2008-08-02T02:22:39 <dreimark> re
2008-08-02T02:22:43 * dreimark feels better
2008-08-02T02:23:02 <dreimark> byeongweon: please fix the pep8 failures in your repository
2008-08-02T02:24:16 <dreimark> you got them listed by running py.test MoinMoin/_tests/test_sourcecode.py
2008-08-02T02:25:30 <dreimark> #print "'%s'" % i.data.replace('\n', '') if not needed anymore that can be removed
2008-08-02T02:33:23 <dreimark> byeongweon: please use for testing of nesting the http://moinmo.in/ReimarBauer?action=AttachFile&do=get&target=UnderConstruction.zip
2008-08-02T02:33:57 <dreimark> install it to your wiki by adding it as attachment to a page and calling install from the AttachFile view
2008-08-02T02:34:24 <dreimark> load it in the gui editor and save it
2008-08-02T02:34:30 <dreimark> then look at the diff
2008-08-02T02:35:31 <dreimark> nesting is broken for parsers
2008-08-02T02:43:49 <dreimark> you should also make a copy in the text editor of HelpOnMoinWikiSyntax
2008-08-02T02:44:18 <dreimark> and save it before you look at it with the gui editor
2008-08-02T02:44:43 <dreimark> compare to the first revision after saving with the gui editor
2008-08-02T02:47:19 <dreimark> e.g. Wiki''''''Name becomes WikiName
2008-08-02T02:47:38 <dreimark> [[http://www.example.com/]]notlinked becomes http://www.example.com/notlinked
2008-08-02T02:49:13 <dreimark> if there are equivalent patterns in tables used that is ok
2008-08-02T02:49:15 <dreimark> e.g.
2008-08-02T02:49:22 <dreimark> ||<#0000FF> blue ||<#00FF00> green ||<#FF0000> red || becomes
2008-08-02T02:49:31 <dreimark> |<bgcolor="#0000ff">blue ||<bgcolor="#00ff00">green ||<bgcolor="#ff0000">red ||
2008-08-02T02:51:02 <dreimark> may be this is related to the nesting fix you did
2008-08-02T02:51:13 <dreimark> {{{#!wiki red/solid becomes
2008-08-02T02:51:29 <dreimark> {{{\n\n#!wiki red/solid
2008-08-02T02:55:07 <dreimark> one empty line would be better I am not sure if any amount of empty lines is ok, ThomasWaldmann please add a comment
2008-08-02T02:56:49 <dreimark> at the beginning of the diff I've overssen another problem
2008-08-02T02:56:54 <dreimark> : Line 9:
2008-08-02T02:56:54 <dreimark> '''Contents''' (up to the 2nd level)
2008-08-02T02:57:00 <dreimark> <<TableOfContents(2)>>
2008-08-02T02:57:12 <dreimark> is now '''Contents''' (up to the 2nd level) <<TableOfContents(2)>>
2008-08-02T02:57:26 <dreimark> without the linebreak
2008-08-02T02:58:22 <dreimark> this eems to happen on every browser call which is done in a new line
2008-08-02T02:59:20 <dreimark> this is equivalent but should it changed ?
2008-08-02T02:59:50 <dreimark> s/browser/parser/
2008-08-02T03:00:00 * dreimark should sleep
2008-08-02T03:00:09 <dreimark> s/browser/makro/
2008-08-02T03:11:36 <dreimark> dennda: if I use the memory backend, do you or PawelPacana think it is difficult to dump it by an action to e.g. a mercurial or sql snapshot ?
2008-08-02T03:12:12 <dreimark> or wahtever
2008-08-02T03:13:51 <dreimark> dennda: how can I get the underlay please pointme to the docs
2008-08-02T03:15:32 <dreimark> PawelPacana: where do I have to look for the essential things to run the repo, e.g. which mercurial version
2008-08-02T03:16:02 * dreimark anyway good night
2008-08-02T08:53:26 <dreimark> moin
2008-08-02T10:07:51 <zenhase> moin
2008-08-02T10:19:11 <dennda> moin
2008-08-02T10:19:18 <dennda> dreimark: seriously, I am astonished
2008-08-02T10:19:36 <dennda> How do you do that? go to bed at 3:16 am and be awake again at 8:53?!?
2008-08-02T10:23:05 * dennda always sleeps 10 hours, give or take, and I hate it
2008-08-02T10:28:47 <zenhase> hmm
2008-08-02T10:28:49 <dreimark> dennda: no not really I slept from ~9 pm to ~3am had a little break
2008-08-02T10:29:20 <dennda> ah I see
2008-08-02T10:29:26 <zenhase> i went to bed at 2:00 and was awake again at 8:00
2008-08-02T10:29:32 <dennda> I was hoping there was a special trick I could learn
2008-08-02T10:29:53 <dennda> I waste so much time sleeping... And no, multiple clocks put to hardly reachable locations do not help
2008-08-02T10:29:55 <xorAxAx> real men dont need any sleep
2008-08-02T10:30:07 <zenhase> dennda: more sport?
2008-08-02T10:30:29 <dennda> zenhase: probably. had little the last few weeks due to my exams
2008-08-02T10:30:38 <dennda> anyway, /me takes a shower
2008-08-02T10:30:47 * dreimark sometimes solves difficult problems while sleeping
2008-08-02T10:30:52 <dennda> dreimark: I will think about your questions then
2008-08-02T10:38:10 * dreimark seems not awake
2008-08-02T10:38:41 <dreimark> re.split('@|[$\.]',mail[::-1] ,2) how can I revert that now
2008-08-02T10:40:53 <dreimark> reverse
2008-08-02T10:40:58 <dreimark> reverse
2008-08-02T10:43:02 <dreimark> bbl
2008-08-02T10:51:16 <dennda> dreimark: As for the "dump": I think it'd be technically possible. In that action you'd just have to create the backend you want to dump to, and then iterate over all items and all revisions and transfer the content. That's not too hard. You basically use a migrationscript there. I just think it makes little sense to have such an action. Could you describe a more... general use case?
2008-08-02T10:51:51 <dennda> Because if we allow that, people will start using it. And then people will start complaining "oh, my important file's gone! how can I restore it? what? I can't?!? moin is teh suck!"
2008-08-02T10:51:55 <dennda> :)
2008-08-02T10:53:45 <xorAxAx> dennda: a dump is a mirroring into another bakcend, like migration :)
2008-08-02T10:59:10 <dennda> xorAxAx: Didn't I say that? I think I did :)
2008-08-02T10:59:28 <dennda> """...You basically use a migrationscript there...."""
2008-08-02T11:00:18 <dennda> I think he wants to work on a memory backend for some hours, then, when he has enough content in it, call an action that puts it into a more persistant backend
2008-08-02T11:00:47 <dennda> boy is it ho
2008-08-02T11:00:48 <dennda> t
2008-08-02T11:00:54 <dennda> I need air conditioning
2008-08-02T11:46:17 <dreimark> dennda: I guess wikisync, packages together with the memory backend works, that means you don't need any file io.
2008-08-02T11:47:30 <dreimark> you can setup a machine which gives one the possibility to play with MM and if he likes it he could transfer his data to a real instance
2008-08-02T11:49:18 <ThomasWaldmann> * SECURITY UPDATE: arbitrary code execution via multiple integer overflows.
2008-08-02T11:49:30 <dreimark> how much ram can we use from google webapl.
