MoinMoin: MoinMoinChat/Logs/moin-dev/2009-06-04

2009-06-04T00:00:27  <amartani> I don't know how "cancel" should work.. google docs just don't have any cancel, but you can revert to some revision..
2009-06-04T00:06:08  <amartani> ThomasWaldmann: about the except, I just thought that, if there is any problem on connecting to mobwrite server, it should return an empty answer, which doesn't do anything on the editor..
2009-06-04T00:07:22  *** grzywacz has quit IRC
2009-06-04T00:08:07  <ThomasWaldmann> amartani: except even catches SyntaxErrors iirc
2009-06-04T00:09:44  <amartani> ok, I'm changing it..
2009-06-04T00:09:56  <amartani> sorry..
2009-06-04T00:19:31  <amartani> dreimark: about meeting, on tuesday, I have classes until 10:30 pm (CEST)
2009-06-04T00:27:55  *** amartani has quit IRC
2009-06-04T00:28:22  *** amartani has joined #moin-dev
2009-06-04T00:29:11  *** dimazest has quit IRC
2009-06-04T00:38:57  <amartani> ThomasWaldmann: should I store who is editing some item on it's metadata?
2009-06-04T00:53:12  *** dimazest has joined #moin-dev
2009-06-04T01:00:01  <ThomasWaldmann> the more interesting question is "when do you save"?
2009-06-04T01:00:31  <ThomasWaldmann> and "who saves"
2009-06-04T01:05:50  <amartani> well, I was thinking about just add a button to save but continue editing, and perhaps an option to auto-save..
2009-06-04T01:08:32  <amartani> about who saves, if I have who was editing the doc, it could be a list of editors..
2009-06-04T01:11:36  <ThomasWaldmann> moin usually stores userid, ip and hostname of editor
2009-06-04T01:11:50  <ThomasWaldmann> but until now, there was only one of them
2009-06-04T01:12:43  <ThomasWaldmann> so I suggest you just use a CSV-style thing for now
2009-06-04T01:16:03  *** moinBot` has joined #moin-dev
2009-06-04T01:17:40  *** moinBot has quit IRC
2009-06-04T01:18:57  <amartani> ok, but how can I store the current editors? I was thinking about items metadata, but it looks like it only alters metadata on save...
2009-06-04T01:21:19  <ThomasWaldmann> yes, save creates a new data and metadata revision (in parallel)
2009-06-04T01:23:04  <ThomasWaldmann> amartani: you need the editor list in the javascript code anyway, for showing presence
2009-06-04T01:23:32  <ThomasWaldmann> so, the js could update the document's form data with the list of editors
2009-06-04T01:24:15  <ThomasWaldmann> and if a POST gets triggered, it also POSTs those hidden field(s)
2009-06-04T01:28:14  <amartani> ThomasWaldmann: but how would I get the list to send to the client? I have to store this somehow.. also, we can't trusting this data..
2009-06-04T01:29:03  <ThomasWaldmann> well, you only can have one client do the POST for one revision save, right?
2009-06-04T01:29:52  <ThomasWaldmann> of course only that data is trustworthy
2009-06-04T01:30:14  <ThomasWaldmann> the list of "other editors" could be faked, so it is rather informative
2009-06-04T01:31:53  <ThomasWaldmann> for now, you could also just use the one editor that does the post
2009-06-04T01:34:07  <amartani> yes, but I still need the editors list to show presence..
2009-06-04T01:40:08  <ThomasWaldmann> that's another thing
2009-06-04T01:41:12  <ThomasWaldmann> maybe have a look at the session cookie
2009-06-04T01:41:25  <ThomasWaldmann> not sure whether we store the user's name there
2009-06-04T01:50:09  <amartani> I could also open a fake "page" on mobwrite to store the editors..
2009-06-04T01:52:45  <ThomasWaldmann> and a chat window :)
2009-06-04T03:07:28  *** dimazest has quit IRC
2009-06-04T03:07:54  *** dimazest has joined #moin-dev
2009-06-04T03:08:32  *** dimazest has left #moin-dev
2009-06-04T03:54:14  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5890:e52c1a245a3a 2.0-storage-editor-amartani/MoinMoin/action/mobwrite.py: Real-time editor: Catch only socket error.
2009-06-04T03:54:17  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5891:526f56554947 2.0-storage-editor-amartani/MoinMoin/templates/modify_text.html: Real-time editor: Enable debug on mobwrite.
