MoinMoin: MoinMoinChat/Logs/moin-dev/2008-06-17

2008-06-17T08:03:29  <ThomasWaldmann> moin
2008-06-17T08:15:25  <TheSheep> moin
2008-06-17T08:20:53  <dreimark> moin
2008-06-17T08:37:45  <dreimark> doodle is to complicated for me, how can I change a date now?
2008-06-17T08:38:23  <dreimark> ah ok, got it
2008-06-17T09:04:31  <xorAxAx> "change a date"?
2008-06-17T09:05:09  <xorAxAx> ah, modify afterwards, ok
2008-06-17T09:05:17  <dreimark> well I entered the name first and needed to find afterwards the modify button
2008-06-17T09:26:02  <CIA-64> Johannes Berg <johannes AT sipsolutions DOT net> default * 3708:0ab086afb8b6 1.7/MoinMoin/userprefs/oidserv.py: add openid server settings dialog to allow removing trusted websites
2008-06-17T09:50:12  <ThomasWaldmann> last minute changes :)
2008-06-17T09:50:45  <johill> well it's a new userprefs plugin so can hardly cause regressions :)
2008-06-17T09:51:10  <johill> so I figured I'd just add it there as it's useful for the openid server code
2008-06-17T09:52:12  <johill> good thing I did the userprefs plugin system, that's so much nicer
2008-06-17T09:52:49  <ThomasWaldmann> yeah :)
2008-06-17T09:54:29  <ThomasWaldmann> you store the trusted roots as base64?
2008-06-17T10:13:16  <xorAxAx> johill: hi, did you see my msg?
2008-06-17T10:34:14  <johill> xorAxAx: your msg? no, doesn't look like it
2008-06-17T10:34:37  <xorAxAx> johill: sigh, doesnt your client store highlights?
2008-06-17T10:34:38  <johill> ThomasWaldmann: yeah, because I didn't want to rely on rogue RPs not sending bad characters
2008-06-17T10:34:54  <johill> it actually sends them to me by email
2008-06-17T10:35:13  <johill> was it a query though?
2008-06-17T10:35:20  <xorAxAx> no, in this channel
2008-06-17T10:35:57  <johill> odd, when?
2008-06-17T10:37:40  <xorAxAx> yesterday evening
2008-06-17T10:37:48  <xorAxAx> 23:31 CEST
2008-06-17T10:37:53  <johill> oh that
2008-06-17T10:38:01  <johill> somehow I thought that came from thomas
2008-06-17T10:38:42  <johill> I did indeed see that
2008-06-17T10:40:24  <johill> except I didn't see the rest of what you typed and hence had no clue what you were talking about
2008-06-17T10:40:35  <johill> you should start to get used to the fact that irc is a lossy medium and just add stuff to the wiki instead
2008-06-17T10:40:47  <johill> oh wait, of course! chatlogs! how could I be that stupid ;)
2008-06-17T10:45:53  <xorAxAx> johill: well, you didnt need the context and obviously it wasnt lossy in this case
2008-06-17T10:46:23  <xorAxAx> but you ignored the message. could have happened witha wiki edit or email as well :)
2008-06-17T10:47:19  <johill> no, I actually read the email I got about it but decided to ignore it because it only said "please add your dates" and not even where to
2008-06-17T10:47:28  <xorAxAx> umm
2008-06-17T10:47:44  <xorAxAx> indeed, should have put the url into the msg
2008-06-17T10:48:15  <johill> so I figured I'd either get a wiki changeset by email, or had no idea what to do anyway, and forgot about the message right away
2008-06-17T10:48:23  <johill> :)
2008-06-17T10:49:02  <johill> just saying. I didn't go back and read the irc from around that time
2008-06-17T10:50:25  * xorAxAx -> uni
2008-06-17T12:39:06  <dreimark> gizmach: ping
2008-06-17T12:48:35  <dreimark> ThomasWaldmann: I do have a failing test in test_packages tests/test_packages.py[6] ..F.s.
