MoinMoin: MoinMoinChat/Logs/moin-dev/2006-06-10

2006-06-10T00:01:08  <ThomasWaldmann> I just need to refactor it for 1.6, it was written for 1.3 or 1.5.
2006-06-10T00:01:24  <nwp> what is the difference between his "security string" and a password?
2006-06-10T00:01:54  <ThomasWaldmann> it is like your cookie_secret, but not site wide, but user configurable
2006-06-10T00:02:01  <ThomasWaldmann> and regenerated often
2006-06-10T00:02:29  <ThomasWaldmann> (on logout, on password mail, ...)
2006-06-10T00:02:40  <nwp> what does that gain?
2006-06-10T00:03:06  <ThomasWaldmann> e.g. if you login on a internet cafe computer and forget to logout
2006-06-10T00:03:41  <ThomasWaldmann> next time you logout somewhere else, the old cookie will be invalid
2006-06-10T00:04:45  <nwp> the easy way to do that is to store the login time for the session in the user prefs and in the cookie.
2006-06-10T00:05:04  <nwp> it's more comprehensible than changing a random meaningless thing at various times
2006-06-10T00:05:35  <ThomasWaldmann> expiry?
2006-06-10T00:05:56  <nwp> the cookie is valid until either you expire it, logout, or login again
2006-06-10T00:06:18  <nwp> at the moment I think there's timed expiry in the code isn't there anyway
2006-06-10T00:06:51  <ThomasWaldmann> so if i login from computer A at 10:00, expiry is +12h
2006-06-10T00:06:59  <nwp> or whatever you configure
2006-06-10T00:07:08  <ThomasWaldmann> so what can I do from computer B to make that invalid?
2006-06-10T00:07:13  <nwp> login
2006-06-10T00:07:34  <ThomasWaldmann> why does that make the old cookie invalid?
2006-06-10T00:07:35  <nwp> doesn't happen at the moment but it's fairly trivial to add
2006-06-10T00:07:52  <nwp> and like I said, much easier to understand than using a separate random string
2006-06-10T00:08:11  <nwp> you could generate a random session id at login, but...
2006-06-10T00:08:54  <nwp> this "security string" basically only needs to be *something* that is an identifier for the session that creates it
2006-06-10T00:09:24  <nwp> so you have a "current_session" pref which identifies it, and all others are invalid
2006-06-10T00:09:31  <nwp> current_session being set on login
2006-06-10T00:10:41  <nwp> login would make the old cookie invalid by updating the "current_session" in the prefs
2006-06-10T00:11:10  <nwp> the problem with this is that it means a user can't have two browsers open to the same site
2006-06-10T00:11:19  <nwp> which actually can be a real problem
2006-06-10T00:11:32  <ThomasWaldmann> that's just another (better :) name for security_string :)
2006-06-10T00:12:06  <nwp> yes... and tidying up the concept a little to work out that login is probably the only time you should be changing it
2006-06-10T00:12:47  <ThomasWaldmann> what if you change it on logout? 2 browsers logging in ...
2006-06-10T00:12:52  <nwp> but to use it or not should be optional because of the poor developers who are trying to make a theme work in two different browsers ;-)
2006-06-10T00:13:24  <nwp> um... don't have the code in front of me. You might need to zero it on logout
2006-06-10T00:13:55  <ThomasWaldmann> no, it just must be different on logout
2006-06-10T00:14:11  <nwp> trying to remember what happens on logout at the moment
2006-06-10T00:16:26  <ThomasWaldmann> btw, why do you store username in the cookie?
2006-06-10T00:17:04  <nwp> hang on, just reading code to remind myself what it actually does do
2006-06-10T00:17:43  <ThomasWaldmann> maybe hg pull to read current code :)
2006-06-10T00:18:40  <nwp> that takes forever! ;-)
2006-06-10T00:18:53  <ThomasWaldmann> why?
