2009-08-07T00:05:55 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6443:4f42b89e8e48 2.0-storage/MoinMoin/ (3 files in 2 dirs): fix some tests
2009-08-07T00:05:56 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6444:dc7c8125067c 2.0-storage/ (5 files in 5 dirs): merged main
2009-08-07T00:06:08 <ThomasWaldmann> 1 failure left
2009-08-07T00:07:10 <dennda> :-)
2009-08-07T00:07:17 <dennda> and that's a sourcode one?
2009-08-07T00:07:43 <dennda> i guess you mean the other one
2009-08-07T00:16:10 <dennda> it would be great if we had some roadmap btw
2009-08-07T00:17:59 <ThomasWaldmann> for?
2009-08-07T00:19:56 <dennda> the last few days of this years gsoc
2009-08-07T00:20:54 <ThomasWaldmann> well, make one :)
2009-08-07T00:22:46 <dennda> ok, sqla backend only from now on :-)
2009-08-07T00:23:03 <dennda> sound sane? :P
2009-08-07T00:25:25 <ThomasWaldmann> sigh
2009-08-07T00:25:57 * ThomasWaldmann .o(read your wiki page, check what's done, see what's not done, ...)
2009-08-07T00:26:45 <ThomasWaldmann> do we have the tagging script yet?
2009-08-07T00:28:00 <dennda> for maint I can write one. for migration I don't see the point
2009-08-07T00:30:04 <ThomasWaldmann> we need to tag the syspages or they won't work
2009-08-07T00:31:13 <dennda> and how do you decide what is a syspage when the data comes from 1.x?
2009-08-07T00:32:52 <ThomasWaldmann> we convert the master wiki content into mm2 backend content
2009-08-07T00:32:59 <ThomasWaldmann> and then we tag all pages?
2009-08-07T00:33:46 <dennda> yes, sure, that's for maintenance. I don't argue this
2009-08-07T00:34:08 <dennda> I thought you were talking about tagging old syspages from a users pages folder
2009-08-07T00:34:44 <dennda> i've no idea how we could discover what has been a syspage and what not reliably (the most reliable thing I see would be to compare it with a list of pagenames, which is prone to error and PITA)
2009-08-07T00:50:25 <ThomasWaldmann> ok, so one of your tasks for tomorrow is to create a english only syspages.xml
2009-08-07T00:50:44 <ThomasWaldmann> (from current 1.9)
2009-08-07T00:51:42 <ThomasWaldmann> ehrm, for today
2009-08-07T00:51:59 <dennda> ok
2009-08-07T00:55:41 <dennda> using 1.9 underlay.tar I assume
2009-08-07T00:59:08 <ThomasWaldmann> untar it and run make pagepacks (in 1.9)
2009-08-07T00:59:18 <ThomasWaldmann> then take the underlay
2009-08-07T01:00:08 <ThomasWaldmann> or rather: first install en, then take underlay
2009-08-07T01:06:22 <dennda> will take a look tomorrow
2009-08-07T01:06:26 <dennda> too late again already
2009-08-07T01:06:27 <dennda> gn
2009-08-07T01:12:59 *** grzywacz has quit IRC
2009-08-07T01:18:22 *** tpfennig has quit IRC
2009-08-07T03:09:18 *** birkenfeld has quit IRC
2009-08-07T03:09:19 *** amartani has quit IRC
2009-08-07T03:11:40 *** birkenfeld has joined #moin-dev
2009-08-07T03:44:30 *** amartani has joined #moin-dev
2009-08-07T06:03:04 <CIA-38> Alexandre Martani <amartani AT gmail DOT com> default * 6548:249b6724c3dc 2.0-storage-editor-amartani/MoinMoin/ (3 files in 3 dirs): Real-time editor: port conflict detection and merge from 1.9. failing on some cases for now.
2009-08-07T08:31:01 <ThomasWaldmann> dennda: it has to be HelpOnAccessControlLists and the page contents needs some fixes
2009-08-07T08:31:30 <ThomasWaldmann> e.g. there is no wikidicts cache any more
2009-08-07T08:31:52 <ThomasWaldmann> the hier. acl part doesn't seem uptodate
2009-08-07T08:32:33 <ThomasWaldmann> "new in 1.6" doesn't make much sense any more in 2.0
2009-08-07T08:36:40 <ThomasWaldmann> think about th Trusted: part of the default acl, compare it with 1.9
2009-08-07T08:38:06 <ThomasWaldmann> remove references to autoadmin
2009-08-07T08:43:44 <ThomasWaldmann> docs/CHANGES update
2009-08-07T08:56:51 <dennda> yes
2009-08-07T08:59:09 <dreimark> moin
2009-08-07T09:07:14 <ThomasWaldmann> moin
2009-08-07T09:11:40 <dennda> moin :-)
2009-08-07T09:15:21 <dreimark> dimazest: please add a CHANGES about your work
2009-08-07T09:18:10 <dreimark> .oO in 2.0-storage tests finished: 1228 passed, 84 failed, 83 skipped in 39.77 seconds
2009-08-07T09:18:52 <ThomasWaldmann> dreimark: is your repo uptodate?
2009-08-07T09:20:10 <ThomasWaldmann> hg tip
2009-08-07T09:21:04 <dreimark> changeset: 6444:dc7c8125067c
2009-08-07T09:21:31 <ThomasWaldmann> does hg up -C change anything?