2008-06-17T12:48:57  <ThomasWaldmann> I don't :)
2008-06-17T12:52:36  * dreimark checks on the ramdisk
2008-06-17T12:52:55  <dreimark> I don't have it there too
2008-06-17T12:59:42  <ThomasWaldmann> (I don't use a ramdisk btw)
2008-06-17T13:04:04  * dreimark tries a fresh testwiki
2008-06-17T13:06:19  <dreimark> ThomasWaldmann: no problem in a fresh test wiki
2008-06-17T14:58:51  <gizmach> dreimark: here I am, was at college
2008-06-17T15:00:33  <xorAxAx> gizmach: did you read my msg already?
2008-06-17T15:01:29  <gizmach> xorAxAx: not yet
2008-06-17T15:01:30  <gizmach> sec
2008-06-17T15:02:02  <gizmach> ah I see
2008-06-17T15:04:06  <gizmach> what about the days I'm not sure
2008-06-17T15:14:19  <gizmach> added, but for some I'm not sure
2008-06-17T15:21:36  <TheSheep> you can't plan for somehting you're not sure of, right?
2008-06-17T15:22:20  <gizmach> I didn't add OK if I'm not sure
2008-06-17T15:22:36  <xorAxAx> hmm,why arent you sure? too busy next to soc?
2008-06-17T15:22:57  <dennda> Maybe she has a very unhealthy sleeping rhythm
2008-06-17T15:23:18  <xorAxAx> alarm clocks exist :)
2008-06-17T15:23:20  <gizmach> xorAxAx: no, just dono if I will have to go to the college for an hour
2008-06-17T15:23:23  <gizmach> xorAxAx: :)
2008-06-17T15:23:28  <gizmach> no this is ok
2008-06-17T15:23:45  <gizmach> but probably I can be here for some days all day
2008-06-17T15:25:01  <xorAxAx> and i think we will rather "override" the students red boxes than the mentors' ones
2008-06-17T15:27:48  <dennda> xorAxAx: They fail for me
2008-06-17T15:28:00  <xorAxAx> dennda: hmm? :)
2008-06-17T15:28:39  <gizmach> sorry for mentioning this it got a bad content
2008-06-17T15:28:51  <gizmach> everything is ok to me
2008-06-17T15:29:13  <dennda> xorAxAx: I always have two alarm clocks set and push them under the bed as far as I can. Sounds like a good plan, except it doesn't work
2008-06-17T15:30:11  <TheSheep> just getting up as soon as you hear an alarm clock usually works
2008-06-17T15:30:20  <TheSheep> and making your bed rigth away
2008-06-17T15:30:22  <xorAxAx> dennda: then they are too near
2008-06-17T15:31:15  <dennda> xorAxAx: Room space is limited. And even if it weren't, the further you put these nasty devices away the less you are affected by the signals they emit
2008-06-17T15:31:59  <dennda> johill: rather # {id : {revision_id : (revision_data, {revision_metadata})}}
2008-06-17T15:32:02  <xorAxAx> dennda: then they are not powerful enough
2008-06-17T15:32:23  <dennda> xorAxAx: Yes, I need one made out of steel, non-interruptible and emitting a terrific noise
2008-06-17T15:32:31  <dennda> (and electric shocks)
2008-06-17T15:32:40  <xorAxAx> yes. let your neighbours know when your day begins
2008-06-17T15:35:45  <dennda> See? It isn't easy to find a good solution
2008-06-17T15:38:38  * dennda works on the MemoryBackend...
2008-06-17T15:40:20  <gizmach> bbl shower
2008-06-17T15:42:29  * dreimark removed some spam from MM
2008-06-17T15:44:11  <dreimark> xorAxAx: do you have an idea how to add unit tests to the data/plugins paths
2008-06-17T15:50:09  <dreimark> the same test performs well called from MoinMoin/parser/_tests and from data/plugins/parser/_test I always get AccessDenied: You are not allowed to delete this page!
2008-06-17T15:59:17  <dennda> johill: How are we supposed to load Item metadata lazily if __getitem__ and __setitem__ are not to be overridden by the backend implementor?
2008-06-17T15:59:58  <dennda> Sure, I can create an Item and pass the metadata directly to that items _metadata attribute. I actually even did that, but then noticed it's not very lazy, is it?