2006-06-10T00:19:11  <nwp> always seems to
2006-06-10T00:19:30  <nwp> probably latency
2006-06-10T00:19:36  <ThomasWaldmann> the initial clone take some time, as it has to transfer xx mb
2006-06-10T00:19:51  <ThomasWaldmann> but pull should be fast after that
2006-06-10T00:22:28  <nwp> wasn't too bad ;-)
2006-06-10T00:23:03  <nwp> definitely a good move reorganising auth.py to a directory, btw
2006-06-10T00:23:17  <nwp> much easier to keep sync with local changes
2006-06-10T00:23:22  <nwp> when they are in a separate file
2006-06-10T00:25:14  <ThomasWaldmann> pull again, i had still some changes
2006-06-10T00:25:50  <ThomasWaldmann> the stuff is still in one file, but I might tear it soon
2006-06-10T00:29:01  <nwp> OK, yes, looks like we need to do something on logout to invalidate cookie
2006-06-10T00:29:13  <nwp> (on the server, rather than trusting browser)
2006-06-10T00:29:36  <nwp> or do we
2006-06-10T00:29:44  <nwp> hang on, what does that help?
2006-06-10T00:30:19  <nwp> only case it helps is where a user can't trust their browser to zap the cookie, isn't it?
2006-06-10T00:30:31  <nwp> whatever, it's a good idea
2006-06-10T00:31:13  <ThomasWaldmann> or if you used a machine you cant reach any more
2006-06-10T00:31:32  <fpletz> ThomasWaldmann: nearly works now. I'll just have to parse and replace the prefixes manully, which I'll implement tomorrow morning along with some minor cleanups.. too tired to go on now, though ;)
2006-06-10T00:31:40  <nwp> but if you didn't logout, then what you do on logout has no effect
2006-06-10T00:31:41  <fpletz> good night, guys
2006-06-10T00:31:51  <ThomasWaldmann> gn fpletz :)
2006-06-10T00:32:34  <ThomasWaldmann> nwp: if we calculate a new user secret on logout on another machine, the existing cookie on first machine is invalidated
2006-06-10T00:32:53  <ThomasWaldmann> so logout is global logout, for all machines, all browsers
2006-06-10T00:32:53  <nwp> but before you logout on the other machine, you have to log in on it :-)
2006-06-10T00:33:11  <ThomasWaldmann> and?
2006-06-10T00:33:28  <nwp> session id would change on login, so all other ids are invalid
2006-06-10T00:33:38  <nwp> but yes, logout should be complete
2006-06-10T00:33:48  <ThomasWaldmann> if you do it on login you cant use 2 browsers
2006-06-10T00:34:06  <nwp> eek
2006-06-10T00:34:21  <nwp> that's getting evil though
2006-06-10T00:34:27  <ThomasWaldmann> hehe
2006-06-10T00:34:40  <nwp> and evil => confusion => bugs
2006-06-10T00:35:11  <nwp> it is kind of seductive though...
2006-06-10T00:35:18  <ThomasWaldmann> i think documenting logout as "global logout" can be understood by users
2006-06-10T00:35:23  <nwp> yes
2006-06-10T00:35:32  <ThomasWaldmann> some users might even EXPECT that :)
2006-06-10T00:35:32  <nwp> it's the programmers I'm thinking of
2006-06-10T00:35:55  <nwp> so...
2006-06-10T00:36:04  <nwp> session_id is examined on login
2006-06-10T00:36:10  <nwp> if it doesn't exist, it's created
2006-06-10T00:36:18  <nwp> if it does (from last logout), it is used
2006-06-10T00:36:37  <nwp> on logout, id for next session is generated
2006-06-10T00:36:45  <nwp> aha
2006-06-10T00:36:56  <nwp> but what if logout is by timeout?
2006-06-10T00:37:14  <nwp> or by closing the browser?
2006-06-10T00:37:31  <nwp> then you would just re-use the last id
2006-06-10T00:37:37  <ThomasWaldmann> closing browser does not logout
2006-06-10T00:38:41  <ThomasWaldmann> and if we receive an expired cookie, it is just invalid and does not do a session
2006-06-10T00:38:41  <nwp> OK one thing that definitely should be added is the expiry time in the cookie
2006-06-10T00:38:54  <nwp> I even commented on it in the code :-o
2006-06-10T00:39:18  <nwp> yes, problem is "when does a new session id get generated if the user does not perform an explicit logout?"
2006-06-10T00:39:56  <nwp> I think you have to do it on login and hence prevent use of two browsers
2006-06-10T00:41:04  <ThomasWaldmann> hmm
2006-06-10T00:41:28  <nwp> hence make this check optional, default enabled
2006-06-10T00:41:55  <nwp> and anyone who really cares can disable it and be a little less secure
2006-06-10T00:43:19  <nwp> and blank/zero stored id on logout (when logged out there is no valid session, so it should indicate this)
2006-06-10T00:43:49  <nwp> I have to shower and have breakfast now... got visitor coming in a bit.
