxAx> "it is italian"
2006-06-09T22:00:42  <xorAxAx> deep-throzen pizza?
2006-06-09T22:01:14  <ThomasWaldmann> any comments before i tear auth stuff?
2006-06-09T22:06:03  <fpletz> xorAxAx: perhaps the google publicity package.. stickers, lanyard, cd with free google software (toolbar, picasa, ...)
2006-06-09T22:06:13  <fpletz> and if we're lucky a mug oder some google teddy ;)
2006-06-09T22:07:27  <xorAxAx> hehe
2006-06-09T22:07:44  <xorAxAx> or a copy of the chinese law
2006-06-09T22:07:59  <fpletz> lol
2006-06-09T22:57:47  * ThomasWaldmann killed moin_cookie
2006-06-09T23:53:59  <nwp> ThomasWaldmann: the name bit was just to be able to share  code with moin_cookie. but I think configurable name is probably a good thing
2006-06-09T23:56:49  <ThomasWaldmann> i removed moin_cookie. it is now called moin_login and just creates a user object for moin_session to do the session cookie setting.
2006-06-09T23:56:57  <nwp> ThomasWaldmann: cool
2006-06-09T23:57:10  <nwp> I think that's the Right Thing
2006-06-09T23:57:28  <ThomasWaldmann> and the cookie name is a constant MOIN_SESSION = 'MOIN_SESSION' in auth module
2006-06-09T23:58:18  <nwp> people using package managers won't find that very configurable ;-)
2006-06-09T23:58:19  <ThomasWaldmann> i am currently trying to make it more secure (== decipher Frankie Chows stuff)
2006-06-09T23:58:30  <nwp> what was his problem?
2006-06-09T23:59:14  <ThomasWaldmann> http://moinmoin.wikiwikiweb.de/MoinMoinBugs/Cookie_is_not_secure_enough
2006-06-09T23:59:20  <ThomasWaldmann> http://moinmoin.wikiwikiweb.de/MoinMoinPatch/SecurityString
2006-06-09T23:59:56  <ThomasWaldmann> your stuff is half way to his

MoinMoin: MoinMoinChat/Logs/moin-dev/2006-06-09 (last edited 2007-10-29 19:10:58 by localhost)