2008-06-17T16:00:54  <xorAxAx> dreimark: hmm, can you rephrase your question?
2008-06-17T16:01:27  <dennda> With revisions there is no such problem (as far as I can tell atm)
2008-06-17T16:39:54  <dreimark> xorAxAx: I do wonder why if I copy for example test_Hits from MM/macro/_tests to wiki/data/plugins/macro/_tests
2008-06-17T16:40:20  <dreimark> that it fails because of Access denied in nuke_page
2008-06-17T16:40:46  <xorAxAx> dreimark: the conftest file is in MoinMoin/ so it doesnt work for wiki/*
2008-06-17T16:42:01  <dreimark> it does work for a copy of test_CSV and others if py.test ist started from the toplevel dir
2008-06-17T16:42:17  <dreimark> and it does check jabber* too
2008-06-17T16:43:30  <dreimark> it does not work because of wrong import dependencies if it is called in a data/plugins/_tests dir
2008-06-17T16:44:17  <xorAxAx> waldi: ping
2008-06-17T16:44:24  <xorAxAx> waldi: http://www.doodle.ch/participation.html?pollId=6yeqmzb4v823mqqz
2008-06-17T16:45:06  <xorAxAx> zenhase: http://www.doodle.ch/participation.html?pollId=6yeqmzb4v823mqqz
2008-06-17T16:45:17  <xorAxAx> together with my student, you are the remaining persons
2008-06-17T16:45:30  <xorAxAx> given that this is a full time job, you seem to have large delays
2008-06-17T16:45:56  <xorAxAx> dreimark: "input dependencies"?
2008-06-17T16:46:07  <dreimark> hmm, may be the test is not running in the testwiki, will check that in some mins.
2008-06-17T16:46:28  <xorAxAx> i wonder how it works for you at all without conftest
2008-06-17T16:46:36  <dreimark> xorAxAx: I meant "import" from MoinMoin ...
2008-06-17T16:46:44  <xorAxAx> dreimark: thats called python path
2008-06-17T16:46:53  <xorAxAx> probably you create instantiate the request class manually
2008-06-17T16:46:58  <xorAxAx> but that would be the first mistake
2008-06-17T16:47:35  <dreimark> not me it is probably done by conftest / py.test
2008-06-17T16:47:36  <xorAxAx> s/create//
2008-06-17T16:47:44  <xorAxAx> yes, but conftest doesnt work in wiki/*
2008-06-17T16:48:08  <xorAxAx> only in the subdirectories of the dir where it is
2008-06-17T16:49:04  <dreimark> hmm, then it doesn't work in jabber* too but tests from there are tried too
2008-06-17T16:49:59  <dreimark> bbl ~5mins
2008-06-17T16:53:38  <CIA-64> Christopher Denter <moin GUESSWHAT the DASH space DASH station PERIOD com> default * 4028:161eb6f9edea 1.8-storage-cdenter/MoinMoin/storage/ (__init__.py backends/memory.py): committing first UNDER CONSTRUCTION snapshot of the memorybackend. (this is a heavy commit, sorry)
2008-06-17T16:55:09  <dennda> johill: comments appreciated :)
2008-06-17T16:57:26  <dreimark> backends right written :)
2008-06-17T16:58:39  <dennda> dreimark: hum? :)
2008-06-17T17:00:03  <dreimark> I often read the last times beckends, 6:40 < ThomasWaldmann> gizmach: someone might kill you if you continue talking about beckends
2008-06-17T17:00:33  <dreimark> 16:40
2008-06-17T17:00:39  <dennda> hehe :)
2008-06-17T17:01:01  <dennda> well, it's the end of the back...
2008-06-17T17:01:04  <dennda> ok let's just move along
2008-06-17T17:01:08  <dennda> :)
2008-06-17T17:02:41  <dreimark> xorAxAx: e.g. jabberbot/_tests/test_xmppbot.py
2008-06-17T17:02:51  <xorAxAx> dreimark: whats about that?