2006-06-10T00:44:35  <nwp> I can I'll have another look in an hour or so
2006-06-10T00:44:57  <ThomasWaldmann> ok, maybe I will still be there and not ZzzZzzz
2006-06-10T01:01:24  <ThomasWaldmann> nwp: we could also reuse the last session id if there was no logout. expired cookies will be expired nevertheless.
2006-06-10T09:27:26  <ThomasWaldmann> moin
2006-06-10T11:02:24  <fpletz> ThomasWaldmann: is the linksearch case-sensitive?
2006-06-10T11:02:41  <fpletz> I would suggest not ;)
2006-06-10T11:04:39  <fpletz> because there is a difference in the current implementations for moinSearch and xapian.. while the regular search _is_ case sensitive, xapian isn't (all lowercase)
2006-06-10T11:05:57  <ThomasWaldmann> hmm
2006-06-10T11:06:15  <ThomasWaldmann> i guess we have to post-process those results then
2006-06-10T11:06:33  <fpletz> well, I can make xapian case sensitive if that's an issue
2006-06-10T11:07:25  <ThomasWaldmann> doesnt that conflict with it's prefix stuff?
2006-06-10T11:07:26  <fpletz> the results are currently post-process by moinSearch.. so when i search for linkto:frontpage, I get tons of results from xapian but nothing will be shown as moinSearch filters those out
2006-06-10T11:07:56  <fpletz> no, specs say, that in this case we have to insert a : between prefix and the word
2006-06-10T11:08:36  <fpletz> http://svn.xapian.org/trunk/xapian-applications/omega/docs/termprefixes.txt?view=co
2006-06-10T11:08:53  <fpletz> 4. paragraph
2006-06-10T11:09:16  <ThomasWaldmann> ok, but how do you want to index stuff then?
2006-06-10T11:10:06  <fpletz> just like now but now lowercasing the links
2006-06-10T11:10:25  <fpletz> XLINKTOfrontpage -> XLINKTO:FrontPage
2006-06-10T11:10:25  <fpletz> s/now/not/
2006-06-10T11:11:16  <fpletz> well, if case is in fact an issue.. personally, I think it's not
2006-06-10T11:12:16  <ThomasWaldmann> hmm, do we have code so we can special treat the links when indexing?
2006-06-10T11:12:52  <ThomasWaldmann> ah forget that, we use links cache iirc
2006-06-10T11:13:01  <fpletz> as far as I can see this would be pretty easy..
2006-06-10T11:13:09  <fpletz> which links cache? ;)
2006-06-10T11:13:17  <ThomasWaldmann> pagelinks cache
2006-06-10T11:14:04  <fpletz> mmh, dunno.. the current code adds all pagenames returned by page.getPageLinks(request)
2006-06-10T11:14:32  <ThomasWaldmann> yes and this uses the pagelinks cache :)
2006-06-10T11:15:05  <ThomasWaldmann> then please do that case-sensitive, so we don't generate wrong results
2006-06-10T11:15:20  <fpletz> ok, will do
2006-06-10T11:29:18  <fpletz> mmh, everything seems to work now.. I surely forgot something ;)
2006-06-10T11:30:43  <ThomasWaldmann> oh, finally some commits :)
2006-06-10T11:31:54  <ThomasWaldmann> does anybody know how to add some sys.path.insert(0, packagedir) to distutils setup.py?
2006-06-10T11:32:08  <fpletz> yeah, did already some pushes yesterday ;)
2006-06-10T11:32:24  <ThomasWaldmann> packagedir should correspond to what the user set with --home or --prefix or whatever
2006-06-10T11:33:12  <ThomasWaldmann> i need that for "moin" script to find non-standard MoinMoin/ location
2006-06-10T11:33:45  <fpletz> mmh
2006-06-10T11:34:18  <xorAxAx> ThomasWaldmann: how are moin and setup.py related?