2009-08-07T09:23:29 <ThomasWaldmann> tests finished: 1798 passed, 2 failed, 91 skipped in 48.69 seconds
2009-08-07T09:23:34 <dreimark> 0 files updated, 0 files merged, 0 files removed, 0 files unresolved
2009-08-07T09:23:39 <dennda> strange
2009-08-07T09:23:49 <dennda> ah ok I thought you already fixed the sourcecode thing
2009-08-07T09:23:51 <dennda> just a sec
2009-08-07T09:24:16 <dennda> we need to find out why. otherwise we may see user problems we cannot explain...
2009-08-07T09:24:21 <dennda> dreimark: nopaste the failures please
2009-08-07T09:24:24 * dreimark prepares a pastebin
2009-08-07T09:25:26 <dreimark> ThomasWaldmann: you have some more skipped than me, why?
2009-08-07T09:25:47 <ThomasWaldmann> i also have some more passed
2009-08-07T09:25:53 <dennda> hehe
2009-08-07T09:27:13 <dennda> tests finished: 2131 passed, 1 failed, 91 skipped in 48.75 seconds
2009-08-07T09:27:27 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6445:fabfd20f5e1b 2.0-storage/MoinMoin/script/migration/backend.py: storage: pep8fix
2009-08-07T09:29:14 <dreimark> http://moinmo.in/ReimarBauer?action=AttachFile&do=view&target=output.txt
2009-08-07T09:30:04 <dennda> wtf is that garbage in between?
2009-08-07T09:30:13 <dennda> dreimark: python, py.test, mercurial versions?
2009-08-07T09:30:54 <dreimark> Mercurial Distributed SCM (version 1.2.1)
2009-08-07T09:31:06 <dreimark> Python 2.6.2 (release26-maint, Apr 19 2009, 01:58:18)
2009-08-07T09:31:28 <dennda> did you prematurely abort the test run?
2009-08-07T09:31:55 <dennda> PawelPacana: ^
2009-08-07T09:32:16 <dreimark> where does py.test tells it version?
2009-08-07T09:32:28 <dennda> it doesn't have an option for that
2009-08-07T09:32:32 <dennda> the bloody thing
2009-08-07T09:34:19 <dreimark> python-codespeack-lib is 0.9.1-3ubuntu1
2009-08-07T09:34:41 <dennda> i have the same py.test then
2009-08-07T09:34:55 <dennda> and the same python
2009-08-07T09:35:06 * dreimark updates mercurial
2009-08-07T09:35:09 <dennda> but hg 1.3
2009-08-07T09:35:18 <dreimark> will have 1.3.1
2009-08-07T09:35:42 <dennda> that still doesn't explain why there's other failures as well
2009-08-07T09:36:42 <dreimark> if I remove wiki and data from MoinMoin/_tests
2009-08-07T09:36:50 <dennda> dreimark: did you stop the tests before they had completely run through?
2009-08-07T09:36:58 <dreimark> is that all to do?
2009-08-07T09:37:01 <dreimark> dennda: no
2009-08-07T09:37:59 <dennda> didn't you test that on a fresh clone?
2009-08-07T09:38:10 <dennda> I'm unsure which files are used
2009-08-07T09:38:17 <dreimark> not yet but soon
2009-08-07T09:38:24 <dreimark> tests finished: 1838 passed, 6 failed, 91 skipped in 58.34 seconds
2009-08-07T09:38:35 <dreimark> result after mercurial update
2009-08-07T09:38:48 <dennda> just do: cd ..; hg clone 2.0-storage test_2.0-storage && cd test_2.0-storage; py.test
2009-08-07T09:39:20 <dennda> that should get rid of all files that aren't there by default
2009-08-07T09:39:46 <dennda> ok, what failures are those?
2009-08-07T09:40:07 <dennda> PawelPacana: there seems to be a problem with the hg backend if it is run on a version < 1.3
2009-08-07T09:40:44 <dreimark> I think that is known limitations, but the code should tell that and not trying to run
2009-08-07T09:41:17 <dreimark> dennda: The failing tests likly because of the missing merge from 1.9
2009-08-07T09:41:38 <dennda> dreimark: why do we have only one failure then?
2009-08-07T09:41:52 <dreimark> because of your filesystems
2009-08-07T09:42:04 <dreimark> and the granularity
2009-08-07T09:42:51 <dennda> that is on zfs then?
2009-08-07T09:42:52 <dreimark> I have the same tests failing as we have had in 1.9 before the comparision was fixed
2009-08-07T09:43:11 <dennda> aha ok
2009-08-07T09:43:21 <dennda> so a merge should fix those problems?
2009-08-07T09:43:24 <dreimark> http://hg.moinmo.in/moin/1.9/rev/27d5d0b875a9
2009-08-07T09:43:28 <dreimark> yes
2009-08-07T09:44:05 <dennda> ok
2009-08-07T09:44:10 <dreimark> the problem is that cache_mtime is float and page_mtime can be integer
2009-08-07T09:44:23 * dennda has no idea
2009-08-07T09:44:28 <dennda> about that
2009-08-07T09:44:33 <ThomasWaldmann> omg
2009-08-07T09:46:25 <dreimark> yeah we are bitten by that shit
2009-08-07T09:46:29 <ThomasWaldmann> dreimark: please remove math import
2009-08-07T09:48:03 <dreimark> ok, can do and compare just int ?