2008-06-17T17:03:18  <dreimark> why can that be tested if conftest is in MoinMoin
2008-06-17T17:05:47  <xorAxAx> because its unrelated to moin's conftest
2008-06-17T17:06:00  <xorAxAx> jabberbot is on the uppermost level because its completly decoupled
2008-06-17T17:06:41  <dreimark> ok
2008-06-17T17:11:50  <gizmach> re
2008-06-17T17:29:30  <johill> dennda: we have something like backend._get_item_metadata(), no?
2008-06-17T17:35:05  <dennda> johill: I will answer that question in a second, but I just got struck by something else: You once sent me some mockup code on locking metadata, and there you check value to be of type str, tuple or unicode. I thought values can be arbitrary python objects?
2008-06-17T17:36:22  <dennda> (e.g. ints)
2008-06-17T17:37:24  <johill> we discussed that already ;)
2008-06-17T17:37:40  <johill> or was that with xorAxAx?
2008-06-17T17:37:44  <johill> he didn't like the implied pickling
2008-06-17T17:37:44  <xorAxAx> :-)
2008-06-17T17:37:47  <dennda> can't remember
2008-06-17T17:38:02  <johill> i.e. if you can assign arbitrary stuff, the backend needs to pickle objects
2008-06-17T17:38:09  <dennda> sure
2008-06-17T17:38:22  <johill> xorAxAx now says this sucks for indexing, a point I haven't fully understood
2008-06-17T17:38:37  <dennda> otoh I don't want to do str(myint), int(mystr_which_used_to_be_an_int) all the time
2008-06-17T17:38:38  <johill> in my opinion, the backend, if it wants to index, has a better chance if it knows the type of the value
2008-06-17T17:39:22  <xorAxAx> johill: no, we concluded that the problem is orthogonal
2008-06-17T17:39:34  <xorAxAx> and that you both need to think about a way to facilitate bridging
2008-06-17T17:39:39  <dennda> xorAxAx: could you please explain what you mean by saying orthogonal?
2008-06-17T17:39:40  <xorAxAx> s/bridging/indexing/
2008-06-17T17:39:55  <xorAxAx> dennda: that the problem of indexing and types is there regardless whether its pickled or not
2008-06-17T17:40:03  <xorAxAx> on this level
2008-06-17T17:41:46  <johill> well yeah, you will not be able to index by arbitrary objects
2008-06-17T17:41:59  <johill> not even python supports that (ie. what can be used as dict keys)
2008-06-17T17:42:49  <johill> xorAxAx: I think I don't remember a real conclusion
2008-06-17T17:43:06  <xorAxAx> johill: we concluded that this must be thought about
2008-06-17T17:43:10  <xorAxAx> considering indexing first
2008-06-17T17:43:17  <xorAxAx> maybe we concluded uniliterally
2008-06-17T17:43:18  <xorAxAx> :-)
2008-06-17T17:43:23  <johill> yeah, indexing we have to think about
2008-06-17T17:43:24  <johill> heh
2008-06-17T17:43:27  <dennda> restricting this to str, tuple and unicode just caused a bug in my MemoryBackend which is why I asked in the first place since I was under the impression that we wanted to allow arbitrary objects when I wrote the corresponding parts of the MB
2008-06-17T17:43:58  <johill> restricting it to immutable types does have advantages though
2008-06-17T17:44:05  <dennda> could one of you define what you mean with indexing so we have the same term in mind?
2008-06-17T17:44:14  <johill> it means that people cannot alter a list in a read-only metadata dict
2008-06-17T17:44:22  <johill> s/a list/an object/
2008-06-17T17:44:24  <dennda> ints are immutable O:-)
2008-06-17T17:44:47  <johill> yeah
2008-06-17T17:45:21  <johill> but many objects are not, so allowing arbitrary objects might lead to hard-to-debug bugs when people don't fully understand the immutability properties
2008-06-17T17:45:38  <johill> (yeah, I know, I always assume the worst in people using an API)
2008-06-17T17:45:40  <dennda> does any of you disagree if I just add int to the list of allowed types?