2008-08-02T11:49:39 <ThomasWaldmann> (taken from Python 2.5 security update)
2008-08-02T11:50:08 <dennda> ThomasWaldmann: Thanks for sharing. Is a fix available already or has it just been discovered?
2008-08-02T11:50:29 <ThomasWaldmann> just coming in via ubuntu updater
2008-08-02T11:50:31 <dennda> Or what list / site did you get that from?
2008-08-02T11:50:34 <dennda> Ah ok
2008-08-02T12:19:49 <CIA-53> Byeongweon [tasyblue@gmail.com] default * 3953:513613f25bbf 1.8-guieditor-mbyeongweon/MoinMoin/ (2 files in 2 dirs): fix source code with PEP8
2008-08-02T12:53:47 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4214:611202bfec43 1.8-dom-bblank/MoinMoin/converter2/creole_in.py:
2008-08-02T12:53:47 <CIA-53> Creole input converter - Move lines iterator into methods
2008-08-02T12:53:47 <CIA-53> It may change during the run of the nowiki/parser block.
2008-08-02T12:54:21 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4215:fc8d4a61cc13 1.8-dom-bblank/MoinMoin/converter2/ (_tests/test_creole_in.py creole_in.py): Creole input parser - Parse nowiki/parser sections with our own type directly
2008-08-02T13:30:40 <gizmach> moin, back in zagreb again
2008-08-02T13:34:56 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4216:c16d9155c615 1.8-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Move lines iterator into methods
2008-08-02T13:36:53 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4217:5e595ac59d53 1.8-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input parser - Parse nowiki/parser sections with our own type directly
2008-08-02T13:37:43 <xorAxAx> byeongweon: sigh, you still havent commented/updated the issues/tasks section on the wiki page
2008-08-02T13:38:00 <xorAxAx> neither the bug reports
2008-08-02T13:38:14 <gizmach> dreimark: did u look at my last push and why tests are failing
2008-08-02T13:39:10 <ThomasWaldmann> gizmach: welcome back
2008-08-02T13:39:32 <ThomasWaldmann> gizmach: btw, I pushed a changeset with more todo into your public repo
2008-08-02T13:39:36 <byeongweon> xorAxAx: I started that job but while doing that job other work to do happened. please wait a little more :)
2008-08-02T13:39:44 <gizmach> ThomasWaldmann: ok I will pull that
2008-08-02T13:39:46 <gizmach> thx
2008-08-02T13:40:10 <ThomasWaldmann> try to make not a mess this time when merging it :)
2008-08-02T13:40:20 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4402:f12f8e022e14 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile: When adding an Attachment to a page, use a buffer to write the file to storage
2008-08-02T13:41:14 <gizmach> ThomasWaldmann: ok :)
2008-08-02T13:43:19 <dennda> hm...
2008-08-02T13:48:54 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4218:b3e49b77dcf4 1.8-dom-bblank/MoinMoin/converter2/creole_in.py: Creole input converter - Output unsupported nowiki/parser sections as error
2008-08-02T13:48:57 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4219:cf2f225a6427 1.8-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Output unsupported nowiki/parser sections as error
2008-08-02T13:48:57 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4220:e688f0081d78 1.8-dom-bblank/MoinMoin/converter2/moinwiki_in.py: Moin Wiki input converter - Support css classes in nowiki/parser sections
2008-08-02T13:49:08 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3911:348c9b6e0033 1.8/MoinMoin/parser/ (5 files):
2008-08-02T13:49:08 <CIA-53> ParserBase (and other parsers deriving from it): cleanup / refactorings:
2008-08-02T13:49:08 <CIA-53> source code reformattings / rearrangements, PEP8 fixes
2008-08-02T13:49:09 <CIA-53> using booleans instead of 0/1, use raw strings for regexes
2008-08-02T13:49:10 <CIA-53> reduce code duplication
2008-08-02T13:49:12 <CIA-53> better attr names
2008-08-02T13:49:14 <CIA-53> introduce tabwidth constant as a class member
2008-08-02T14:14:40 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4403:35d0a7eb37f2 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile: Get rid of shutil import. Output proper size. (Via rewriting _write_stream)
2008-08-02T14:26:55 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4221:6161417fcec4 1.8-dom-bblank/MoinMoin/converter2/creole_in.py:
2008-08-02T14:26:55 <CIA-53> Creole input converter - Simplify iterator usage
2008-08-02T14:26:55 <CIA-53> Actually it is possible to use "for ... in ...", it just maps to the next
2008-08-02T14:26:55 <CIA-53> method.
2008-08-02T14:33:48 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3912:01ba9b2e96f6 1.8/MoinMoin/parser/_ParserBase.py: ParserBase: refactoring: reduce code duplication, remove unused code, rearrange
2008-08-02T14:46:10 * dennda wonders why moin doesn't allow comments when adding an attachment
2008-08-02T14:46:42 <xorAxAx> dennda: "allow"?
2008-08-02T14:46:50 <xorAxAx> dennda: well, nobody ever added it to the action
2008-08-02T14:46:59 <xorAxAx> there is no fundamental disallowance
2008-08-02T14:47:02 <dennda> that's what I meant :)
2008-08-02T14:47:02 <mmihaljevic> ThomasWaldmann: did you try to run the tests in my repo? all tests are failing for same thing (the change I done in security/__init__.py in may) but when I try to play with wiki everything was working :|
2008-08-02T14:47:08 <dennda> badly expressed, though
2008-08-02T14:49:57 <ThomasWaldmann> mmihaljevic: not yet. do you have uncommitted stuff?
2008-08-02T14:50:26 <mmihaljevic> ThomasWaldmann: only test names changes but this was also before that
2008-08-02T14:52:52 * mmihaljevic checks everything on running wiki and it seemd working
2008-08-02T14:52:57 <ThomasWaldmann> that was a yes/no question
2008-08-02T14:58:20 <mmihaljevic> yes
2008-08-02T14:58:28 <mmihaljevic> I will push that in a sec
2008-08-02T14:58:52 <mmihaljevic> Now I don't have any uncommited stuff ThomasWaldmann
2008-08-02T14:59:04 <mmihaljevic> ah pardon
2008-08-02T15:00:04 <mmihaljevic> now it's ok
2008-08-02T15:00:04 <CIA-53> MelitaMihaljevic default * 3993:7c01fe50a124 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/ (8 files): changed test names to test_X_group.py
2008-08-02T15:00:08 <mmihaljevic> I tried to push as a root
2008-08-02T15:06:04 <ThomasWaldmann> ?
2008-08-02T15:08:15 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4404:31fbc5ba4080 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile.py: When adding an attachment, don't use editlog, but store the meta-information (hostname, user, time) in the attachment-item directly, using storage api directly
2008-08-02T15:09:03 <ThomasWaldmann> mmihaljevic: i don't see a clean merge commit
2008-08-02T15:11:20 <ThomasWaldmann> mmihaljevic: did you pull first and redo your changes afterwards?
2008-08-02T15:13:43 <mmihaljevic> I redo my changes before and i pull and update
2008-08-02T15:15:07 <ThomasWaldmann> doesn't mercurial complain if you update a dirty workdir?
2008-08-02T15:15:29 <waldi> it will merge it
2008-08-02T15:15:42 <dennda> try
2008-08-02T15:15:51 <dennda> and conflict if it didn't work automatically
2008-08-02T15:16:02 <waldi> it even conflicts identical changes ...
2008-08-02T15:16:12 <dennda> oh, does it?