2006-06-10T11:34:46  <ThomasWaldmann> http://moinmoin.wikiwikiweb.de/MoinMoinBugs/BinMoinCannotImport1.5.3
2006-06-10T11:35:35  <ThomasWaldmann> moin command works for std site-packages location, but not for custom install locations
2006-06-10T11:39:18  <xorAxAx> isnt "our" code generating the scripts?
2006-06-10T11:46:58  <ThomasWaldmann> sure
2006-06-10T11:47:18  <ThomasWaldmann> but i still need to know where it is going to be installed to
2006-06-10T11:52:39  <ThomasWaldmann> could be build_dir
2006-06-10T12:08:23  <ThomasWaldmann> no, it's a temp dir
2006-06-10T13:05:03  <ThomasWaldmann> xorAxAx: you could do a similar modification to the win32 stuff as my last changeset
2006-06-10T13:06:22  <xorAxAx> i dont see how i can fix it
2006-06-10T13:06:34  <xorAxAx> its still not doing what you wanted to do initially :)
2006-06-10T13:07:12  <xorAxAx> and the code that you currently generate is never valid (because python != python2.x) or would help (because site.py adds it already to sys.path)
2006-06-10T13:12:29  <ThomasWaldmann> it is just a sample and the comment tells it has to be fixed
2006-06-10T13:12:46  <ThomasWaldmann> you could do at least the same for win32
2006-06-10T13:18:03  <xorAxAx> it would look exactly the same and i cannot give a sensible sample path
2006-06-10T13:18:08  <xorAxAx> so i dont see the point
2006-06-10T13:19:25  <ThomasWaldmann> the point is giving users something easy to modify
2006-06-10T13:19:39  <ThomasWaldmann> they can find out the path, but they can't code in python
2006-06-10T13:20:09  <xorAxAx> thats the wrong approach IMHO
2006-06-10T13:20:26  <ThomasWaldmann> feel free to do it the same or better :)
2006-06-10T13:20:29  <xorAxAx> you have to do quite some tricks to use that moin.bat script
2006-06-10T13:20:42  <xorAxAx> (because python doesnt put the scripts directory into the PATH)
2006-06-10T13:21:10  <ThomasWaldmann> that's not our fault
2006-06-10T13:21:14  <xorAxAx> ThomasWaldmann: sure, i dont know how to fix this problem (its a) python and b) distutils)
2006-06-10T13:22:12  <ThomasWaldmann> what does %%$ and %%* mean?
2006-06-10T13:22:52  <xorAxAx> %% = quoted %
2006-06-10T13:22:57  <xorAxAx> %* = all parameters
2006-06-10T13:23:05  <xorAxAx> i dont know %$ ... reading docs ...
2006-06-10T13:23:38  <ThomasWaldmann> must be the same for 4dos
2006-06-10T13:25:52  <xorAxAx> not documented
2006-06-10T13:26:37  <ThomasWaldmann> i guess some set PYTHONPATH=... would do it for win32
2006-06-10T13:51:37  <ThomasWaldmann> fpletz: can't one just use : in general?
2006-06-10T13:52:35  <fpletz> ThomasWaldmann: sure, because moin is the only searcher
2006-06-10T13:54:37  <ThomasWaldmann> maybe that's easier than querying pattern[0] every time
2006-06-10T13:55:10  <fpletz> ok
2006-06-10T13:58:50  <ThomasWaldmann> moin richardb :)
2006-06-10T13:58:59  <richardb> helo
2006-06-10T13:59:52  <fpletz> hi richardb
2006-06-10T14:00:17  <richardb> Hi.  Looks like you've been active. :)
2006-06-10T14:00:28  <fpletz> yup ;)
2006-06-10T14:06:04  <richardb> Woohoo - it works. ;-)
2006-06-10T14:06:36  <richardb> Haven't got time to hang around, but that's a good start fpletz.
2006-06-10T14:06:41  <richardb> Keep it up. ;-)
2006-06-10T14:07:09  <fpletz> yes, I'm going to implement the stemming
2006-06-10T14:07:35  <richardb> Which python bindings to snowball are you going to use?
2006-06-10T14:08:16  <fpletz> dunno yet, I'm currently working something else
2006-06-10T14:08:17  <richardb> I really should pick one of the bindings, and make an official version.
2006-06-10T14:12:21  <ThomasWaldmann> fpletz: maybe wait until xapian has utf-8 stemmers included
2006-06-10T14:13:08  <fpletz> ThomasWaldmann: when will that happen? with utf-8 capable qp?