2009-08-07T09:48:49 <dennda> I think that's what already happens if you convert it to an int
2009-08-07T09:48:57 <dennda> i.e. int(3.9999) -> 3
2009-08-07T09:49:32 <dennda> doesn't work for int(3.99999999999999999999999999999999999999999999999) though :-)
2009-08-07T09:49:37 <dreimark> besides that it is because of the type change
2009-08-07T09:50:06 <dennda> yes sure
2009-08-07T09:50:24 *** devilsadvocate has joined #moin-dev
2009-08-07T09:50:29 * dreimark fixes that
2009-08-07T09:58:25 <CIA-38> Reimar Bauer <rb.proj AT googlemail DOT com> default * 4928:6c4c93f896f8 1.9/MoinMoin/datastruct/backends/ (wiki_dicts.py wiki_groups.py): replaced math.floor by int
2009-08-07T10:00:57 <ThomasWaldmann> dreimark: tests are ok for you in 1.9?
2009-08-07T10:01:10 <dreimark> yes, verified
2009-08-07T10:01:44 <ThomasWaldmann> ok, then i merge 1.9>2.0 now
2009-08-07T10:06:34 <dreimark> ThomasWaldmann: we have to find a better solution for that fix otherwise if the granularity is more than 1 sec it will fail again
2009-08-07T10:07:16 <dreimark> that may be a reason for the cache invalidations by changing group pagesd in the past
2009-08-07T10:09:53 <dreimark> (if we know the granularity it can be used for checking)
2009-08-07T10:30:38 <ThomasWaldmann> the only fs i know with more than 1s is fat32 and iirc that has other problems also
2009-08-07T10:31:09 <dreimark> ext2 too
2009-08-07T10:35:11 <ThomasWaldmann> afair 1s
2009-08-07T10:38:39 <dennda> Hm, I logged into 1.9 as superuser, but neither does SystemPagesSetup work nor does LanguageSetup offer me links to install from
2009-08-07T10:38:49 <dennda> (as it clearly states it would)
2009-08-07T10:39:46 <dennda> and it still says Immutable Page for me
2009-08-07T10:40:21 <dennda> superuser = ["dennda"]
2009-08-07T10:40:27 <dreimark> dennda: do you see the action in the text ?
2009-08-07T10:41:13 <dreimark> uptodate underlay?
2009-08-07T10:41:46 <dennda> the one shipped with 1.9 in the repo
2009-08-07T10:41:51 <dennda> let me update the branch
2009-08-07T10:43:19 <dreimark> and in DE mode every user would be superuser (so normally you don't need to configure this)
2009-08-07T10:44:50 <dennda> well, at least with the new underlay languagesetup got an update
2009-08-07T10:45:18 <dreimark> you have run make pagepacks from top level dir ?
2009-08-07T10:45:40 <dreimark> before that (otherwise you don't have pagepacks)
2009-08-07T10:46:04 <dennda> ah ic
2009-08-07T10:47:07 <dennda> dreimark: ok, I did that and it seems to have worked. but still no syspages action
2009-08-07T10:47:20 <dreimark> it is in the text
2009-08-07T10:47:26 <dreimark> not in more actions
2009-08-07T10:47:59 <dreimark> You can install help and system page packages for any language you want to support for your wiki user
2009-08-07T10:48:13 <dreimark> click on "install help and system page packages"
2009-08-07T10:48:42 <dreimark> we should render it as button
2009-08-07T10:48:50 <dennda> ah
2009-08-07T10:49:13 <dennda> Installed MoinMaster page bundle English_all_pages.zip.
2009-08-07T10:49:14 <dennda> :-)
2009-08-07T10:49:45 <dennda> great
2009-08-07T10:51:18 <dennda> thanks
2009-08-07T10:59:17 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 4936:1c8da2f463ee 1.9/MoinMoin/ (6 files in 5 dirs): merged moin/1.8
2009-08-07T11:15:22 <CIA-38> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 6467:b5659e3c25c2 2.0-storage/ (63 files in 13 dirs): merged moin/1.9
2009-08-07T11:26:48 * dreimark has 2.0-storage tests finished: 1853 passed, 1 failed, 91 skipped in 70.83 seconds
2009-08-07T11:26:54 <dennda> great :-)
2009-08-07T11:29:23 <dennda> why is there such a difference
2009-08-07T11:30:22 <ThomasWaldmann> maybe dimazest could have a look at the failing test in 2.0. it looks a bit like the test doesn't expect something the code is doing.
2009-08-07T11:30:37 <dennda> hm I should pull
2009-08-07T11:31:20 <ThomasWaldmann> not sure why that works in 1.9, though
2009-08-07T11:33:17 <dennda> I have more than 2130 tests passing
2009-08-07T11:33:25 <dennda> why does dreimark only have 1853?
2009-08-07T11:38:47 <ThomasWaldmann> some stuff is skipped if you don't have the libs. e.g. jabber.
2009-08-07T11:39:29 <ThomasWaldmann> also, how much pep8 tests it runs depends on file timestamps
2009-08-07T11:40:21 <dennda> ah
2009-08-07T11:44:03 <dennda> Do i need to add my tagging script (which is below MoinMoin/script/maint/ as tagsyspages.py) to some list so it can be invoked with py?
2009-08-07T11:44:10 <dennda> *moin.py
2009-08-07T11:46:22 <ThomasWaldmann> could you write it a bit more generic and name it just tag or setmeta?