2008-08-02T15:16:20 <dennda> never had that problem, afair
2008-08-02T15:16:38 <dennda> if your updated changeset only affects a different file there should be no problem
2008-08-02T15:18:33 <mmihaljevic> ThomasWaldmann: I did pull and it said I need only to update
2008-08-02T15:18:36 <mmihaljevic> so I done update
2008-08-02T15:18:44 <mmihaljevic> and it didn't complain anythng
2008-08-02T15:20:12 <ThomasWaldmann> well, maybe you should get used to the normal mercurial work flow after working 2 months with it
2008-08-02T15:20:57 <ThomasWaldmann> that means you usually want to do "clean commits"
2008-08-02T15:21:19 <ThomasWaldmann> e.g. if you worked on some specific topic and you are done, you commit that
2008-08-02T15:21:56 <ThomasWaldmann> if you pull some changeset from another repo, you have to merge them and the you do a commit with just the merge and nothing else.
2008-08-02T15:23:59 <ThomasWaldmann> mmihaljevic: ok, so now please remove all those **FOO after fixing them.
2008-08-02T15:24:32 <mmihaljevic> ThomasWaldmann: ok
2008-08-02T15:24:39 <mmihaljevic> ThomasWaldmann: well I pulled from my repo
2008-08-02T15:24:59 <ThomasWaldmann> and use grep to make sure you didn't miss one
2008-08-02T15:25:21 <mmihaljevic> ThomasWaldmann: in every other cases I merged with another repo (with clean comit) except that one time when I messed something
2008-08-02T15:25:25 <mmihaljevic> ThomasWaldmann: ok
2008-08-02T15:25:46 <ThomasWaldmann> another repo means "any other repo"
2008-08-02T15:25:55 <ThomasWaldmann> including your public one
2008-08-02T15:26:42 <mmihaljevic> ThomasWaldmann: ok
2008-08-02T15:30:34 <xorAxAx> http://code.google.com/soc/2008/psf/appinfo.html?csaid=AFE1EA0F0C8D886F
2008-08-02T15:30:41 <xorAxAx> somebody integrating crunchy nd moin
2008-08-02T15:35:41 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3913:ab09415de20b 1.8/MoinMoin/parser/_ParserBase.py: ParserBase: use less confusing names, avoid calculating len(text) repeatedly
2008-08-02T15:40:07 <ThomasWaldmann> bah, ParserBase only opens code_line, but never closes :|
2008-08-02T15:47:45 * ThomasWaldmann looks at code_ formatter and gets blind :|
2008-08-02T16:02:47 <dennda> damn it
2008-08-02T16:02:55 <dennda> I was boiling water and forgot to set an alarm clock
2008-08-02T16:03:01 <dennda> now guess what happened ;)
2008-08-02T16:03:44 <waldi> no more water?
2008-08-02T16:04:00 <dennda> correct
2008-08-02T16:04:15 <xorAxAx> also works with noodles
2008-08-02T16:04:16 <dennda> only a salty pot
2008-08-02T16:04:38 <dennda> Yes, I was going to cook noodles, but luckily I didn't insert them yet :)
2008-08-02T16:07:41 <dennda> vim is trying to fool me
2008-08-02T16:08:48 <dennda> buggered editor
2008-08-02T16:17:51 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4405:9765359f63fb 1.8-storage-cdenter/MoinMoin/action/AttachFile.py:
2008-08-02T16:17:51 <CIA-53> storage: action/AttachFile:
2008-08-02T16:17:51 <CIA-53> * Adding metadata key to attachment-revisions indicating they actually *are* attachments
2008-08-02T16:17:51 <CIA-53> * Rewrite _get_files to get all attachments of a page (using MoinMoin.search)
2008-08-02T16:17:51 <CIA-53> * The latter BREAKS the list of attachments shown on a page for now. (It was bogus anyway.)
2008-08-02T16:20:52 * dennda is gone, cooking
2008-08-02T16:32:30 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3914:bd0aa9e93ed6 1.8/MoinMoin/parser/ (_ParserBase.py text_python.py): ParserBase / text_python parser: only use a single request.write
2008-08-02T16:34:29 <ThomasWaldmann> dennda: what is the difference between an attachment, a page, and a item?
2008-08-02T16:34:54 <dennda> ThomasWaldmann: I'm eating at the moment. Will explain in a minute
2008-08-02T16:39:21 <dennda> ThomasWaldmann: First of all, there is no seperate Attachment class, thus no attachment objects
2008-08-02T16:40:20 <dennda> When you add an attachment to a page, a new item is created in the backend. The name of the item is composed as follows: pagename + "/" + filename
2008-08-02T16:40:58 <dennda> Then, a new revision on that item is created, and the revision filled with the attachments data
2008-08-02T16:41:12 <dennda> that's all
2008-08-02T16:42:14 <dennda> i.e., if you want to test that, do it as follows:
2008-08-02T16:44:01 <dennda> pull my latest changes, create a new empty data/ dir and use the FSBackend as data_backend on that data/ directory. then run the wikiserver, open a page, say, FrontPage, click attachments, select a file and upload it. You will then have a new folder in your data/ folder with some number. in that folder, you will have a file name and rev.0. in name you'll find pagename + "/" + filename, and the contents (and some metadata) in rev.0
2008-08-02T16:45:23 <dennda> Was that precise enough to understand? :)
2008-08-02T16:46:13 <xorAxAx> why would he want to know the backend details? :)
2008-08-02T16:47:21 <dennda> ok then let me put it this way: attachments don't exist on the storage layer.
2008-08-02T16:47:56 <dennda> neither on any other layer. some "special" items/revisions are just presented as "attachments" to the user
2008-08-02T16:49:01 <dennda> better, xorAxAx? :)
2008-08-02T16:49:03 <ThomasWaldmann> dennda: i hope you are clear about that this can be only an intermediary solution
2008-08-02T16:49:22 <dennda> ThomasWaldmann: please elaborate
2008-08-02T16:49:43 <ThomasWaldmann> in the end, there will be NO difference between attachments and pages, they are all just items with different mimetype
2008-08-02T16:50:11 <dennda> well, that's the way I am just implementing it
2008-08-02T16:50:29 <ThomasWaldmann> you told you mark them as attachments
2008-08-02T16:52:52 <dennda> When generating the list of attachments of a page, you want to check whether it is a normal page by looking up the mimetype, and if not, add it to the list of attachments?
2008-08-02T16:53:10 <dennda> "it" being all items that may be attachments
2008-08-02T16:55:38 <dennda> that's basically the same I am doing right now, except I don't look if it has an appropriate mime-type to render it as a normal page, but just check if I set that "format" = "attachment" flag
2008-08-02T16:55:44 <dennda> you can replace that easily, however
2008-08-02T16:56:17 <dennda> should be enough to just adjust the search.term
2008-08-02T16:56:27 <dennda> (and store the mimetype in metadata, of course)
2008-08-02T16:57:36 <cortana> the change in http://hg.moinmo.in/moin/1.7/rev/a7db3c488083 was only made to multiconfig.py not wikiconfig.py
2008-08-02T16:57:56 <cortana> ditto the change in 1.8
2008-08-02T16:59:52 <ThomasWaldmann> dennda: please just use real mimetypes
2008-08-02T17:00:19 <dennda> ThomasWaldmann: Ok, I just need to figure out how to get the mimetype when one uploads an attachment
2008-08-02T17:00:40 <ThomasWaldmann> mimetypes lib
2008-08-02T17:00:57 <dennda> I'll check
2008-08-02T17:01:11 <ThomasWaldmann> (and make sure you import wikiutil first)
2008-08-02T17:03:06 <ThomasWaldmann> cortana: true
2008-08-02T17:09:12 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3784:2044213dcb13 1.7/wiki/config/wikiconfig.py: prevent CategoryTemplate being listed as a category (it is a Template, but matched also the category regex) - added to sample wikiconfig
2008-08-02T17:12:03 <dennda> ThomasWaldmann: Of what mimetype must it be to be rendered as a normal page?