2006-06-10T14:13:18  <richardb> utf-8 stemmers will be in 1.0
2006-06-10T14:13:25  <richardb> But when that will be is hard to say.
2006-06-10T14:13:30  <richardb> Probably within a year. ;-)
2006-06-10T14:13:32  <ThomasWaldmann> maybe before hurd final release :)
2006-06-10T14:13:55  <fpletz> lol, ok
2006-06-10T14:13:57  <ThomasWaldmann> fpletz: did you test your qp stuff
2006-06-10T14:14:05  <richardb> Olly doesn't want to break existing databases for a non major release: hence the delay.
2006-06-10T14:14:18  <richardb> (utf-8 qp will be at same time)
2006-06-10T14:14:27  <fpletz> yes, but we should probably include some tests
2006-06-10T14:17:21  <richardb> If you want to do stemming now, I recommend using PyStemmer, and just catch ImportError and fallback to no stemming.
2006-06-10T14:18:48  <richardb> Ah.
2006-06-10T14:19:04  <richardb> On second thoughts, not.  it doesn't seem to compile with recent GCCs.
2006-06-10T14:19:19  <fpletz> oh..
2006-06-10T14:19:45  <richardb> I wonder if PySnowballStemmer works.
2006-06-10T14:20:48  <richardb> looks better.
2006-06-10T14:21:19  <richardb> Hmm, works, but no documentation.
2006-06-10T14:21:27  <richardb> I really need to sort this out.
2006-06-10T14:25:00  <richardb> Oh, this is a terrible wrapper.
2006-06-10T14:25:02  <richardb> *sigh*
2006-06-10T14:25:14  <ThomasWaldmann> fpletz: could we have utf-8 encoding in some Query object __init__?
2006-06-10T14:27:33  <fpletz> ThomasWaldmann: yeah, ok, I'll derive a new class from xapian.Query
2006-06-10T14:28:15  <ThomasWaldmann> with similar magic as the others have
2006-06-10T14:28:59  <ThomasWaldmann> and maybe have some __init__(self, ..., encoding=UNICODE_ENCODING)
2006-06-10T14:29:23  <richardb> Okay - I'm shocked at how bad one of the python wrappers for snowball is, and the other one looks okay but has bit-rotted and doesn't compile.
2006-06-10T14:29:42  <richardb> I'll have to fix this up in the near future, but there aren't any useful python bindings now.
2006-06-10T14:30:04  <richardb> (Though someone said something about Zope having some.  Hmm)
2006-06-10T14:31:23  <richardb> fpletz: I may get this done this afternoon.  If not, won't be done for a while.
2006-06-10T14:31:42  <ThomasWaldmann> coding C and Python at the same time might have some negative effect on the python code X)
2006-06-10T14:31:49  <richardb> Getting the infrastructure in place to tell the indexer what language to stem in is probably the hard bit, anyway.
2006-06-10T14:32:10  <richardb> I've been coding C/Python/Javascript pretty much simultaneously for the last two years...
2006-06-10T14:32:20  <fpletz> richardb: ok
2006-06-10T14:32:21  <richardb> It mainly has a negative effect on your sanity.
2006-06-10T14:32:31  <richardb> Particularly the javascript.. ;-)
2006-06-10T14:32:31  <ThomasWaldmann> lol
2006-06-10T14:34:02  <fpletz> lol, did you use a library for js? it's not that bad if you use mochikit, looks more phythonic then ;)
2006-06-10T14:34:14  * ThomasWaldmann makes some index for xapian.wikiwikiweb.de
2006-06-10T14:36:52  <ThomasWaldmann> fpletz: we should have the possibility of a common farm index_dir
2006-06-10T14:37:25  <ThomasWaldmann> each wiki's index having its cfg.siteid in the filename
2006-06-10T14:37:42  <ThomasWaldmann> per default, each wiki uses its own index only
2006-06-10T14:38:19  <ThomasWaldmann> but optionally, it should have other wiki's names in the index list
2006-06-10T14:38:53  <ThomasWaldmann> (similar to how user_dir is handled)
2006-06-10T14:39:06  <richardb> Note that xapian is happy to search across several indexes simultaneously.
