2008-03-22T00:00:13 <dennda> yep, reading the whole chapter now
2008-03-22T00:00:24 <johill> you can stop when they get into this N_ stuff
2008-03-22T00:00:36 <johill> we don't do that, we do the first way
2008-03-22T00:01:08 <dennda> ok
2008-03-22T00:37:52 <johill> hm
2008-03-22T00:38:04 <johill> anybody know if I can have a nested function access variables from the outer scope?
2008-03-22T00:38:07 <johill> def a():
2008-03-22T00:38:10 <johill> i.e.
2008-03-22T00:38:11 <johill> var1 = 7
2008-03-22T00:38:14 <johill> def b():
2008-03-22T00:38:18 <johill> var1 += 1
2008-03-22T00:38:21 <johill> b()
2008-03-22T00:38:26 <johill> assert var1 == 8
2008-03-22T00:38:34 <johill> like that it doesn't work
2008-03-22T00:38:38 <johill> but is it possible?
2008-03-22T00:40:37 <johill> guess not
2008-03-22T00:40:40 <ThomasWaldmann> johill: are you working on the query parser?
2008-03-22T00:41:46 <johill> not really
2008-03-22T00:41:56 <johill> I'm extending parse_quoted_separated to be able to handle it
2008-03-22T00:42:53 <johill> so in a way, yes
2008-03-22T00:43:46 <dreimark> probably yes, because it is an interpreter
2008-03-22T00:43:59 <dreimark> but I don't know the trick
2008-03-22T00:44:53 <johill> dreimark: the internet says you can't
2008-03-22T00:50:15 * ThomasWaldmann looks at the 1.6 qp
2008-03-22T00:53:52 <dreimark> http://iwiwdsmi.blogspot.com/2008/01/python-variable-scope-notes.html
2008-03-22T00:54:16 <dreimark> last example
2008-03-22T00:54:59 <dreimark> johill:
2008-03-22T00:55:03 <johill> that'd work
2008-03-22T01:06:04 <dreimark> we should check if the python parser has problems with % in a line
2008-03-22T01:07:43 <johill> hm?
2008-03-22T01:09:56 <johill> good thing I have so many parse_quoted_separated tests :)
2008-03-22T01:21:15 <johill> phew
2008-03-22T01:21:19 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3399:b6d727809985 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py):
2008-03-22T01:21:19 <CIA-39> introduce wikiutil.parse_quoted_separated_ext
2008-03-22T01:21:19 <CIA-39> This new function can now handle space-separated arguments
2008-03-22T01:21:19 <CIA-39> (pass None as separator) as well as bracketing. Includes tests.
2008-03-22T01:21:41 <johill> that will make a query parser trivial :)
2008-03-22T01:22:34 <johill> almost
2008-03-22T01:22:38 <johill> one thing is still missing
2008-03-22T01:23:22 <johill> xorAxAx: still around?
2008-03-22T01:23:55 <johill> I'm wondering whether regex:(abc) really is right
2008-03-22T01:24:02 <johill> imho it should be regex:"(abc)"
2008-03-22T01:32:02 <dreimark> http://moinmo.in/ReimarBauer/UnderConstruction
2008-03-22T01:32:15 <dreimark> first converted version, needs some fine tuning
2008-03-22T01:32:50 <dreimark> will continue later
2008-03-22T01:34:26 <johill> wow
2008-03-22T01:34:36 <johill> not bad
2008-03-22T01:34:38 <johill> dennda: check that out
2008-03-22T01:34:45 <johill> if you read German, that is
2008-03-22T01:34:56 <dennda> I do
2008-03-22T01:35:19 <dennda> dreimarks link?
2008-03-22T01:35:24 <johill> http://moinmo.in/ReimarBauer/UnderConstruction
2008-03-22T01:35:25 <johill> yeah
2008-03-22T01:35:51 <dennda> looks promising
2008-03-22T01:35:59 <johill> wow this is confusing
2008-03-22T01:35:59 <dennda> I will definitely have to check it
2008-03-22T01:36:13 <johill> my parser converts a:b:c to ('a', 'b=c')
2008-03-22T01:36:14 <johill> wtf?