2008-08-02T17:12:40 <dennda> text/plain?
2008-08-02T17:14:45 <TheSheep> text/moin-wiki
2008-08-02T17:15:18 <dennda> ok
2008-08-02T17:15:26 * dennda tries if johill's search.term is intuitive
2008-08-02T17:16:22 <dennda> hm, seems to be the case
2008-08-02T17:18:06 <dennda> johill: ping
2008-08-02T17:22:45 <xorAxAx> ThomasWaldmann: well, the current gui semantics and the old storage semantics match, the new storage semantics is different
2008-08-02T17:22:55 <xorAxAx> so both sides have to be modified to go through the new api
2008-08-02T17:23:17 <xorAxAx> (i.e. the moin17 backend and the attachfile code for example)
2008-08-02T17:23:36 <xorAxAx> thats a bit weird on the attachfile side but easiest as long there is no concrete new plan
2008-08-02T17:23:51 <xorAxAx> and it will give a good feeling whether the new attachment semantics are bearable
2008-08-02T17:29:01 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4406:0e3e7de1b3c3 1.8-storage-cdenter/MoinMoin/ (PageEditor.py action/AttachFile.py): storage: action/AttachFile.py: Use MimeTypes to check if an Item/Revision should be handled as an Attachment or a regular moin page. (removing seperate 'format' = 'attachment' flag.
2008-08-02T17:44:49 <dennda> wtf @ http://hg.moinmo.in/moin/1.8/file/bd0aa9e93ed6/MoinMoin/action/AttachFile.py -- lines 311 ff.
2008-08-02T17:52:09 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4407:23c149d7bf60 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile: First step to fix the attachment list. The list now shows all Attachments of a page. Not working: Size, last edit-time, zip-related functionality
2008-08-02T17:54:42 <xorAxAx> dennda: hmm?
2008-08-02T17:54:53 <xorAxAx> its just a bit copy&pasty :)
2008-08-02T17:55:15 <dennda> it decreases moins efficiency! :)
2008-08-02T17:55:26 <xorAxAx> hmm?
2008-08-02T17:55:37 <dennda> an additional, unneeded if-check
2008-08-02T17:55:53 <xorAxAx> ah
2008-08-02T17:56:09 <xorAxAx> such things happen in copy&pasty blocks of code
2008-08-02T17:56:15 <dennda> yes
2008-08-02T17:56:21 <dennda> second time I sped up moin today
2008-08-02T17:56:40 <dennda> time python -c "import shutil"
2008-08-02T17:57:09 <dennda> After I am finished with it, moin will be the fastest wiki out there, hehe :D
2008-08-02T17:57:30 <xorAxAx> sometimes you cannot see the wood for the trees anymore in such pasty parts
2008-08-02T17:57:46 <dennda> yeah
2008-08-02T18:03:09 <dennda> umm
2008-08-02T18:03:25 * dennda is trying to find where moin gets its timezone-related information from
2008-08-02T18:03:32 <xorAxAx> from the userprofile
2008-08-02T18:03:45 <dennda> What if I am viewing the page anonymously?
2008-08-02T18:03:47 <xorAxAx> thats why the user object is used to output times
2008-08-02T18:03:53 <xorAxAx> tamps
2008-08-02T18:03:59 <dennda> tamps?
2008-08-02T18:04:04 <xorAxAx> then its in the config of course
2008-08-02T18:04:07 <xorAxAx> timestamps
2008-08-02T18:04:12 <xorAxAx> why do you ask?
2008-08-02T18:04:25 <dennda> (I got a difference by two hours as anonymous user. I'll try what happens if I adjust that as user)
2008-08-02T18:05:07 <xorAxAx> by default moin is showing GMT
2008-08-02T18:05:20 <dennda> ah yes indeed
2008-08-02T18:05:20 <xorAxAx> but its not telling you that (which might be better probably)
2008-08-02T18:05:20 <dennda> great
2008-08-02T18:05:30 <xorAxAx> anyway, thats not storage related :)
2008-08-02T18:06:50 <dennda> you want to see when an attachmant was changed last time, don't you? (in the attachment list)
2008-08-02T18:07:47 <xorAxAx> probably
2008-08-02T18:07:59 <dennda> see?
2008-08-02T18:08:08 <dennda> I just wanted to make sure I was calling the function correctly :)
2008-08-02T18:09:58 <xorAxAx> but its not so important :)
2008-08-02T18:10:08 <xorAxAx> because people might want to rewrite that some time anyway
2008-08-02T18:10:13 <dennda> If it didn't work you'd have told me to fix it :)
2008-08-02T18:10:31 <dennda> well yeah, it was just easy to fix
2008-08-02T18:12:07 <xorAxAx> ah, didnt know that there is this feature already :)
2008-08-02T18:13:24 <dennda> which?
2008-08-02T18:13:35 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4408:11561b60e46c 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile.py: Fix displaying of an attachments size and last time of change. (With this changeset, you must delete your old data directory from FSBackend and create it again)
2008-08-02T18:14:18 <dennda> damnit, must tidy up a bit. bbl
2008-08-02T18:14:29 <dennda> or semi-afk, rather :)
2008-08-02T18:14:50 <xorAxAx> dennda: the times
2008-08-02T18:14:51 <xorAxAx> tamps
2008-08-02T18:15:16 <xorAxAx> dennda: why do you have to "delete" something?
2008-08-02T18:15:29 <xorAxAx> dennda: you only changed the frontend code, how can that break the on disk format?
2008-08-02T18:16:15 <dennda> xorAxAx: The new code assumes that there is specific metadata already on every item. That kind of metadata was just added
2008-08-02T18:16:44 <dennda> So your old Items don't have it yet, thus leading to a KeyError
2008-08-02T18:16:55 <xorAxAx> thats bad
2008-08-02T18:17:02 <dennda> Why is that bad?
2008-08-02T18:17:06 <xorAxAx> every piece of code must handle missing metadata keys
2008-08-02T18:17:14 <xorAxAx> because in some cases they are not available
2008-08-02T18:17:21 <dennda> yes
2008-08-02T18:17:31 <dennda> they are not available in the case one made a programming error
2008-08-02T18:17:40 <xorAxAx> no, also in migration cases
2008-08-02T18:17:49 <xorAxAx> where the data was not available
2008-08-02T18:18:07 <dennda> nope, this specific metadata will (and must) be added then
2008-08-02T18:18:13 <dennda> but I can handle it if you like
2008-08-02T18:18:26 <xorAxAx> well, you could log a warning or note
2008-08-02T18:18:30 <xorAxAx> but please dont break
2008-08-02T18:18:33 <dennda> I just won't see any programming error related to this if I produce one :)
2008-08-02T18:18:38 <dennda> ok
2008-08-02T18:18:48 <xorAxAx> its not a programming error, the idea is that some keys are optional
2008-08-02T18:18:54 <xorAxAx> because None is not a valid value, is it?