2006-06-10T14:39:12  <fpletz> right, putting that on my TODO list
2006-06-10T14:39:15  <richardb> (See Database.addDatabase())
2006-06-10T14:39:23  <fpletz> oh, interesting
2006-06-10T14:39:30  <richardb> Scales happily up to at least 20 databases.
2006-06-10T14:39:46  <richardb> Or more, depending on load.
2006-06-10T14:40:13  <ThomasWaldmann> fpletz: and in the index, we should have the siteid, so that we can do interwiki links in search results
2006-06-10T14:40:31  <ThomasWaldmann> (this should be checked if it works, then)
2006-06-10T14:40:57  <fpletz> ah, ok
2006-06-10T14:41:09  <ThomasWaldmann> s/siteid/interwikiname/ maybe better
2006-06-10T14:43:04  <ThomasWaldmann> i guess i trash that _ url stuff now
2006-06-10T14:46:42  <Kepplar> hey
2006-06-10T14:46:45  <Kepplar> python question
2006-06-10T14:47:07  <Kepplar> how do i find the type of a object?
2006-06-10T14:47:34  <fpletz> type(obj)? ;)
2006-06-10T14:47:40  <Kepplar> ok, returns?
2006-06-10T14:47:54  <richardb> A type object.
2006-06-10T14:48:17  <Kepplar> er hassle
2006-06-10T14:48:22  <Kepplar> i need it as a string really
2006-06-10T14:48:40  <richardb> Are you trying to test if something is an instance of a particular class?
2006-06-10T14:48:48  <richardb> If so, use "isinstance"
2006-06-10T14:50:29  <richardb> There isn't necessarily a unique string describing a type.
2006-06-10T14:50:58  <richardb> Might be worth reading the documentation on the types module:
2006-06-10T14:51:01  <richardb> http://docs.python.org/lib/module-types.html
2006-06-10T14:51:20  <ThomasWaldmann> Kepplar: often you want to use instanceof()
2006-06-10T14:51:35  <ThomasWaldmann> ehrm, isinstance
2006-06-10T14:52:10  <Kepplar> yep
2006-06-10T14:52:27  <Kepplar> im just giving the flatfile storage engine the ability to work out where to put something based on the type of the item revision
2006-06-10T14:52:31  <Kepplar> its a hack
2006-06-10T14:52:49  <richardb> Definitely sounds like a place to use "isinstance"
2006-06-10T14:52:57  <Kepplar> but this logic cannot be stored in the abstact layer as some systems will store them in the same kind of place
2006-06-10T14:53:07  <Kepplar> the flatfile system is just legacy
2006-06-10T14:53:14  <Kepplar> yep, thanks for that
2006-06-10T14:55:03  <ThomasWaldmann> what does the item type have to do with "where to put something"?
2006-06-10T14:56:10  <ThomasWaldmann> you mean attachments vs. wiki pages for the old stuff?
2006-06-10T14:56:20  <Kepplar> user
2006-06-10T14:56:21  <Kepplar> page
2006-06-10T14:56:21  <Kepplar> etc
2006-06-10T14:56:56  <Kepplar> data/pages   /data/user
2006-06-10T14:57:11  <ThomasWaldmann> xapian.wikiwikiweb.de now has an index and xapian_search = 1
2006-06-10T14:57:33  <fpletz> great
2006-06-10T14:57:55  <ThomasWaldmann> will add some documents...
2006-06-10T14:58:46  <Kepplar> Schumacher on the second row of the grid ^_^
2006-06-10T14:58:52  * Kepplar is a McClaren fan
2006-06-10T15:05:52  <ThomasWaldmann> http://xapian.wikiwikiweb.de/TestWiki - with doc/xls/pdf/sxw stuff
2006-06-10T15:15:45  <fpletz> mmh, some there are some problems.. I'll look into it
2006-06-10T15:15:53  <fpletz> seems
2006-06-10T15:19:30  <ThomasWaldmann> pdf doesnt find the plugin
2006-06-10T15:20:00  <ThomasWaldmann> it uses application_octet_stream although there is application_pdf
2006-06-10T15:21:01  <ThomasWaldmann> same for msword
2006-06-10T15:24:11  <Kepplar> does wiki utils have a item revision number generator?
2006-06-10T15:24:19  <Kepplar> as i cant hold it as an int because of the leading 0's
2006-06-10T15:24:25  <Kepplar> nor can i incriment it as a string?