2009-06-04T03:54:19  <CIA-19> Alexandre Martani <amartani AT gmail DOT com> default * 5892:20ef89d32ed7 2.0-storage-editor-amartani/MoinMoin/action/ (_tests/test_mobwrite.py mobwrite.py): Real-time editor: Add nullify method
2009-06-04T06:47:00  *** rkerr has quit IRC
2009-06-04T08:28:55  <ThomasWaldmann> moin
2009-06-04T08:32:14  *** grzywacz has joined #moin-dev
2009-06-04T09:14:14  *** grzywacz has quit IRC
2009-06-04T09:17:06  <dreimark> moin
2009-06-04T09:31:54  *** dimazest has joined #moin-dev
2009-06-04T09:32:10  <dimazest> moin
2009-06-04T10:36:51  *** amartani has quit IRC
2009-06-04T10:41:09  <ThomasWaldmann> dennda: the mimetype assert you added is at the wrong place
2009-06-04T10:43:54  <dennda> oh?
2009-06-04T10:48:51  <ThomasWaldmann> reading the line afterwards helps :)
2009-06-04T10:50:17  <dennda> you mean in case somebody stored meta[MIMETYPE] = None?
2009-06-04T10:50:39  <ThomasWaldmann> no, in case that is not None
2009-06-04T10:51:10  <ThomasWaldmann> you're panicking it too early :)
2009-06-04T10:52:40  <dennda> so you'd want another mimetype = meta.get(MIMETYPE, mimetype)\n assert mimetype is not None \n newrev[MIMETYPE] = mimetype?
2009-06-04T10:54:08  <ThomasWaldmann> not "another", only that
2009-06-04T10:55:15  <dennda> ok I'll do that later
2009-06-04T10:55:26  <dennda> that SA backend is gonna be trickier than I expected
2009-06-04T10:55:58  <ThomasWaldmann> btw, there is LOTS else to do before the SA backend :)
2009-06-04T10:56:42  <dennda> it'd help me if we could exactly outline what "LOTS" is
2009-06-04T10:57:28  <ThomasWaldmann> well, one goal is to make it production (or at least beta test ready)
2009-06-04T10:57:48  <ThomasWaldmann> i guess you can easily find out that this might not work with the current code
2009-06-04T10:58:21  <ThomasWaldmann> a SA backend is not required for beta test, therefore this was scheduled after midterm
2009-06-04T10:59:04  <ThomasWaldmann> but docs, working system pages, practically tested code, less broken tests, ... should be done before beta
2009-06-04T10:59:10  <dennda> I agree to postpone the SABackend, but please let's be precise on what to do now
2009-06-04T11:04:20  <ThomasWaldmann> also, we need to switch to permanent storage. mem backend is nice for playing and for having something fresh, but for a bit more than trivial practical testing rather unusable.
2009-06-04T11:05:11  <ThomasWaldmann> so, just try to practically USE a mm 2.0 wiki
2009-06-04T11:05:53  <dennda> I don't remember exactly if there was any backend stable enough to do that. I guess johill's fs backend is the candidate here?
2009-06-04T11:07:55  <ThomasWaldmann> yes, for now use that for rw data. and I guess you could also use fs17 for ro data.
2009-06-04T11:08:14  <dennda> btw, do you feel like fixing the testcases you broke? you know better what you did than I :-)
2009-06-04T11:08:34  <ThomasWaldmann> sure :)
2009-06-04T11:08:51  <dennda> great
2009-06-04T11:08:54  <ThomasWaldmann> but you could fix the stuff that crashes in backend. like all the wsgiapp tests
2009-06-04T11:09:15  <ThomasWaldmann> likely there is simply no backend init for that stuff.
2009-06-04T11:09:33  <ThomasWaldmann> E           item = self.backend.get_item(itemname)
2009-06-04T11:09:33  <ThomasWaldmann> >           AttributeError: 'NoneType' object has no attribute 'get_item'
2009-06-04T11:10:35  <ThomasWaldmann> also you need to fix the failing acl tests
2009-06-04T11:11:07  <ThomasWaldmann> and work on user.may...
2009-06-04T11:11:47  <ThomasWaldmann> and now please check your project page and update the todos / done there
2009-06-04T11:14:01  <ThomasWaldmann> dennda: btw: py.test | less   --> /^E
2009-06-04T11:14:09  <ThomasWaldmann> n n n
2009-06-04T11:17:01  <dennda> ThomasWaldmann: I guess I'm doing something wrong. What should that give me?