2008-08-02T18:19:22 <dennda> these keys are set by the very same code that reads them again. if they are not there something dirty happened
2008-08-02T18:19:36 <xorAxAx> ?
2008-08-02T18:19:46 <xorAxAx> every part of moin could add an item
2008-08-02T18:20:07 <dennda> hm.. ok
2008-08-02T18:20:13 <dennda> I will handle it when I get back home
2008-08-02T18:20:20 <dennda> my fathers birthday is tomorrow, mixed that up
2008-08-02T18:22:15 <xorAxAx> %-)
2008-08-02T19:25:33 * dennda is on his way back home, bbl
2008-08-02T19:28:04 <xorAxAx> ah, hah
2008-08-02T20:02:25 <byeongweon> xorAxAx: may I ask you a question?
2008-08-02T20:03:27 <xorAxAx> byeongweon: yes, sure
2008-08-02T20:03:41 <byeongweon> I'm changing fixed bugs status now.
2008-08-02T20:03:47 <byeongweon> about this bug http://moinmo.in/MoinMoinBugs/GuiEditorLineBreakInLists
2008-08-02T20:04:18 <byeongweon> I found I have mistake sorting this bug. this is unfixed.
2008-08-02T20:04:42 <byeongweon> what I want ask is first. is this bug?
2008-08-02T20:05:12 <byeongweon> and second if that true. should I insert br macro when user input shift+enter?
2008-08-02T20:05:30 <byeongweon> that's what I want to know.
2008-08-02T20:08:55 <xorAxAx> hmm, yeah, i think [[BR]] would be correct
2008-08-02T20:09:11 <xorAxAx> and i think its a bug, yes
2008-08-02T20:09:26 <ThomasWaldmann> <<BR>> please
2008-08-02T20:09:28 <byeongweon> ok. I should insert <<BR>> macro?
2008-08-02T20:09:34 <byeongweon> hehe. I think that.
2008-08-02T20:16:57 <xorAxAx> well, we mean the same :)
2008-08-02T20:29:06 * dennda drops in
2008-08-02T20:39:32 <dennda> Does windows allow creating filenames with a "/"?
2008-08-02T20:40:20 <xorAxAx> no
2008-08-02T20:40:23 <dennda> ok
2008-08-02T20:53:23 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4409:8a4da1657bb4 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile.py: Don't show an Exception to the user if an item doesn't have certain metadata
2008-08-02T20:57:01 <CIA-53> Byeongweon [tasyblue@gmail.com] default * 3954:e0301e5bf019 1.8-guieditor-mbyeongweon/MoinMoin/converter/text_html_text_moin_wiki.py: fix gui editor line break in list problem MoinMoinBugs/GuiEditorLineBreakInLists
2008-08-02T21:19:28 <dennda> _("foobar") immediately calls gettext in order to get the string translated, correct?
2008-08-02T21:20:19 <johill> via some weird stuff, yes
2008-08-02T21:20:29 <dennda> ah hi johill
2008-08-02T21:20:34 <johill> you usually use _ = request.getText
2008-08-02T21:20:40 <dennda> yeah
2008-08-02T21:20:41 <johill> you need to, in fact, to get strings labeled
2008-08-02T21:20:58 <dennda> just have error = _("asdasdasdasd") and want to do if error != ""
2008-08-02T21:21:14 <dennda> johill: seen my latest changesets?
2008-08-02T21:21:17 <johill> why would you want to do things like that?
2008-08-02T21:21:19 <johill> no, just got online
2008-08-02T21:21:20 <dennda> There's one thing I wanted to ask
2008-08-02T21:21:25 <dennda> ok
2008-08-02T21:21:29 <dennda> no worries, not too important
2008-08-02T21:21:58 <dennda> checking if an item exists and otherwise displaying that error message to the user
2008-08-02T21:21:59 <johill> sure shoot
2008-08-02T21:22:28 <johill> should revision size be an additional metadata thing?
2008-08-02T21:22:35 <johill> i.e. rev.size
2008-08-02T21:22:39 <johill> rather then rev['size']?
2008-08-02T21:22:46 <johill> it seems like we don't really need to store it
2008-08-02T21:23:06 <xorAxAx> yeah
2008-08-02T21:23:32 <dennda> how else do you want to cheaply get to know its size?
2008-08-02T21:23:39 <johill> well the fs backend can stat() it
2008-08-02T21:23:43 <johill> and the sql backend can do some sql magic
2008-08-02T21:23:58 <johill> and another backend can just keep the size as metadata if it needs to
2008-08-02T21:24:05 <dennda> I am not too sure you can do that easily and cheaply in any kind of backend
2008-08-02T21:24:18 <johill> well if it can't it can keep it at commit() time
2008-08-02T21:24:28 <johill> and calculate it during write()
2008-08-02T21:24:29 <dennda> so why don't do it right here? we got that information anyway and it won't change
2008-08-02T21:24:50 <johill> because then we'd have it for the info action too
2008-08-02T21:25:33 <dennda> hm, agreed
2008-08-02T21:25:47 <dennda> I already forgot again how that computes it %-)
2008-08-02T21:25:56 <dennda> johill: My question is mimetype related
2008-08-02T21:25:59 <johill> it reads it I think, via Page
2008-08-02T21:26:01 <johill> ok
2008-08-02T21:26:06 <dennda> just a sec
2008-08-02T21:26:20 <dennda> I am storing the mimetype of an attachment when it is saved, in metadata
2008-08-02T21:26:27 <dennda> maybe you have already seen that
2008-08-02T21:27:23 <dennda> I know that normal moin pages are supposed to have the mimetype text/moin-wiki
2008-08-02T21:27:36 * johill adds note to dennda's page about metadata keys
2008-08-02T21:27:42 <johill> ok
2008-08-02T21:27:50 <johill> what's the question then?
2008-08-02T21:27:56 <dennda> So I use these informations to tell attachments and normal wiki pages apart when I need to make a list of attachments
2008-08-02T21:28:04 <dennda> I'm coming to that, promise :)
2008-08-02T21:28:14 <johill> ok :)
2008-08-02T21:28:45 <dennda> To be complete, it sounds logical to store a mimetype text/moin-wiki for normal pages, too
2008-08-02T21:28:59 <johill> except, that's not really true for rst or creole pages
2008-08-02T21:29:01 <dennda> Thus, I edited PageEditor.py and added that to the list of stored metadata-values
2008-08-02T21:29:15 <johill> so you'd have to do something based on the format pragma
2008-08-02T21:29:23 <dennda> but with that enabled, it shows up in the editor as processing instruction
2008-08-02T21:29:28 <dennda> The question is: why? :)
2008-08-02T21:30:00 <johill> well or use application/x-moin-wiki-any?
2008-08-02T21:30:10 <dennda> http://hg.moinmo.in/moin/1.8-storage-cdenter/rev/0e3e7de1b3c3
2008-08-02T21:30:21 <dennda> see the very first changes there
2008-08-02T21:30:27 <dennda> when I uncomment that line, the error occurrs
2008-08-02T21:30:42 <johill> what error?
2008-08-02T21:30:50 <johill> anyhow, text/moin-wiki is just wrong
2008-08-02T21:30:57 <dennda> 21:29 < dennda> but with that enabled, it shows up in the editor as processing instruction
2008-08-02T21:31:00 <johill> mimetypes need to be registered with the ietf (or iana?)