2008-06-17T17:46:25  <johill> no, but then you should probably add long and various other types too
2008-06-17T17:46:25  <dennda> a goddamn it, missed my bus
2008-06-17T17:46:38  <johill> wrt. indexing
2008-06-17T17:46:55  <johill> the user backend for example has itemname=12345.4321.123
2008-06-17T17:47:06  <johill> and item['name']='JohannesBerg'
2008-06-17T17:47:08  <dennda> ok what have we got there? str, tuple, unicode, int, long, bool
2008-06-17T17:47:15  <dennda> float
2008-06-17T17:47:18  <johill> float double complex?
2008-06-17T17:47:35  <dennda> I wonder if anybody needs complex numbers as metadata keys
2008-06-17T17:47:44  <dennda> same goes for float and double
2008-06-17T17:47:47  <johill> heh
2008-06-17T17:47:52  <dennda> and bool
2008-06-17T17:48:19  <johill> dunno
2008-06-17T17:48:33  <dennda> me neither
2008-06-17T17:48:37  <dennda> but when in doubt, leave it out
2008-06-17T17:48:41  <johill> let's go back to indexing, I want to go again soon
2008-06-17T17:48:54  <gizmach> ThomasWaldmann : if I put into my local config that I use ldap for auth, MM is authenticated to ldap server, but in ldap_groups.py in wikidicts i need to use some ldap information (ldap object, bind_dn,...)
2008-06-17T17:49:04  <johill> yeah but adding the type doesn't hurt and if somebody later needs it in a plugin it gets complicated
2008-06-17T17:49:26  <johill> I also forgot tuples containing tuples, so you'd have to check the type recursively
2008-06-17T17:49:35  <dennda> I hold the bet that that is never going to happen
2008-06-17T17:49:37  <gizmach> do I need to import ldap_login again
2008-06-17T17:49:40  <dennda> why do we need tuples?
2008-06-17T17:49:45  <gizmach> or dreimark
2008-06-17T17:50:12  <ThomasWaldmann> gizmach: no
2008-06-17T17:50:53  <johill> dennda: tuples? instead of lists, to list up say openid_trusted_roots
2008-06-17T17:51:09  <ThomasWaldmann> gizmach: i suggest if user x logs in via ldap, then ldap auth queries also all group where x is member and stores them to moin cache (see caching module)
2008-06-17T17:51:26  <dennda> johill: nevermind, I was thinking why we'd need them as keys, not as value
2008-06-17T17:51:38  <dennda> then I do see a point for float, double, bool and complex
2008-06-17T17:51:39  <dennda> ok
2008-06-17T17:51:43  <dennda> I'll check recursively
2008-06-17T17:51:50  <johill> ah, no, keys are just strings I think
2008-06-17T17:51:52  <gizmach> ThomasWaldmann: ok
2008-06-17T17:51:57  <dennda> yes
2008-06-17T17:51:58  <dennda> they are
2008-06-17T17:52:04  <johill> ok
2008-06-17T17:52:06  <dennda> getting tired :)
2008-06-17T17:52:09  <johill> so about indexing (try #2)
2008-06-17T17:52:17  <dennda> #2?
2008-06-17T17:52:30  <johill> we need to be able to look up based on a metadata property
2008-06-17T17:52:36  <johill> say to find the user item based on the 'name' metadata
2008-06-17T17:53:00  <gizmach> Then ldap_groups is practically not needed if I store groups while logging in
2008-06-17T17:53:35  <dennda> ah
2008-06-17T17:53:51  <dennda> ok, stupid approach: iterate over all useritems and check if the name matches
2008-06-17T17:54:14  <johill> yeah, way too slow
2008-06-17T17:54:22  <johill> hence, we need to build an index
2008-06-17T17:55:01  <johill> optionally, of course, since sql might do that at the db level; it should be a backend that defers everything but searching to a backend below it
2008-06-17T17:58:25  <ThomasWaldmann> gizmach: you need some ldap groups code to check that cache you have built
2008-06-17T17:58:46  <gizmach> ThomasWaldmann: well true
2008-06-17T17:59:06  <gizmach> then lots of coffie and not sleeping night
2008-06-17T17:59:11  <dennda> johill: the type double doesn't even exist, does it?