2006-06-10T15:24:50  <ThomasWaldmann> fpletz: there's a break missing X)
2006-06-10T15:25:03  <fpletz> ThomasWaldmann: in my code? ;)
2006-06-10T15:25:16  <ThomasWaldmann> no, in mine :)
2006-06-10T15:25:29  <fpletz> ;)
2006-06-10T15:25:29  <ThomasWaldmann> but you may fix it :)
2006-06-10T15:25:41  <ThomasWaldmann>         for modulename in mt.module_name():
2006-06-10T15:25:41  <ThomasWaldmann>             try:
2006-06-10T15:25:41  <ThomasWaldmann>                 execute = wikiutil.importPlugin(request.cfg, 'filter', modulename)
2006-06-10T15:25:54  <ThomasWaldmann>   ---> break
2006-06-10T15:26:04  <ThomasWaldmann> in Xapian.py
2006-06-10T15:27:32  <fpletz> ok
2006-06-10T15:29:13  <Kepplar> oh god
2006-06-10T15:29:22  <Kepplar> im forgotten my ssh key passphrase
2006-06-10T15:29:25  <Kepplar> ive
2006-06-10T15:29:48  <ThomasWaldmann> you should commit more often :)
2006-06-10T15:30:10  <fpletz> just pushed it to the repo ;)
2006-06-10T15:30:47  <Kepplar> i reset my box earlier
2006-06-10T15:30:53  <Kepplar> it caches the passphrase
2006-06-10T15:31:11  <Kepplar> tbh i forgot there was even one
2006-06-10T15:31:13  <Kepplar> i never use keys
2006-06-10T15:36:52  <ThomasWaldmann> submit a new key
2006-06-10T15:37:34  <ThomasWaldmann> fpletz: works :)
2006-06-10T15:37:57  <fpletz> ThomasWaldmann: yep, great ;)
2006-06-10T15:38:37  <ThomasWaldmann> I'll do some umlaut tests
2006-06-10T15:39:41  <fpletz> japanese characters worked on my box
2006-06-10T15:46:37  <fpletz> ThomasWaldmann: seems you built an index with --mode=add on an existing db
2006-06-10T15:46:42  <fpletz> if not, it's a bug ;)
2006-06-10T15:49:04  <fpletz> yep, I guess you did
2006-06-10T15:50:11  <ThomasWaldmann> oops
2006-06-10T15:51:00  * ThomasWaldmann recreates index
2006-06-10T15:53:28  <ThomasWaldmann> ok, better now
2006-06-10T16:02:50  <Kepplar> yea, will do
2006-06-10T16:03:01  <Kepplar> 1 more week till i move to cambridge :)
2006-06-10T16:05:39  <fpletz> what's in cambridge? :p
2006-06-10T16:05:52  <Kepplar> home :)
2006-06-10T16:06:16  <fpletz> oh.. I see ;)
2006-06-10T16:06:32  <Kepplar> one more week left here with my industrial placement you see
2006-06-10T16:06:56  <Kepplar> and its a free week so I can work on moin (already done most of the handovers \o/)
2006-06-10T16:07:07  <fpletz> ah, ok
2006-06-10T16:07:41  <fpletz> mmh, only 28 new summer-accepted mails.. is everyone still sleeping? ;)
2006-06-10T16:08:17  <Kepplar> heh, not been checking for a while
2006-06-10T16:08:23  <Kepplar> been working on code =)
2006-06-10T16:09:38  <mvirkkil> For future reference: http://siretart.tauware.de/dapper-mercurial/
2006-06-10T16:09:53  <mvirkkil> siretart built a package for dapper
2006-06-10T16:12:17  <xorAxAx> its really a shame that its not in the hotfixes suite yet
2006-06-10T16:17:17  <mvirkkil> hotfixes suite?
2006-06-10T16:23:32  <xorAxAx> dapper-updates
2006-06-10T16:30:49  <mvirkkil> xorAxAx: I'm just creating a bug re that on behalf of siretart
2006-06-10T16:32:15  <mvirkkil> xorAxAx: So it will probably be available through there too.
2006-06-10T16:40:34  <ThomasWaldmann> fpletz: I would like to merge xapian stuff into 1.6-main, ok?