2009-06-04T11:18:40  <ThomasWaldmann> an easy way to see all errors?
2009-06-04T11:19:10  <dennda> ahhh
2009-06-04T11:19:17  <dennda> that's supposed to actually be an ^
2009-06-04T11:19:24  <dennda> I sent ctrl+e
2009-06-04T11:19:44  <dennda> which put it into ignore EOF or something
2009-06-04T11:19:52  <ThomasWaldmann> no, that's a start-of-line
2009-06-04T11:20:10  <dennda> sure
2009-06-04T11:20:31  <dennda> I quite like the --pdb option
2009-06-04T11:20:49  <dennda> did a tutorial on pdb the other day
2009-06-04T11:24:38  <dennda> ThomasWaldmann: did you see my user.may comment?
2009-06-04T11:25:54  <ThomasWaldmann> where?
2009-06-04T11:27:53  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5883:927310d84ef6 2.0-storage/MoinMoin/ (5 files in 2 dirs): skip some currently broken tests
2009-06-04T11:31:26  <dennda> ThomasWaldmann: yesterday, 16:21
2009-06-04T11:44:54  <ThomasWaldmann> dennda: iirc the user code needs the USER backend, not the data backend, right?
2009-06-04T11:48:05  <dennda> Well. If we want to make user.may use data_backend._may we need to have that data_backend first. The problem with that is that the data_backend (speaking of the AMW) needs the user himself so it has something to check against
2009-06-04T11:50:44  <ThomasWaldmann> that shouldn't be a real problem as long as you do not need to call "may" while setting up that stuff
2009-06-04T11:52:12  <dennda> i.e. you want me to lazy-load it?
2009-06-04T11:52:30  <dennda> that may work
2009-06-04T11:56:02  <ThomasWaldmann> maybe just patching in the backend / backend_may attr into user would work also
2009-06-04T11:58:38  <dennda> in wsgiapp? no
2009-06-04T11:58:48  <dennda> the tests construct their own user objects
2009-06-04T12:03:54  <ThomasWaldmann> a good idea would be to unify that stuff somehow, so it runs same code
2009-06-04T12:09:40  *** amartani has joined #moin-dev
2009-06-04T12:18:08  <dennda> hmpf
2009-06-04T12:18:19  <dennda> cfg.data_backend for the tests is None
2009-06-04T12:18:36  <dennda> *only* for the tests I think
2009-06-04T12:22:45  <ThomasWaldmann> bbl
2009-06-04T12:26:00  <CIA-19> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 5884:12ff0bf27e7f 2.0-storage/MoinMoin/_tests/test_wsgiapp.py: storage: first attempt to fix wsgi tests
2009-06-04T12:26:35  <dennda> ThomasWaldmann: when you come back, please look at that. There's still some problem of the following form:
2009-06-04T12:26:47  <dennda> >       assert ('Content-Type', 'text/html; charset=utf-8') in [('Content-Type', 'text/html'), ('Vary', 'Cookie, User-Agent, Accept-Language')]
2009-06-04T12:32:03  <dreimark> dimazest: on what do you work currently?
2009-06-04T12:32:28  <dimazest> on tests
2009-06-04T12:32:43  <dimazest> and i'm trying to get rid of request.dicts
2009-06-04T12:32:48  <dreimark> dennda: what is the minimum meta I should care for package installer. is mimetype and contentype enough?
2009-06-04T12:32:56  <dreimark> dimazest: ok
2009-06-04T12:33:19  <dennda> package installer?
2009-06-04T12:34:14  <dreimark> http://moinmo.in/HelpOnPackageInstaller e.g. PackagePages
2009-06-04T12:35:12  <xorAxAx> nearly 2 weeks passed and no single sign of work of my student :-(
2009-06-04T12:35:19  <dreimark> dennda: ^^^
2009-06-04T12:35:45  <dennda> dreimark: what exactly do you want to do/store? I don't know that package concept
2009-06-04T12:36:15  <dimazest> and caching seems to be more complicated
2009-06-04T12:36:38  <dreimark> dennda: the old one stored a page or an attachment to a zip file and instructions how to install it.
2009-06-04T12:36:42  <dennda> xorAxAx: get a whip
2009-06-04T12:37:02  <dreimark> dimazest: what are the plans for MyDict pages?
2009-06-04T12:37:13  <dennda> dreimark: so you want to store a zipfile?