2008-06-17T17:59:25  <dennda> :)
2008-06-17T17:59:29  <johill> dennda: dunno, might not :)
2008-06-17T17:59:36  <johill> xorAxAx: and actually, here it shows that it isn't orthogonal
2008-06-17T18:00:08  <johill> xorAxAx: a user can have many openIDs, but it would be good to be able to search by any of them, and index for that
2008-06-17T18:00:23  <xorAxAx> johill: how is that related to pickling? :)
2008-06-17T18:00:46  <johill> well if the front-end that stores the openIDs pickles, the backend that does the indexing has to unpickle them again
2008-06-17T18:01:26  <johill> hmmm
2008-06-17T18:01:43  <johill> it seems we need to allow the set type too to allow the backend to lose the order?
2008-06-17T18:02:14  <johill> I'll think about it some more while I skate to the store :)
2008-06-17T18:02:49  <dennda> don't b0rk yourself
2008-06-17T18:08:45  <dennda> http://paste.pocoo.org/show/72074/
2008-06-17T18:08:46  <dennda> ok?
2008-06-17T18:21:46  <dreimark> gizmach: do_user_browser may need it too
2008-06-17T18:22:15  <gizmach> dreimark: huh?
2008-06-17T18:22:48  <dreimark> there are some places in the code where we currently do
2008-06-17T18:22:55  <dreimark> isgroup = request.cfg.cache.page_group_regexact.search
2008-06-17T18:23:12  <dreimark> groupnames = request.rootpage.getPageList(user='', filter=isgroup)
2008-06-17T18:24:00  <dreimark> gizmach: additional to the answers you got for "Then ldap_groups is practically not needed"
2008-06-17T18:24:17  <gizmach> ah I see
2008-06-17T18:24:33  <gizmach> I didn't know for what is it
2008-06-17T18:24:45  * dreimark was late
2008-06-17T18:24:50  <gizmach> dreimark: ok I will leave it then
2008-06-17T18:25:28  * gizmach is first adding user groups to ldap auth
2008-06-17T18:35:41  <dreimark> gizmach: please do some commits today
2008-06-17T18:35:56  <gizmach> dreimark: will do
2008-06-17T18:36:27  <gizmach> just finishing the test to see if it works then will add that to ldap auth if it's ok and commit
2008-06-17T18:41:33  <CIA-64> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3709:aae5351b8a1e 1.7/ (MoinMoin/request/__init__.py docs/CHANGES): allow trusted users to go beyond surge action limits, but log their names if it happens
2008-06-17T18:42:26  <johill> dennda: yeah looks right
2008-06-17T18:42:40  <dennda> just fixing a few bugs
2008-06-17T18:42:48  <dennda> and then write tests
2008-06-17T18:43:04  <dennda> don't look at the MB yet, please
2008-06-17T18:43:23  <dennda> there are a few things that I need to iron out and you'd waste your time telling me that things I already fixed are wrong :)
2008-06-17T18:45:49  <johill> :)
2008-06-17T19:01:38  <dreimark> johill: should be anonymous_session_lifetime mentioned in CHANGES because of killing anonymous_session ?
2008-06-17T19:04:21  <CIA-64> Christopher Denter <moin GUESSWHAT the DASH space DASH station PERIOD com> default * 4029:3ab3ae1d7336 1.8-storage-cdenter/MoinMoin/storage/ (__init__.py backends/memory.py):
2008-06-17T19:04:21  <CIA-64> fixed quite a bunch of bugs in both, the abstract backend and the specific
2008-06-17T19:04:21  <CIA-64> memorybackend. also changed the types allowed for metadata-values. you may now
2008-06-17T19:04:21  <CIA-64> use any standard immutable python type known to mankind, including tuples and
2008-06-17T19:04:23  <CIA-64> recursively.
2008-06-17T19:04:47  <dennda> johill: you can review it now if you like
2008-06-17T19:05:17  <dennda> Quite a few things already work, including: getting and creating items and revisions (but not committing items), metadata, etc.