2008-08-02T21:31:03 <johill> ah!
2008-08-02T21:31:07 <johill> yeah
2008-08-02T21:31:11 <dennda> blame TheSheep for that! :D
2008-08-02T21:31:13 <johill> that's the dumb metadata code in wikiutil
2008-08-02T21:31:18 <dennda> he told me it was text/moin-wiki
2008-08-02T21:31:29 <johill> add_metadata_to_body
2008-08-02T21:31:37 <johill> I haven't seen a registration for that
2008-08-02T21:31:54 <johill> anyway, look at add_metadata_to_body
2008-08-02T21:32:02 <dennda> neither did I. I trusted him on that :)
2008-08-02T21:32:15 <johill> and make a list there so it uses only those PIs we used to support
2008-08-02T21:32:22 <johill> and ignores the rest of the metadata
2008-08-02T21:32:48 <johill> TheSheep: well, where is the registration for that mimetype then?
2008-08-02T21:33:20 <johill> :)
2008-08-02T21:33:49 <dennda> Did I already say that editing AttachFile.py is much more fun than Page.py, PageEditor.py and user.py?
2008-08-02T21:33:52 <TheSheep> johill: in the parser directory of MoinMoin, no?
2008-08-02T21:34:55 <johill> dennda: :)
2008-08-02T21:35:31 <johill> TheSheep: well, have you read the relevant RFCs? they require assignment from the IANA for non x- mimetypes
2008-08-02T21:35:35 <dennda> johill: How to get a list of all PI's that are supported?
2008-08-02T21:35:55 <johill> dennda: heh good question, look in Page.py I suspect
2008-08-02T21:35:59 <dennda> doesn't matter too much, does it?
2008-08-02T21:36:03 <TheSheep> johill: I didn't pick this name, I just reported how it currently works
2008-08-02T21:36:13 <dennda> hm, great
2008-08-02T21:36:15 <johill> parse_processing_instructions
2008-08-02T21:36:19 <dennda> ok
2008-08-02T21:36:27 <johill> TheSheep: ok, I think it's wrong though
2008-08-02T21:36:37 <johill> regardless though, this is just the "moin" format
2008-08-02T21:36:38 <TheSheep> johill: #fromat wiki is mapped to text/moin-wiki in wikiutil.py
2008-08-02T21:36:44 <johill> and hence we shouldn't use it for all editable pages
2008-08-02T21:36:48 <dennda> shall I change it to text/x-moin-wiki then?
2008-08-02T21:36:50 <johill> since we have rst too etc
2008-08-02T21:37:01 <xorAxAx> the stored mimetype obviously depends on the default mimetype
2008-08-02T21:37:03 <xorAxAx> of the wiki
2008-08-02T21:37:07 <xorAxAx> its easy :)
2008-08-02T21:37:27 <johill> xorAxAx: well if you write #format rst it shouldn't be text/x-moin-wiki or whatever tho
2008-08-02T21:37:45 <xorAxAx> yeah, default or PI
2008-08-02T21:38:05 <johill> hence, I think currently it makes sense to identify attachments explicitly
2008-08-02T21:38:20 <johill> until we actually unify editing of pages and attachments
2008-08-02T21:38:27 <xorAxAx> ?
2008-08-02T21:38:47 <xorAxAx> sounds dirty
2008-08-02T21:38:54 <johill> attachfile _is_ dirty
2008-08-02T21:39:04 <dennda> on another note: we already have a metadata key "deleted" for deleted items. what are correct values? True / False I assume...
2008-08-02T21:39:04 <johill> but right now, we still want to be able to enumerate attachments
2008-08-02T21:39:23 <johill> dennda: actually, I meant it to be just "present"/"not present", but doesn't matter much to me
2008-08-02T21:39:49 <xorAxAx> johill: but you are suggesting this workaround because the non-attachfile parts are not yet enhanced
2008-08-02T21:39:51 <dennda> ok, I'll do an explicit True
2008-08-02T21:40:03 <xorAxAx> johill: attachfile itself is "fine"
2008-08-02T21:40:15 <johill> xorAxAx: sort of, but no, attachfile isn't really "fine"
2008-08-02T21:40:26 <johill> because it tries to "list attachments"
2008-08-02T21:40:29 <johill> which really doesn't make sense
2008-08-02T21:40:29 <xorAxAx> dennda is making it superb :)
2008-08-02T21:40:35 <xorAxAx> sure, johill
2008-08-02T21:40:38 <xorAxAx> it exists by definition
2008-08-02T21:40:43 <johill> and should just be "list subitems with their mimetypes"
2008-08-02T21:40:45 <xorAxAx> and this is not an action problem
2008-08-02T21:40:54 <xorAxAx> but a GUI design problem
2008-08-02T21:40:57 <johill> exactly
2008-08-02T21:41:01 <xorAxAx> you change peoples workflows
2008-08-02T21:41:04 <johill> and we don't actually have time to redesign the GUI
2008-08-02T21:41:06 <xorAxAx> and need to find an equivalent
2008-08-02T21:41:12 <johill> or rewrite it, for that matter
2008-08-02T21:41:25 <xorAxAx> and the new definition is sane if the the idea of throwing together namespaces is assumed to be fine
2008-08-02T21:41:35 <xorAxAx> so i dont see any problem, except for the namespace merge probably :)
2008-08-02T21:41:46 <johill> oh I agree that the new definition is fine in principle
2008-08-02T21:41:52 <johill> but storing all editable pages as a single mimetype isn't
2008-08-02T21:42:00 <johill> because format should be equivalent to mimetype
2008-08-02T21:42:07 <xorAxAx> how about using a transitional mimetype?
2008-08-02T21:42:13 <johill> therefore, the attachfile code would have to have knowledge about all the other mimetypes
2008-08-02T21:42:21 <xorAxAx> "some-wiki-page-with-unknown-format"
2008-08-02T21:42:35 <xorAxAx> i think thats the cleanest solution in the set concept
2008-08-02T21:42:36 <johill> yeah that's why I was suggesting application/x-some-wiki-format
2008-08-02T21:42:52 <johill> or text/
2008-08-02T21:43:00 <TheSheep> Iwonder, this is not relevant, but I was wondering about it since a while. If I have a broken browser that doesn't guess file's mimetype correctly, and I upload an attachment with wrong mimetype, can I change that mimetype later somehow?
2008-08-02T21:43:08 <johill> I guess it's the other way around, it explicitly identifies _non_ attachments then
2008-08-02T21:43:21 <xorAxAx> TheSheep: yes, thats a question of the metadata editing
2008-08-02T21:43:25 <johill> TheSheep: the browser doesn't guess the mimetype anyway
2008-08-02T21:43:33 <johill> moin does, and it sucks at it
2008-08-02T21:43:39 <xorAxAx> TheSheep: in a perfect world, there is a view to the metadata editing process that allows you to change mimetypes
2008-08-02T21:43:49 <johill> except revisions are actually immutable ;)
2008-08-02T21:43:58 * dennda likes your search.term
2008-08-02T21:44:01 <TheSheep> johill: I always thought you get the mimetype with the form data
2008-08-02T21:44:05 <xorAxAx> yeah, thomas also sells hard drives :)
2008-08-02T21:44:10 <johill> TheSheep: I don't think so
2008-08-02T21:44:18 <johill> TheSheep: afaik moin guesses based on the extension
2008-08-02T21:44:19 <dennda> quite easy to use. didn't even need to look up how to use it
2008-08-02T21:44:22 <johill> TheSheep: it really should use magic
2008-08-02T21:44:30 <johill> magic(5)
2008-08-02T21:44:41 <TheSheep> johill: I worded it worng: I *know* you get it, I just didn't know whether moin uses that information or not
2008-08-02T21:44:42 <johill> and maybe let the user set it like bugzilla
2008-08-02T21:44:57 <johill> TheSheep: oh ok, well, I don't think it does
2008-08-02T21:45:03 <TheSheep> johill: thanks
2008-08-02T21:45:28 <johill> dennda: :)
2008-08-02T21:45:33 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4410:3fe96f5bf02b 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile.py: Make _do_del() work.