2009-08-07T11:46:51 <dennda> sure
2009-08-07T11:47:06 <dennda> I made it operate on an xml file, would that be ok?
2009-08-07T11:48:29 <ThomasWaldmann> why not on the wiki?
2009-08-07T11:49:56 <ThomasWaldmann> it could be just an arg parser, for item in ..., get item, get last rev, modify meta, create rev
2009-08-07T11:49:58 <dennda> can be done as well... what I do now is somewhat hacky though. I clone the data to be tagged to a membackend, tag there by accessing the internal revision metadata data structure and serialize back...
2009-08-07T11:50:28 <dennda> aha so you want new revisions for that, ok
2009-08-07T11:50:55 <ThomasWaldmann> without hack, it won't work otherwise
2009-08-07T11:52:10 <ThomasWaldmann> if that script is run on master wiki, you can make diffs between tagged revs
2009-08-07T11:52:31 <ThomasWaldmann> (once we have a tag display UI)
2009-08-07T11:53:12 <dennda> ok
2009-08-07T11:53:38 <dennda> same question still: do I need to add it somewhere so it can be executed?
2009-08-07T11:53:59 <ThomasWaldmann> we also can generalize that is_syspage, sysp_version stuff later to be just generic tags, like categories also
2009-08-07T11:54:27 <ThomasWaldmann> no, the script stuff just loads it from the dir
2009-08-07T12:05:21 <ThomasWaldmann> dennda: how about a moin xml load/save <xmlfile|->
2009-08-07T12:06:08 <dennda> ju
2009-08-07T12:06:09 <dennda> p
2009-08-07T12:07:33 <dennda> hm
2009-08-07T12:09:25 <ThomasWaldmann> one could even think about a --root option for giving a name for backend lookup
2009-08-07T12:09:40 <dennda> ThomasWaldmann: If I call it setmeta so you can set the metadata of the revision to a given --key and --value, how do I know how I should interpret what I get via value type-wise? That is not that trivial
2009-08-07T12:10:29 <dennda> i could test different things and catch exceptions, but --value 3 will work for int, float and string alike (and when it comes to nested tuples, you can totally forget it anyway)
2009-08-07T12:16:41 <ThomasWaldmann> moin meta --set=<key> --value=<value> or --delete=<key>
2009-08-07T12:17:59 <dennda> yes sure
2009-08-07T12:18:11 <dennda> but when saving that value, how do I know what type it's supposed to be?
2009-08-07T12:19:16 <ThomasWaldmann> and --select=<pattern> with default=all
2009-08-07T12:19:40 <dennda> where <pattern> is...?
2009-08-07T12:20:08 <ThomasWaldmann> a regex
2009-08-07T12:20:27 <ThomasWaldmann> for the item names
2009-08-07T12:20:28 <dennda> i think we need some eval() magic for the type
2009-08-07T12:21:12 <ThomasWaldmann> yeah
2009-08-07T12:26:22 <ThomasWaldmann> btw, i have the slight feeling that the user backend cloning problem you found yday is a shortcoming of clone and the storage api
2009-08-07T12:26:34 <dennda> how so?
2009-08-07T12:26:49 <ThomasWaldmann> it is likely because it all thinks in revision history, and item meta is not revisioned
2009-08-07T12:27:07 <dennda> hm perhaps there's a bug in clone
2009-08-07T12:27:13 <dennda> i will check that
2009-08-07T12:27:31 <ThomasWaldmann> no revision -> no item
2009-08-07T12:27:56 <dennda> ah yes
2009-08-07T12:29:15 <dennda> i see why
2009-08-07T12:44:30 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6468:1620afaf44c9 2.0-storage/MoinMoin/storage/__init__.py: storage: Improve exception message
2009-08-07T12:44:33 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6469:fbc8e56ff90e 2.0-storage/MoinMoin/storage/__init__.py: storage: Fix copy_item to work with revision-less items. Fixes user migration.
2009-08-07T12:47:50 <ThomasWaldmann> did you try that ^^
2009-08-07T12:48:06 <dennda> yes
2009-08-07T12:48:39 <ThomasWaldmann> i mean cloning a user backend
2009-08-07T12:48:59 <dennda> $ ls ../1.9/wiki/data/user/
2009-08-07T12:49:00 <dennda> 1249587866.02.11565
2009-08-07T12:49:12 <dennda> cat instance/user/86/name
2009-08-07T12:49:12 <dennda> 1249587866.02.11565
2009-08-07T12:49:38 <ThomasWaldmann> hmm, i thought clone only clones what history() tells?
2009-08-07T12:50:22 <dennda> nope, source.iteritems() and then copy item
2009-08-07T12:51:23 <ThomasWaldmann> ah, ok, maybe we had that at some point in the past
2009-08-07T12:51:51 <ThomasWaldmann> but oviously copy_item misses tests
2009-08-07T12:51:57 <dennda> jup
2009-08-07T12:54:04 <ThomasWaldmann> ok, so we read items from fs19(underlay), write to fs (e.g.), then we tag run moin meta --set="is_syspage" --value=True
2009-08-07T12:54:57 <ThomasWaldmann> the moin xml --save="underlay_complete.xml" (for example)
2009-08-07T12:55:07 <dennda> you need to run the meta thing twice
2009-08-07T12:55:20 <dennda> --set="syspage_version" --value=1
2009-08-07T12:55:27 <ThomasWaldmann> yeah, when tagging a release
2009-08-07T12:56:24 <ThomasWaldmann> if you have that stuff finished until later, I can modify the page packager
2009-08-07T12:58:34 <dennda> will try
2009-08-07T12:58:50 <ThomasWaldmann> ACL privileges
2009-08-07T12:58:51 <ThomasWaldmann> * Make methods check for the proper privileges.