2008-06-17T19:05:23  <dennda> now is a good time to write some tests
2008-06-17T19:06:46  <johill> ok I'll take a look after I reviewed these wireless patches for 802.11w
2008-06-17T19:10:08  <dennda> ok
2008-06-17T19:10:16  <dennda> no need to hurry
2008-06-17T19:10:22  <dennda> my brain is somewhat dead anyway now
2008-06-17T19:12:24  <johill> dennda: Item._lock shouldn't have an underscore but be calede lock_metadata, no?
2008-06-17T19:14:07  <johill> dennda: you misunderstood what a search term is meant to be
2008-06-17T19:14:50  <johill> dennda: look at MoinMoin/search/term.py (but feel free to put low on your todo list right now)
2008-06-17T19:14:53  <dennda> well, yes, but a) what else do you want to lock on an item? afair we concluded that committing revisions on an item doesn't need locks, and b) you wrote that and I just lazily copy-pasted it :)
2008-06-17T19:16:04  <johill> heh ok then I was probably drugged or asleep ;)
2008-06-17T19:16:19  <johill> underscore methods are internal so that can't be with an underscore
2008-06-17T19:17:09  <johill> "Itemnames must be of type str,"
2008-06-17T19:17:15  <johill> they can also be unicode, of course
2008-06-17T19:18:03  <johill> dennda: can we put that into gobby for a minute?
2008-06-17T19:18:31  <dennda> sure
2008-06-17T19:18:37  <dennda> dunno if the session still runs
2008-06-17T19:18:44  <johill> doesn't look like
2008-06-17T19:18:52  <dennda> ok, just a sec
2008-06-17T19:19:24  <dennda> now
2008-06-17T19:33:38  * ThomasWaldmann tries xapian 1.0.6
2008-06-17T19:43:53  <gizmach> re
2008-06-17T19:44:30  * ThomasWaldmann .oO(works)
2008-06-17T19:47:36  * gizmach done a test to get all groups certain user is in it and it passed
2008-06-17T20:03:30  <CIA-64> Christopher Denter <moin GUESSWHAT the DASH space DASH station PERIOD com> default * 4030:228036c97610 1.8-storage-cdenter/MoinMoin/storage/ (__init__.py backends/memory.py): committing a bunch of refactorings that johill pointed out to me, thanks! a few other things in there are still going to be refactored
2008-06-17T20:12:17  <ThomasWaldmann> dennda: your commit comments are not understandable out of context
2008-06-17T20:13:10  <dennda> ThomasWaldmann: you mean "in there"?
2008-06-17T20:13:25  <dennda> Or at all?
2008-06-17T20:13:48  <ThomasWaldmann> imagine we merge that into main branch
2008-06-17T20:15:05  <dennda> ok I will try to be more precise
2008-06-17T20:25:49  <dreimark> TheSheep: I completly missed encoding of attachments in arnica
2008-06-17T20:26:37  <dreimark> will do some refactoring in the next days related to that
2008-06-17T20:31:05  <dreimark> ThomasWaldmann: hmm I think I do want a logging.error msg if in case of OSError in caching.remove. I do think pass is too silent.
2008-06-17T20:36:02  <johill> dreimark: make sure to check for "-ENOENT" then though (i.e. file is already gone)
2008-06-17T20:41:27  <dreimark> johill: ENONET = "Machine is not on the network" http://docs.python.org/lib/module-errno.html
2008-06-17T20:42:09  <ThomasWaldmann> ent not net
2008-06-17T20:48:24  <dreimark> oh yeah
2008-06-17T20:49:00  <johill> yeah, even though my typing is lazy at times, that was intentional ;)
2008-06-17T20:49:18  <johill> see, I meant to type "even though my typing is lousy at times"
2008-06-17T20:49:38  <johill> EBRAINTOOFASTFORFINGERS
2008-06-17T20:55:05  <dreimark> bbl
2008-06-17T22:28:15  <gizmach> any ideas how to solve this http://paste.pocoo.org/show/72204/
2008-06-17T23:11:26  <zenhase> re
2008-06-17T23:11:39  <zenhase> xorAxAx: sorry, missed that one :o
2008-06-17T23:11:48  <zenhase> i added myself
2008-06-17T23:12:16  <xorAxAx> monday and tuesday are completly non-soc?