2008-08-02T21:46:04 <dennda> johill: although the usage example you gave me yesterday was wrong :)
2008-08-02T21:46:15 <johill> dennda: so, maybe use "mimetype" "text/x-unidentified-wiki-format" or whatever like that for non-attachments, which is then used in conjunction with "format"
2008-08-02T21:46:18 <johill> heh
2008-08-02T21:46:28 <TheSheep> johill: I guess upload doesn't have to be fast, so you can do all sort of content detection and stuff
2008-08-02T21:46:33 <johill> I forgot the NOT?
2008-08-02T21:46:49 <johill> TheSheep: magic(5) usually just reads a few bytes of the file header before giving up
2008-08-02T21:47:10 <dennda> johill: No, you didn't mention I needed to pass an re-object, rather than a string representing a re
2008-08-02T21:47:10 <TheSheep> johill: it actually executes arbitrary code on a file
2008-08-02T21:47:19 <johill> dennda: ah, ok
2008-08-02T21:47:25 <johill> TheSheep: well, sort of arbitrary, yes
2008-08-02T21:47:31 <dennda> why unidentified?
2008-08-02T21:47:38 <johill> TheSheep: typically though, patterns aren't that
2008-08-02T21:47:39 <TheSheep> johill: so, for starters, you need a file... :)
2008-08-02T21:47:42 <dennda> we know that it's moin, don't we?
2008-08-02T21:47:48 <johill> TheSheep: not with the python implementation
2008-08-02T21:48:01 <johill> dennda: no, it could be rst, creole, ...
2008-08-02T21:48:05 <johill> anything moin has a parser for
2008-08-02T21:48:05 <TheSheep> johill: wow, there is a python implementation?
2008-08-02T21:48:22 <TheSheep> johill: I've been looking for it
2008-08-02T21:48:54 <johill> I thought there was
2008-08-02T21:48:58 <johill> it's been a while since I looked
2008-08-02T21:49:14 <TheSheep> there is 'mimetypes' module in stdlib, but it guesses based on extensions
2008-08-02T21:49:24 <TheSheep> there is also imghdr for guessing image types
2008-08-02T21:49:39 <TheSheep> and I think another one for sound
2008-08-02T21:49:47 * johill looks
2008-08-02T21:50:15 <dennda> yes, mimetypes.guess_mimetype already suggests that it guesses
2008-08-02T21:50:20 <dennda> the docs say the same
2008-08-02T21:51:08 <TheSheep> http://hupp.org/adam/hg/python-magic/ <-- there is something like that, hmm...
2008-08-02T21:51:20 <dennda> johill: I wonder when the first administrator of a forum arrives, complaining that somebody uploaded a 7 GB pr0n and that after deleting that attachment, disk space wasn't freed..
2008-08-02T21:52:26 <TheSheep> I guess it's nontrivial to set up on windows :)
2008-08-02T21:53:00 <johill> TheSheep: there's a libmagic python wrapper that can take buffers rather than a file name
2008-08-02T21:53:15 <johill> dennda: heh yeah we really do need purging, but lets get this stuff working right first
2008-08-02T21:53:55 <dennda> sure
2008-08-02T21:54:37 <dennda> johill: in page.py, processing stuff: pi["foobar"] means that "foobar" can be used as processing instruction? I know too little about the use of those PIs
2008-08-02T21:55:01 <johill> yeah
2008-08-02T21:55:20 <dennda> ok, that means i must write my own list
2008-08-02T21:55:30 <dennda> no problem, I just hope I don't forget one :)
2008-08-02T21:56:14 <dennda> so PI are stored as metadata as well?
2008-08-02T21:56:15 <johill> actually
2008-08-02T21:56:20 <johill> only the verb == stuff you care about
2008-08-02T21:56:28 <johill> yes
2008-08-02T21:56:45 <dennda> why is that?
2008-08-02T21:56:46 <johill> i.e. format, language, refresh, redirect, deprecated, openiduser, pragma
2008-08-02T21:56:51 <dennda> I'd like to understand it a bit better
2008-08-02T21:56:51 <johill> the other stuff is dynamically built
2008-08-02T21:57:12 <dennda> can I use your list there or isn't it complete? :)
2008-08-02T21:57:38 <johill> you probably want tol look at the original code
2008-08-02T21:57:38 <johill> rather than nlanius's branch
2008-08-02T21:57:46 <johill> that list was complete afaik
2008-08-02T21:58:02 <dennda> great, then I can tell people to blame you rather than me if it doesn't work
2008-08-02T21:58:05 <johill> hehe
2008-08-02T21:58:13 <johill> oh well it can't have been complete
2008-08-02T21:58:13 <CIA-53> Florian Krupicka <florian.krupicka@googlemail.com> default * 4031:e1a63fb9bb4d 1.8-wsgi-fkrupicka/MoinMoin/ (web/utils.py wsgiapp.py): Catch configuration errors in the application, not in a debug layer
2008-08-02T21:58:15 <CIA-53> Florian Krupicka <florian.krupicka@googlemail.com> default * 4032:f70003d57e47 1.8-wsgi-fkrupicka/MoinMoin/web/contexts.py: Added http_redirect for compatibility
2008-08-02T21:58:15 <johill> acl is missing for sure
2008-08-02T21:58:16 <CIA-53> Florian Krupicka <florian.krupicka@googlemail.com> default * 4033:6de8b4205754 1.8-wsgi-fkrupicka/MoinMoin/ (script/server/standalone.py web/serving.py wsgiapp.py): Moved serving-related code into MoinMoin.web.serving module
2008-08-02T21:58:45 <dennda> is that the only one missing? :)
2008-08-02T21:58:46 <zenhase> not much code today, since i mostly gathered information for deploying
2008-08-02T21:58:59 <zenhase> i still have some strange bug with mod_wsgi
2008-08-02T21:59:07 <johill> no idea :)
2008-08-02T21:59:09 <dennda> zenhase: which?
2008-08-02T21:59:25 <dennda> johill: Heh, you don't want to be blamed, right? :)
2008-08-02T21:59:26 <zenhase> when using daemon processes that run as another user than apache
2008-08-02T22:00:28 <zenhase> doesn't really work, even if directories of moin belong to the user i am switching the process to
2008-08-02T22:01:12 <dennda> zenhase: i guess you noticed that the knight of wsgi has been online a few times the last few days? maybe he knows
2008-08-02T22:02:19 <zenhase> the knight of wsgi?