2009-08-07T12:59:34 <ThomasWaldmann> have a look again. think about whether rename should check for may(create, destination)
2009-08-07T13:00:13 <ThomasWaldmann> i mean, if a user just has write and thus is not allowed to create new pages, shall he be able to rename a page?
2009-08-07T13:00:29 <ThomasWaldmann> or delete a page (== rename to trashbin). Maybe rather not.
2009-08-07T13:00:56 <ThomasWaldmann> # rip out all the underlay stuff -> check?
2009-08-07T13:01:29 <ThomasWaldmann> # Make get_enduser_backend work with new storage configuration approach -> check?
2009-08-07T13:02:03 <ThomasWaldmann> # make sure no acl-protected pages are available through the cache
2009-08-07T13:02:12 <ThomasWaldmann> "cache" is a bit unclear there.
2009-08-07T13:02:50 <ThomasWaldmann> think about interaction of acl wrapper and acl protected group definition pages ... -> check?
2009-08-07T13:03:47 <ThomasWaldmann> # General documentation (docstrings, wikipages, etc) -- Work in progress -> do! :)
2009-08-07T13:04:50 <dennda> I'm only gonna write docs if you promise that there won't be major changes again with respect to stuff that needs to be documented :D
2009-08-07T13:04:55 <ThomasWaldmann> i guess you can kill the "beta releases until midterm", that obviously didn't work :)
2009-08-07T13:05:23 <dennda> like, for example, storage configuration :-)
2009-08-07T13:05:42 <ThomasWaldmann> don't forget the english-only syspages.xml, i don't want to wait 4mins :)
2009-08-07T13:06:24 <dennda> yes
2009-08-07T13:06:31 <ThomasWaldmann> ah, yeah, configuration: clean up multiconfig, think more about configuration.
2009-08-07T13:06:53 <ThomasWaldmann> e.g. i don't think we want to have "content_acl" somehow hardcoded
2009-08-07T13:07:08 <ThomasWaldmann> that stuff needs to be params or so
2009-08-07T13:08:18 <ThomasWaldmann> review the output of the 1.9 -> 2.0 clone, esp. attachment items, and metadata in general.
2009-08-07T13:08:45 <ThomasWaldmann> make a list of what's still todo (in fs19.py maybe)
2009-08-07T13:09:00 <dennda> what do you mean by not hardcoding it? how exactly should it work?
2009-08-07T13:09:14 <ThomasWaldmann> that's what you should think about :)
2009-08-07T13:09:24 <dennda> for one: if the syntax to include an attachment differed in 1.9, do you rewrite that in fs19?
2009-08-07T13:09:27 <dennda> ah ok
2009-08-07T13:09:37 <ThomasWaldmann> i don't want to read docs like "you need to have cfg.content_acl because we expect that to be so"
2009-08-07T13:09:55 <ThomasWaldmann> -> explicit is better than implicit
2009-08-07T13:10:11 <ThomasWaldmann> anyway, gtg/bbl
2009-08-07T13:11:10 <ThomasWaldmann> read through "router improvement ideas" on your page and clean up
2009-08-07T13:11:50 <dennda> will do
2009-08-07T13:38:31 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6470:a4e24e78cec5 2.0-storage/MoinMoin/script/maint/setmeta.py: storage: Introduce maintenance setmeta script
2009-08-07T13:54:45 <dennda> reading from FS Backend is SO much faster than writing
2009-08-07T13:56:38 <dimazest> dreimark: ok
2009-08-07T13:57:45 <dreimark> hi dimazest
2009-08-07T14:02:32 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6471:81cd529a1deb 2.0-storage/MoinMoin/script/migration/xml.py: storage: Add script to load backend from/save backend to xml data
2009-08-07T14:33:27 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6472:03a96f489b76 2.0-storage/MoinMoin/script/maint/setmeta.py: storage: setmeta script: support only setting the meta data for some items whose names match a given regex
2009-08-07T14:34:59 <dennda> hm what I have seen of that search term stuff kinda rocks
2009-08-07T14:49:38 <dennda> really cool
2009-08-07T14:49:46 <dennda> too bad nobody seems to have looked at it :D
2009-08-07T14:49:56 <dreimark> what do you mean?
2009-08-07T14:50:07 <dennda> MoinMoin.search.term
2009-08-07T14:50:56 <dennda> storage based search
2009-08-07T14:53:24 <dreimark> ah johill's implementation
2009-08-07T14:53:50 <dennda> yes
2009-08-07T15:12:52 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6473:3c1ade901329 2.0-storage/MoinMoin/storage/_tests/test_backends.py: storage: Add test case for backend.clone()
2009-08-07T15:25:35 <dennda> looks like it is implemented correctly now :-)
2009-08-07T15:35:35 <dennda> ThomasWaldmann: the bad thing with setmeta is that it creates two new revisions. one for is_syspage and one for syspage_version. Would be good to have that script that only takes the latest revision again...
2009-08-07T15:36:54 * dennda writes a new reducewiki version
2009-08-07T15:48:47 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6474:d5ea06256e26 2.0-storage/MoinMoin/script/maint/reducewiki.py: storage: Add new reducewiki script.