2008-06-17T23:12:36  <zenhase> uhm no
2008-06-17T23:12:40  <zenhase> they are in the past
2008-06-17T23:12:47  <xorAxAx> ?
2008-06-17T23:12:52  <gizmach> the date
2008-06-17T23:12:53  <xorAxAx> this is about a weekly meeting
2008-06-17T23:12:56  <gizmach> I figure that too
2008-06-17T23:12:57  <zenhase> ah ok
2008-06-17T23:13:00  <gizmach> this week
2008-06-17T23:13:02  <zenhase> :)
2008-06-17T23:13:06  <gizmach> I also looked at the date
2008-06-17T23:13:06  <zenhase> then i i'll change that
2008-06-17T23:13:18  <xorAxAx> "The dates are only exemplarily and depict the weekdays."
2008-06-17T23:14:33  <zenhase> now
2008-06-17T23:14:35  <xorAxAx> mitsuhiko: can you check that as well?
2008-06-17T23:14:57  <gizmach> I will change it
2008-06-17T23:15:02  <mitsuhiko> xorAxAx: link pls
2008-06-17T23:15:05  <zenhase> only fixed dates i have right now is tuesday and thursday, at 15:45-17:15
2008-06-17T23:15:08  <mitsuhiko> haven't bookmarked it
2008-06-17T23:15:43  <xorAxAx> mitsuhiko: http://www.doodle.ch/participation.html?pollId=6yeqmzb4v823mqqz
2008-06-17T23:16:42  <mitsuhiko> oh well
2008-06-17T23:16:52  <zenhase> also i'll try to shift my sleep/work/fun-cycle again, so i get up at 8-9am ... then an extra hour will be added in front of each day ;)
2008-06-17T23:17:17  <zenhase> right now it got kind of shifted towards getting up at half past 10 :o
2008-06-17T23:18:41  <xorAxAx> mitsuhiko: i.e.?
2008-06-17T23:18:53  * gizmach gets at 4 am and that sux
2008-06-17T23:18:55  <xorAxAx> mitsuhiko: are monday/tuesday correct in your schedule?
2008-06-17T23:19:33  <mitsuhiko> monday yes, tuesday i could probably manage to be online 4 PM
2008-06-17T23:19:36  <mitsuhiko> but i can't promise that
2008-06-17T23:19:57  <xorAxAx> ok, i thought you used the same interpretation as zenhase and gizmach :)
2008-06-17T23:20:06  <xorAxAx> then its fine
2008-06-17T23:20:15  <zenhase> :)
2008-06-17T23:21:13  * gizmach changed but she can fit probably in any time
2008-06-17T23:21:58  <dreimark> re
2008-06-17T23:22:15  <dreimark> gizmach: do you have fixed your merge problem ?
2008-06-17T23:22:29  <gizmach> I think I did
2008-06-17T23:22:43  <dreimark> ok
2008-06-17T23:22:45  * gizmach will commit in a half a minute
2008-06-17T23:22:51  <gizmach> we'll see
2008-06-17T23:22:51  <dreimark> very good
2008-06-17T23:23:09  <gizmach> if ThomasWaldmann comes and killes me than I done something wrong
2008-06-17T23:23:29  <dreimark> use  your branch
2008-06-17T23:24:30  <xorAxAx> repository you mean
2008-06-17T23:24:41  <gizmach> yes
2008-06-17T23:24:49  <gizmach> will check it
2008-06-17T23:25:16  <gizmach> it's mine
2008-06-17T23:27:12  <dreimark> xorAxAx: right
2008-06-17T23:27:22  * gizmach crosses her fingers
2008-06-17T23:27:39  <xorAxAx> engineering with hope is a bad approach :-P
2008-06-17T23:27:56  <gizmach> :P
2008-06-17T23:28:29  <CIA-64> MelitaMihaljevic default * 3675:6eb3de127225 1.8-ldapgroups-mmihaljevic/MoinMoin/_tests/ (ldap_testdata.py test_ldapgroups.py): test_ldapgroups.py: new test testUsersInGroup for getting all the groups that certain user is a member of, modified other tests and fix errors; ldap_testdata.py: added a Group record
2008-06-17T23:28:43  <gizmach> damit a storm is comming