2006-06-10T16:41:17  <fpletz> ThomasWaldmann: ok, go ahead
2006-06-10T16:41:44  <ThomasWaldmann> (and after that, remove _ magic)
2006-06-10T16:42:08  <xorAxAx> ThomasWaldmann: should i merge my changes as well?
2006-06-10T16:42:46  <ThomasWaldmann> no need to hurry.
2006-06-10T16:42:57  <ThomasWaldmann> i just want to get rid of lupy stuff :)
2006-06-10T16:43:09  <xorAxAx> and  i want to have the conflict changes tested :)
2006-06-10T16:43:14  <ThomasWaldmann> and the _ i want to change is also in xapian code
2006-06-10T16:43:39  <ThomasWaldmann> not too much at once
2006-06-10T16:45:16  <xorAxAx> ugh, i already get ugly conflicts
2006-06-10T16:45:32  <xorAxAx> conflict changes are unrelated to xapian things
2006-06-10T16:46:11  <xorAxAx> pushing
2006-06-10T16:49:27  * ThomasWaldmann slaps xorAxAx 
2006-06-10T16:49:50  <fpletz> lol
2006-06-10T16:50:11  * xorAxAx slaps ThomasWaldmann :-p
2006-06-10T16:50:41  <xorAxAx> they hadnt been merged for weeks
2006-06-10T16:51:00  <mvirkkil> The dapper package of hg DOESN'T WORK.
2006-06-10T16:51:09  <mvirkkil> so don't go installing it.
2006-06-10T16:51:31  <xorAxAx> you mean that weird backport?
2006-06-10T16:51:44  <mvirkkil> xorAxAx: yeah, that one.
2006-06-10T16:51:55  <mvirkkil> hmm.. should of been clearer.
2006-06-10T16:52:02  <xorAxAx> ok
2006-06-10T16:52:52  <mvirkkil> hould have, even..
2006-06-10T16:53:20  <xorAxAx> shehehehe
2006-06-10T17:22:58  <ThomasWaldmann> <a class="interwiki" href="http://moinmoin.wikiwikiweb.de/Übler%20Dübel"
2006-06-10T17:24:53  <ThomasWaldmann> (the %20 came from wiki markup, too)
2006-06-10T17:27:04  <ThomasWaldmann> hmm, looks like we have no interwiki markup with blanks in target pagename possible
2006-06-10T17:45:46  <mvirkkil> I've been thinkin about implementing a formatter which would generate a docbook, collect all images, create a zip and serve that. Thoughts?
2006-06-10T17:54:10  <xorAxAx> i would make it an action
2006-06-10T18:06:31  <mvirkkil> Ahh... good point :)
2006-06-10T20:00:07  <ThomasWaldmann> re
2006-06-10T21:23:05  <starshine> um I thought a formatter would be for the browser to present, a "packaged" but presentable page would be a pdf, no?
2006-06-10T22:13:16  <ThomasWaldmann>     q_string = ur"((?P<quote>['\"])[^(?P=quote)]*(?P=quote))"
2006-06-10T22:13:24  <ThomasWaldmann> isn't that correct?
2006-06-10T22:13:50  <ThomasWaldmann> for match "some quoted string" or 'some other quoted string'
2006-06-10T22:27:51  <ThomasWaldmann> solved in another way
2006-06-10T23:10:52  <mvirkkil> Ok, so I have my basic formatter that works just great, with the exception that since it's meant to convert single pages, it generates articles.
2006-06-10T23:11:45  <mvirkkil> And I'd need to be able to create books too. This is no problem, but I can't figure out how to switch between the two.
2006-06-10T23:12:16  <mvirkkil> I've been thinkin about creating another formatter, by subclassing the first, but I'm not sure if that's possible.
2006-06-10T23:12:26  <mvirkkil> Since all formatters are called Formatter.
2006-06-10T23:13:07  <mvirkkil> Another idea is to make it a processing instruction, and add that to Page.py
2006-06-10T23:13:35  <mvirkkil> But then we'd have an artifical limitation that a page can only be formatted either as a book or an article.
2006-06-10T23:14:17  <mvirkkil> The only solution I've found is to create DocBookFormatter class, and then subclass that for both formatters, but then I wouldn
2006-06-10T23:14:34  <mvirkkil> wouldn't be sure where to put that DocBookFormatter class.
2006-06-10T23:14:47  <mvirkkil> Any ideas, hints or solutions?