2009-08-07T15:52:26 *** y84 has quit IRC
2009-08-07T15:53:11 <dennda> err
2009-08-07T15:53:31 <dennda> I accidentially hg diff'ed the new syspages as opposed to the old syspages... /me goes and makes a coffee
2009-08-07T15:53:54 * dreimark hates d
2009-08-07T15:54:00 <dreimark> o debug adobe problems
2009-08-07T16:01:44 <dennda> boy why is it pushing that long
2009-08-07T16:01:47 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6475:9c84fa3d0b48 2.0-storage/wiki/syspages.xml: storage: syspages.xml now only contains english pages with syspage metadata
2009-08-07T16:01:54 <dennda> %-)
2009-08-07T16:04:57 *** amartani has quit IRC
2009-08-07T16:06:13 <dennda> hm why does it not preload anymore now
2009-08-07T16:07:59 <dennda> ah
2009-08-07T16:08:16 <dennda> bug in the api
2009-08-07T16:08:19 <dennda> great -_-
2009-08-07T16:16:55 <dennda> ThomasWaldmann: http://paste.pocoo.org/show/133024/ -- I removed the usual moin startup noise
2009-08-07T16:34:55 <dennda> hmmmm
2009-08-07T16:38:19 <dennda> ThomasWaldmann: There's a problem with serialization: If the serialized revision's revno doesn't exactly match what is expected as next_revno, it fails.
2009-08-07T16:38:32 <dennda> (that's the reason why preload is broken atm)
2009-08-07T16:46:22 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6476:bd0bf9c545bc 2.0-storage/MoinMoin/storage/__init__.py: storage: Update storage API docstring. We now support destroy()
2009-08-07T17:04:28 <dreimark> bbl
2009-08-07T17:10:54 <dreimark> dennda: why can I have MoinMoin/storage/_tests/test_backends.py[0]
2009-08-07T17:11:17 <dreimark> there must be tests and they are not skipped
2009-08-07T17:14:05 * dreimark goes home
2009-08-07T17:14:45 <dennda> dreimark: You mean tests aren't run although they should? No idea. Works here...
2009-08-07T17:27:33 <dennda> ThomasWaldmann: http://paste.pocoo.org/show/133031/
2009-08-07T17:27:42 <dennda> that would fix the issue but there'd be a significant problem:
2009-08-07T17:28:30 <dennda> if you have a backend, destroy revs, serialize, unserialize again, urls may have changed (because the empty revs in between wouldn't exist anymore or point to other revs)
2009-08-07T17:29:34 *** tpfennig has joined #moin-dev
2009-08-07T17:29:43 <dennda> that needs more thought
2009-08-07T17:36:56 <dennda> Apart from some non-functional macros and an ever-so-ugly UI the result of the clone looks good, btw
2009-08-07T18:10:28 <dennda> sometimes edit_log_extra / edit_log_comment is set with apparently no value, but perhaps to be that way
2009-08-07T18:26:29 <ThomasWaldmann> dennda: yes, I know
2009-08-07T18:34:36 <dennda> *that has to be that way
2009-08-07T18:35:45 <dennda> ok gibt gutes essen anlaesslich meines vadders geburtstag und die ganze sippe kommt gleich angerannt wenns nach essen riecht... :-)
2009-08-07T19:12:05 *** franklin_ has quit IRC
2009-08-07T19:12:05 *** dreimark has quit IRC
2009-08-07T19:12:43 *** franklin_ has joined #moin-dev
2009-08-07T19:12:43 *** dreimark has joined #moin-dev
2009-08-07T19:43:31 <ThomasWaldmann> ist ja ein dauergeburtstag :D
2009-08-07T19:44:51 * ThomasWaldmann is finished with remote maint
2009-08-07T19:55:40 <dennda> remote maint?
2009-08-07T19:55:46 *** grzywacz has joined #moin-dev
2009-08-07T19:55:53 <dennda> what did you maintain?
2009-08-07T19:56:18 <ThomasWaldmann> some ssl certs for some py software
2009-08-07T19:56:35 <dennda> aka expensive shit
2009-08-07T19:56:55 <ThomasWaldmann> cacert.org :)
2009-08-07T19:57:01 <dennda> ok
2009-08-07T19:57:34 <dennda> if only anybody would acknowledge their trust :)
2009-08-07T19:57:43 * dennda has cacert as well
2009-08-07T20:14:28 * ThomasWaldmann erases disks of moinmo.in-old
2009-08-07T20:23:06 <ThomasWaldmann> http://hg.moinmo.in/moin/2.0-storage/rev/fbc8e56ff90e looks like that was quite a bit insane before
2009-08-07T20:23:38 <dennda> well it was just a bug
2009-08-07T20:24:16 <ThomasWaldmann> that you copy N times the same data? X)
2009-08-07T20:24:49 <ThomasWaldmann> btw, there is a comment missing at the item metadata copying
2009-08-07T20:24:58 <ThomasWaldmann> with the reasoning why it is done like that
2009-08-07T20:25:00 <dennda> you don't
2009-08-07T20:25:16 <dennda> It's in the except clause
2009-08-07T20:25:49 <ThomasWaldmann> ok, so it only gets the same item a thousand times :)
2009-08-07T20:25:59 <dennda> jup
2009-08-07T20:26:52 <ThomasWaldmann> the question is why item meta is only copied for fresh items.