2009-06-04T12:37:31  <dennda> the instructions are contained in that zipfile I suppose
2009-06-04T12:38:02  <dreimark> yes, but I need to extend the extract parameters they are written to a file in the zip
2009-06-04T12:39:13  <dimazest> no plans. what are MyDict pages? are they different from the group pages?
2009-06-04T12:39:31  <dreimark> yes
2009-06-04T12:39:49  <dreimark> http://moinmo.in/HelpOnDictionaries
2009-06-04T12:39:50  <dennda> well, the storage api doesn't require any metadata at all when storing something. the one who stores and retrieves it must be aware of what he wants to put in
2009-06-04T12:41:15  <dreimark> ok, that means if I add items to a zip file I need to know which mimetype it has.
2009-06-04T12:41:55  <dreimark> s/it has/they have/
2009-06-04T12:42:45  <dimazest> dreimark: so i deleted wikidicts module too early?
2009-06-04T12:43:10  <dennda> do you need to know that if you use pythons zip modules without MoinMoin.storage?
2009-06-04T12:43:29  <dennda> if no, then you don't need to know it with MoinMoin.storage either
2009-06-04T12:43:57  <dreimark> dimazest: yes. Some of that code is needed on other places.
2009-06-04T12:45:43  <dreimark> dimazest: for example http://moinmo.in/ReimarBauer/MyDict
2009-06-04T12:45:59  <dreimark> makes me able to write @IMG@ on any other page
2009-06-04T12:46:19  <dreimark> and it becomes expanded to {{attachment:ReimarBauer/Photo/img.png}}
2009-06-04T12:46:57  <dreimark> it is also used for translations
2009-06-04T12:47:01  <dimazest> i see...
2009-06-04T12:50:51  <dennda> dreimark: i.e., if the zipfile itself takes care of that, you can just put the whole zipfile into storage and retrieve it afterwards. I don't know how zipfile works
2009-06-04T12:55:09  <dennda> (you'd need to marshal it, of course)
2009-06-04T13:00:02  <dimazest> dreimark: this code works with dicts, not with groups? http://hg.moinmo.in/moin/1.9-groups-dmilajevs/file/df833ab68df2/MoinMoin/macro/__init__.py#l381
2009-06-04T13:00:57  <dimazest> then, i think, there shuld be request.groups and request.dicts
2009-06-04T13:01:27  <dreimark> yes
2009-06-04T13:02:06  * dimazest would like to have a time machine
2009-06-04T13:02:21  <dreimark> currently I would say yes because at least me would need some time to figure out if it can be done by the groups code
2009-06-04T13:02:39  * dreimark too and a clone of myself
2009-06-04T14:46:43  <dimazest> dreimark: this code also uses dicts not groups? http://hg.moinmo.in/moin/1.9/file/ab70e33b037c/MoinMoin/i18n/__init__.py#l306
2009-06-04T14:47:22  <dimazest> am i right, that groups are defined as * member1 \n * member2
2009-06-04T14:47:43  <dimazest> and dicts item1:: value \n item2:: value
2009-06-04T14:57:26  <dreimark> yes
2009-06-04T14:59:17  <dreimark> you can use theses Vars by e.g. [[GetVal(WikiDict,var1)]]
2009-06-04T14:59:44  <dreimark> and it will write the value then
2009-06-04T14:59:48  <dreimark> dimazest: ^^
2009-06-04T15:00:18  <dreimark> http://moinmo.in/WikiDict
2009-06-04T15:00:35  <dimazest> so, my plan is to do not touch wikidicts at all
2009-06-04T15:00:57  <dimazest> may be rename GroupDict to something else, because it serves dicts, not groups
2009-06-04T15:01:28  <dimazest> and wiki_group backend will be totally independent from wikidicts
2009-06-04T15:01:54  * dimazest thinks taht it would be nice to rename GroupDict
2009-06-04T15:03:57  <dreimark> wiki_group independent ./
2009-06-04T15:04:19  <dreimark> where is GroupDict currently used if it won't be used for groups ?