2008-08-02T22:02:28 <zenhase> actually that must be ianb
2008-08-02T22:02:43 <dennda> you aren't that active on python-forum.de, are you? :)
2008-08-02T22:02:50 <zenhase> no, i am not
2008-08-02T22:03:02 <zenhase> i don't use bulletin-boards
2008-08-02T22:03:21 <zenhase> (or mailing lists as a matter of fact)
2008-08-02T22:03:39 <johill> phone? ;)
2008-08-02T22:04:22 <dennda> If that goddamn phpBB didn't suck that much I'd find his profile within seconds
2008-08-02T22:04:25 <zenhase> yeah, i do use phone, but my preferred media are chatrooms and instant messaging
2008-08-02T22:04:49 <dennda> http://www.python-forum.de/user-600.html
2008-08-02T22:05:05 <zenhase> haha, nice
2008-08-02T22:05:16 <zenhase> didn't know my mentor is called that ;)
2008-08-02T22:05:26 <zenhase> but ok, that kindof fits too
2008-08-02T22:08:46 <johill> dennda: I tend to from MoinMoin.search import term and then use term.NOT() etc, but you can do it this way too if you like it better
2008-08-02T22:09:35 <dennda> johill: oh, ok. The line would have been become a bit long. but now that I divided it to multiple lines I can do that as well
2008-08-02T22:09:44 <dennda> What is the reason for that preference of yours?
2008-08-02T22:10:13 <johill> fewer things to import when you need more things from it :)
2008-08-02T22:10:21 <johill> I was writing tests when developing that preference
2008-08-02T22:10:42 <johill> but yeah I don't care
2008-08-02T22:11:22 <dennda> johill: http://paste.pocoo.org/show/81061/
2008-08-02T22:11:30 <dennda> that's all mentioned in the Page.py method
2008-08-02T22:11:38 <dennda> havn't found "acl"
2008-08-02T22:12:39 <johill> yeah I think it goes via metadata[ACL] already somehow
2008-08-02T22:12:51 <dennda> whatever that means :)
2008-08-02T22:13:06 <johill> http://moinmo.in/HelpOnProcessingInstructions
2008-08-02T22:13:26 <johill> I suppose acls are already special
2008-08-02T22:13:35 <johill> they still need to be added there until we have an ACL editor or something
2008-08-02T22:13:47 <dennda> ok
2008-08-02T22:15:25 <johill> dennda: new_rev[DELETED] = True
2008-08-02T22:15:27 <johill> vs
2008-08-02T22:15:35 <johill> NOT(LastRevisionMetaDataMatch('deleted', True))
2008-08-02T22:15:50 * dennda thinks there should be a common place for constant-definitions
2008-08-02T22:16:04 <byeongweon> xorAxAx: is there any discussion about this one http://moinmo.in/MoinMoinBugs/GuiEditorShouldFollowBrMacroWithLineBreak ?
2008-08-02T22:16:08 <CIA-53> Florian Krupicka <florian.krupicka@googlemail.com> default * 4034:46ad11b00bed 1.8-wsgi-fkrupicka/MoinMoin/web/serving.py: Fixed: variable reference and left out request handler
2008-08-02T22:16:20 <johill> dennda: yeah you seem to already have DELETED so just use it both times
2008-08-02T22:16:28 <johill> dennda: 'xxx correct wrt. gettext'
2008-08-02T22:16:34 <byeongweon> related with new line in table
2008-08-02T22:16:38 <johill> dennda: why not just use 'try: except: else:' there?
2008-08-02T22:16:42 <zenhase> ups, missed something :o
2008-08-02T22:16:57 <johill> anyway, gtg
2008-08-02T22:17:08 <dennda> bye johill. i'll check
2008-08-02T22:17:26 <xorAxAx> byeongweon: i guess its a nice feature
2008-08-02T22:17:43 <xorAxAx> byeongweon: shift+enter doesnt yet insert the <<br>> macro, right? then this would go along with it i think
2008-08-02T22:17:55 <xorAxAx> not sure how hard it is, i dont know fckeditor well enough
2008-08-02T22:18:53 <byeongweon> xorAxAx: I think I can handle that server side using process_br in formatter.
2008-08-02T22:19:24 <xorAxAx> well, you can try it. as long as roundtripping works
2008-08-02T22:19:33 <xorAxAx> its basically fine
2008-08-02T22:20:05 <byeongweon> insert br macro is not a hard problem I think.
2008-08-02T22:20:32 <byeongweon> but table doesn't accept new line is little bit hard problem
2008-08-02T22:21:38 <byeongweon> when I insert new line in table, table doesn't work.
2008-08-02T22:22:10 <byeongweon> what I want to know is
2008-08-02T22:22:28 <byeongweon> is there any discussion about new line in table.
2008-08-02T22:23:40 <xorAxAx> http://moinmo.in/MoinMoinBugs/GuiEditorLineBreak
2008-08-02T22:24:16 <xorAxAx> but nothing more, no
2008-08-02T22:24:22 <xorAxAx> that bug page lists an issue, though
2008-08-02T22:24:35 <xorAxAx> but that shouldnt stop you from implementing above "feature request"
2008-08-02T22:27:35 <byeongweon> xorAxAx: hmm.. do you think MoinMoinBugs/GuiEditorShouldFollowBrMacroWithLineBreak this one also not a bug?
2008-08-02T22:29:20 <xorAxAx> why "not a bug"?
2008-08-02T22:29:23 <byeongweon> xorAxAx: I will away about 10 min :)
2008-08-02T22:29:29 <xorAxAx> all these br issues are kind of bugs
2008-08-02T22:29:36 <xorAxAx> because they often break roundtripping
2008-08-02T22:30:25 <xorAxAx> hmm, the latter bug report with the wiki markup is just a normalization issue
2008-08-02T22:30:30 <xorAxAx> and easily solvable in the converter
2008-08-02T22:30:49 <xorAxAx> you just have to check that you dont generate any empty line in those cases
2008-08-02T22:34:25 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4411:debfed81ecfe 1.8-storage-cdenter/MoinMoin/ (PageEditor.py action/AttachFile.py wikiutil.py): storage: attachfile/pageeditor/wikiutil: Fix to properly handle mimetypes as metadata.
2008-08-02T22:39:08 <dennda> johill: I cannot use try: except: else: there because that'd be ignorant of the first two errors that could occurr
2008-08-02T22:39:10 <byeongweon> xorAxAx: I'm back
2008-08-02T22:42:05 <byeongweon> hmm. then, should I insert new line after br macro?
2008-08-02T22:43:03 <byeongweon> when user open with gui editor.
2008-08-02T22:43:17 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3915:b4bdfaaa79b0 1.8/MoinMoin/parser/_ParserBase.py: ParserBase: do code_line() open and close calls correctly
2008-08-02T22:48:14 <CIA-53> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 4412:bf89a8d02007 1.8-storage-cdenter/MoinMoin/action/AttachFile.py: storage: action/AttachFile: Make _do_del() work better. Fix: Display an appropriate error when trying to delete a file that already has been deleted.
2008-08-02T22:51:26 <xorAxAx> byeongweon: yes, sounds like a nice thing if its purely cosmetical
2008-08-02T22:52:46 <byeongweon> xorAxAx: ok. I will try that during sorting issues and fix bugs work.
2008-08-02T22:53:19 <xorAxAx> cool
2008-08-02T22:53:53 <byeongweon> xorAxAx: it seems like little sensitive work. without unittest is hard :-(
2008-08-02T23:07:12 * dennda has a thinking barrier. Let's do a break MoinMoin: MoinMoinChat/Logs/moin-dev/2008-08-02 (last edited 2008-08-02 00:30:02 by IrcLogImporter)