2009-08-07T20:27:05 <ThomasWaldmann> what's the pros and cons.
2009-08-07T20:30:15 <dennda> to answer that question we need to be aware of the different metadata entries that may be there
2009-08-07T20:30:32 <dennda> tbh, except user stuff nothing comes to my mind immediately
2009-08-07T20:33:40 <ThomasWaldmann> so, assume we have users and dst.clone(src), do we really want to keep the old stuff if it is same id?
2009-08-07T20:34:40 <dennda> you wouldn't expect that at least I guess
2009-08-07T20:34:50 <dennda> so it's somewhat counter-intuitive currently
2009-08-07T20:35:36 <ThomasWaldmann> also, if we would put some stuff from current rev into item metadata (kind of cache it there), it should be also rather the latest stuff, not the old
2009-08-07T20:36:56 <dennda> we're not doing that, but yes
2009-08-07T20:37:37 <ThomasWaldmann> not yet at least :)
2009-08-07T20:37:52 <dennda> any arguments against changing it?
2009-08-07T20:39:20 <ThomasWaldmann> well, you should add above arguments to a comment there and change it
2009-08-07T20:39:40 <ThomasWaldmann> so we at least will know why it is like that when we want to change it back
2009-08-07T20:40:52 <ThomasWaldmann> and mv setmeta meta, s/key/setkey/ and add rmkey or so
2009-08-07T20:42:02 <dennda> aye
2009-08-07T20:44:44 <ThomasWaldmann> that init stuff you do in the xml script, shouldn't that be done at another place?
2009-08-07T20:45:04 *** tpfennig has quit IRC
2009-08-07T20:46:18 <dennda> you mean pull it up to the parent class or something? likely
2009-08-07T20:48:14 <ThomasWaldmann> well, i guess almost all scripts want to be able to use storage
2009-08-07T20:48:36 <ThomasWaldmann> so the request init should be similar to when a normal request runs
2009-08-07T20:49:28 <ThomasWaldmann> btw, for what do you want to use the reducewiki?
2009-08-07T20:50:16 <ThomasWaldmann> (or: do you know what its main application was?)
2009-08-07T20:50:40 <dennda> a) there was such a script before, so I thought better add it again, b) when using the meta script to tag syspages you need to run it twice, which results in two revisions being added, with duplicate content just for adding some simple metadata key/value pair
2009-08-07T20:50:53 <dennda> i guess the idea was to thin out a wiki
2009-08-07T20:51:07 <ThomasWaldmann> ah, how to almost kill firefox: just look at http://hg.moinmo.in/moin/2.0-storage/rev/9c84fa3d0b48
2009-08-07T20:51:18 <dennda> I won't, been there, done that :-)
2009-08-07T20:51:21 <dennda> hg diff is nice as well
2009-08-07T20:51:40 <dennda> files > $size shouldn't be shown
2009-08-07T20:53:29 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6477:d57f321f98bb 2.0-storage/MoinMoin/storage/__init__.py: storage: When copying an item, use the copied item's metadata if there has already been an item present with the same name in the target backend
2009-08-07T20:55:47 <ThomasWaldmann> dennda: btw, "monotone"
2009-08-07T20:55:59 <dennda> i actually looked that up
2009-08-07T20:56:42 <dennda> seems one is the math term and the other one the more colloquial one
2009-08-07T20:56:48 <ThomasWaldmann> dict.leo.org gives "monotone" in the math sense
2009-08-07T20:56:49 <dennda> at least if it's tagged correctly
2009-08-07T20:56:52 <dennda> yes
2009-08-07T20:56:55 <dennda> dict.cc as well
2009-08-07T20:57:26 <ThomasWaldmann> you know how to insult a mathematician?
2009-08-07T20:58:10 <dennda> no
2009-08-07T20:59:15 <ThomasWaldmann> call him convergent. then he'll know you think he's bounded and monotone.
2009-08-07T20:59:52 <dennda> %-)
2009-08-07T21:00:54 <dimazest> dreimark: with py.test version 0.9.2 i have MoinMoin/search/_tests/test_search.py[19] ............F..F...
2009-08-07T21:01:05 <dimazest> it is my linux box
2009-08-07T21:01:08 <dennda> dimazest: we have one group failure in 2.0. can you take a look?
2009-08-07T21:01:23 <dimazest> dennda: sure
2009-08-07T21:02:17 <dimazest> dreimark: in development i use py==1.0.0b1
2009-08-07T21:03:15 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6478:3f267de7135e 2.0-storage/MoinMoin/script/maint/reducewiki.py: storage: pep8 fix
2009-08-07T21:04:16 <ThomasWaldmann> i guess I'll add some \n to the xml at some harmless places
2009-08-07T21:04:26 <dennda> please
2009-08-07T21:04:43 <dennda> I bet the missing whitespace doesn't really save much space anyway
2009-08-07T21:08:39 <dennda> btw: suggested pencils down on tuesday %-)
2009-08-07T21:15:07 <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4954:bd1bbdb40a5f 1.9-xapian-dmilajevs/MoinMoin/ (36 files in 4 dirs): Automated merge with http://hg.moinmo.in/moin/1.9
2009-08-07T21:15:08 <CIA-38> Dmitrijs Milajevs <dimazest@gmail.com> default * 4955:0603dfa4cc49 1.9-xapian-dmilajevs/MoinMoin/search/ (Xapian.py queryparser.py): Groups2009: queryparser.TitleSearch.xapian_term was refactored. Xapian.Index.prefixMap attribute was removed.