2009-06-04T15:05:05  <dimazest> http://hg.moinmo.in/moin/1.9/file/ab70e33b037c/MoinMoin/i18n/__init__.py#l306
2009-06-04T15:05:57  <dimazest> http://hg.moinmo.in/moin/1.9/file/ab70e33b037c/MoinMoin/macro/__init__.py#l389
2009-06-04T15:06:19  <dimazest> this to line cope with DICTS not with GROUPS
2009-06-04T15:06:37  <dimazest> *lines
2009-06-04T15:08:36  <dreimark> I think you can remove GroupDict from wikidicts but have to save Dict
2009-06-04T15:08:49  <dreimark> usecase http://moinmo.in/MoinAPI/Beispiele#wikidicts.Dict.28.29
2009-06-04T15:09:48  <dimazest> i thought about it,i need to check which methods are used, of which classes, this inheritance drives me crazy
2009-06-04T15:10:19  <dreimark> yeah that why we want it to become refactored. debugging this sh*t isn't fun
2009-06-04T15:10:24  <dimazest> first i want to isolate groups and dicts, and then refactor wikidicts
2009-06-04T15:10:32  <dreimark> ok
2009-06-04T15:14:08  <dimazest> ok, i need to go, c u
2009-06-04T15:14:27  <dreimark> hmm, alternative you could do a dicts backend and add it to request.dicts
2009-06-04T15:16:09  <dimazest> :) we will see
2009-06-04T15:16:40  <dimazest> it depends how fast i solve caching problems
2009-06-04T15:17:12  <dreimark> if caching blocks don't forget to ask
2009-06-04T15:18:04  <dimazest> sure
2009-06-04T17:40:49  *** adg has joined #moin-dev
2009-06-04T17:57:20  *** alfonsodg has joined #moin-dev
2009-06-04T18:03:32  *** alfonsodg1 has joined #moin-dev
2009-06-04T18:10:54  *** alfonsodg has quit IRC
2009-06-04T18:33:12  *** rkerr has joined #moin-dev
2009-06-04T18:33:14  <ThomasWaldmann> dennda: it at least does not look obviously wrong :) does it work?
2009-06-04T18:53:46  *** alfonsodg1 has quit IRC
2009-06-04T19:06:22  *** grzywacz has joined #moin-dev
2009-06-04T19:25:51  *** alfonsodg has joined #moin-dev
2009-06-04T20:17:18  <ThomasWaldmann> http://www.google.com/squared/search?q=wiki - wtf?
2009-06-04T20:21:11  *** alfonsodg has quit IRC
2009-06-04T20:26:38  <TheSheep> low confidence
2009-06-04T20:30:06  <dennda> ThomasWaldmann: i meant "look at the tests"
2009-06-04T20:30:17  <dennda> they still fail due to some reason I know nothing about
2009-06-04T20:41:13  <ThomasWaldmann> obviously the charset declaration is missing and there are no pages, so it 404s
2009-06-04T20:42:30  <dennda> i.e., nothing I need to worry about?
2009-06-04T20:42:48  <ThomasWaldmann> no
2009-06-04T20:43:43  <ThomasWaldmann> except that we soon have some pages to render and not just empty memory :)
2009-06-04T21:31:40  <dimazest> now i have 2 branches in my local repo
2009-06-04T21:32:20  <dimazest> one branch is with good commits, but second with wrong, so i would like to get rid from the second
2009-06-04T21:32:27  <dimazest> is it possible?
2009-06-04T21:33:14  <dimazest> i can create repo, which then can be cloned in place of http://hg.moinmo.in/moin/1.9-groups-dmilajevs
2009-06-04T21:38:15  <dimazest> or i cen merge just to have one head in the repo
2009-06-04T22:01:38  <ThomasWaldmann> you can't remove something from a repo, except when using convert or some similar too to rewrite/filter the whole repo
2009-06-04T22:01:45  <ThomasWaldmann> -too
2009-06-04T22:02:14  <ThomasWaldmann> +tool
2009-06-04T22:02:24  <ThomasWaldmann> dimazest:
2009-06-04T22:03:53  <dimazest> ok then i'll merge those two heads
2009-06-04T22:04:52  <dimazest> good night
2009-06-04T22:05:06  <ThomasWaldmann> gn
2009-06-04T22:08:28  <dreimark> re
2009-06-04T22:19:27  <dreimark> ThomasWaldmann: stupid that it fetches this line
2009-06-04T22:20:10  <dreimark> and where did it get moinmoin-1.0 ?
2009-06-04T22:20:37  <dreimark> http://www.svbug.com/administrator/packages&ports/CATEGORY/python.html
2009-06-04T22:22:17  <CIA-19> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5885:10a16fc7b3bf 2.0-storage/MoinMoin/theme/__init__.py: fix default content-type header, include charset
2009-06-04T22:42:40  <ThomasWaldmann> E       assert not can_access
2009-06-04T23:29:12  *** grzywacz has quit IRC
2009-06-04T23:56:28  <dreimark> gn#