2009-08-07T21:15:15 <ThomasWaldmann> btw migration in the sense used in moin means moving from one version to another (a later one)
2009-08-07T21:15:26 <ThomasWaldmann> so, how's the xml stuff related to migration?
2009-08-07T21:16:19 <dennda> the rationale behind that was that you could use it for migration
2009-08-07T21:16:36 <dennda> i didn't know the term was already used consistently in another context
2009-08-07T21:23:53 *** amartani has joined #moin-dev
2009-08-07T21:28:25 <CIA-38> Christopher Denter <moin GUESSWHAT the DASH space DASH station ROUNDTHING com> default * 6479:f806a66a02a3 2.0-storage/MoinMoin/ (3 files in 2 dirs): storage: Rename setmeta script to just meta. Add -r option that 'deletes' the given key by not copying it to the new revision
2009-08-07T21:29:51 <dennda> ThomasWaldmann: how do we fix preload now?
2009-08-07T21:35:11 <ThomasWaldmann> i just wanted ask whether preload is broken :)
2009-08-07T21:37:14 <ThomasWaldmann> dennda: it doesn't do anything for migration, it just restores what's in the xml. so, either create a new xml script dir, or use maint, but dont keep it in migration.
2009-08-07T21:43:23 <dennda> didn't you read what I wrote? (wrt preload)
2009-08-07T21:48:56 <ThomasWaldmann> so you serialized revno > 0?
2009-08-07T21:49:03 <dennda> yes
2009-08-07T21:49:49 <dennda> if you want to fix it to 0 that'd mean that unserialize can only work on clean backends
2009-08-07T21:51:22 <ThomasWaldmann> well, it could work at least in one easy to get situation. currently it doesn't work at all, except if you have N revs already there.
2009-08-07T21:52:28 <dennda> yes, but before doing some fix I'd have to redo perhaps I thought I'd bring the issue up :-)
2009-08-07T21:55:45 <dennda> so fix it to 0?
2009-08-07T21:56:22 <ThomasWaldmann> fix what to 0?
2009-08-07T21:58:11 <dennda> the revnos in syspages
2009-08-07T21:58:24 <dennda> otherwise preload cannot work in the current code
2009-08-07T21:58:51 <ThomasWaldmann> you mean in the xml?
2009-08-07T21:59:23 <dennda> yes
2009-08-07T22:01:32 <ThomasWaldmann> wait
2009-08-07T22:02:19 <ThomasWaldmann> http://pastebin.com/f78c760dc when you redo the xml with rev 0, try that cs
2009-08-07T22:06:49 <ThomasWaldmann> but please run tests and commit code change in separate cs
2009-08-07T22:07:05 <dennda> the question is how to properly redo that with rev 0
2009-08-07T22:07:25 <dennda> hm what does clone do...
2009-08-07T22:13:14 <dennda> ThomasWaldmann: hm, imho unserialize needs to be revision number agnostic
2009-08-07T22:13:25 <dennda> otherwise you just can't provide syspages
2009-08-07T22:13:51 <dennda> each syspage a user has potentially has a different maximum revision number
2009-08-07T22:14:40 <dennda> you're gonna get RevisionMismatchErrors / RevisionAlreadyExistsErrors when unserializing syspages to a wiki that already contains syspages
2009-08-07T22:14:50 <dennda> the same is true for all other kinds of serialized data
2009-08-07T22:14:51 <ThomasWaldmann> yeah, but that does not mean that unserialize doesnt need revno
2009-08-07T22:15:03 <dennda> well it would use item.next_revno
2009-08-07T22:15:13 <dennda> with the downside that links may need to change
2009-08-07T22:16:41 <ThomasWaldmann> http://pastebin.com/f79a475e6
2009-08-07T22:18:34 <dennda> :-)
2009-08-07T22:18:42 <dennda> can you commit that?
2009-08-07T22:18:57 <dennda> what the heck is going on outside
2009-08-07T22:19:05 <dennda> sonds like someone started WW3
2009-08-07T22:19:21 <ThomasWaldmann> iirc it didn't work somehow
2009-08-07T22:19:28 <ThomasWaldmann> or at least it needs more work
2009-08-07T22:32:23 <dennda> ok will check that tomorrow. kinda tired now
2009-08-07T22:34:18 <ThomasWaldmann> look in the repo first :)
2009-08-07T22:34:45 <dennda> what should I see there?
2009-08-07T22:35:01 <dennda> your potential changesets? :-)
2009-08-07T22:35:36 <ThomasWaldmann> maybe
2009-08-07T22:35:47 <dennda> ok
2009-08-07T23:31:53 <dennda> n8
2009-08-07T23:56:23 <ThomasWaldmann> gn dennda
2009-08-07T23:56:24 <ThomasWaldmann> File "/home/tw/w/mm20storage/MoinMoin/storage/backends/fs.py", line 517, in _get_revision_metadata
2009-08-07T23:56:28 <ThomasWaldmann> datastart = struct.unpack('!L', datastart)[0]
2009-08-07T23:56:30 <ThomasWaldmann> error: unpack requires a string argument of length 4 MoinMoin: MoinMoinChat/Logs/moin-dev/2009-08-07 (last edited 2009-08-06 22:15:01 by IrcLogImporter)