2008-03-22T01:36:54 <johill> oh.
2008-03-22T01:37:00 * johill hides
2008-03-22T01:38:10 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3400:b12aba864312 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): fix bug in parse_quoted_separated_ext and add tests that caught the bug
2008-03-22T01:43:34 <dreimark> the converter has had a problem with lines containing %
2008-03-22T01:43:48 <dreimark> i will check those in some hours
2008-03-22T01:44:31 <dreimark> and anchors are mostly broken
2008-03-22T01:49:26 <dreimark> good night
2008-03-22T01:49:40 <johill> night
2008-03-22T02:02:19 <dennda> http://moinmo.in/ReimarBauer/UnderConstruction?action=AttachFile&do=get&target=MoinMoinArchitecture.png
2008-03-22T02:02:27 <dennda> that's not supposed to be a class diagram, is it?
2008-03-22T02:06:10 <dennda> Or is it?
2008-03-22T02:06:12 <dennda> hm
2008-03-22T02:06:16 <dennda> lemme read on
2008-03-22T02:11:14 <dennda> nevermind
2008-03-22T02:12:34 <ThomasWaldmann> dennda: that's a very rough overview
2008-03-22T02:14:14 <dennda> yep
2008-03-22T02:17:59 * ThomasWaldmann found a search bug
2008-03-22T02:18:30 <johill> not surprising
2008-03-22T02:24:18 <johill> wow
2008-03-22T02:24:22 <johill> I cleaned up the function :)
2008-03-22T02:25:25 <dennda> umm
2008-03-22T02:25:53 <dennda> When saving changes to a page, why is it necessary to pass the revision number? Can't you just ommit it?
2008-03-22T02:27:34 <johill> where?
2008-03-22T02:27:50 <johill> I suppose backends support history modification :)
2008-03-22T02:28:29 <dennda> dreimark's link states that
2008-03-22T02:29:31 <dennda> johill: Look for the PageEditor.saveText() section of the text
2008-03-22T02:29:54 <dennda> History modification sounds ugly, doesn't it? :)
2008-03-22T02:31:00 <dennda> You know, you can't just travel back in time and change the past... You'd cause serious paradoxes :)
2008-03-22T02:33:21 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3401:d75e838f4b72 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): wikiutil.parse_quoted_separated_ext: allow multi-key stuff
2008-03-22T02:37:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3402:4b94ff6a4073 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): wikiutil.parse_quoted_separated_ext: make bracketing support indicate which bracket was used
2008-03-22T02:37:12 <johill> ok
2008-03-22T02:37:21 <johill> with this we can implement the search parser
2008-03-22T02:37:57 <johill> oh, wait, no
2008-03-22T02:38:00 <johill> darn
2008-03-22T02:39:12 * johill thinks
2008-03-22T02:42:41 * dennda falls asleep
2008-03-22T02:42:48 <dennda> good night
2008-03-22T02:48:28 <ThomasWaldmann> gn dennda
2008-03-22T03:00:44 <johill> ok
2008-03-22T03:00:45 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3403:d479e393e318 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): wikiutil.parse_quoted_separated_ext: allow for prefixes
2008-03-22T03:00:48 <johill> *now* we can implement the query parser
2008-03-22T03:07:25 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2593:22f5a256238e 1.6/ (3 files in 3 dirs): disabled xapian_stemming by default, added some comments to problematic code enabled by it
2008-03-22T03:48:09 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3404:e979e3cad972 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): parse_quoted_separated_ext: fix unclosed quotes
2008-03-22T03:59:48 <johill> yay
2008-03-22T03:59:53 <johill> ThomasWaldmann: I fixed the queryparser :)
2008-03-22T05:26:54 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3405:6f5e09f14525 1.7/MoinMoin/ (_tests/test_wikiutil.py wikiutil.py): parse_quoted_separated_ext: allow multiple/different quoting characters
2008-03-22T05:26:55 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3406:ccdd2bbf76f7 1.7/MoinMoin/search/ (_tests/test_search.py queryparser.py): reimplement and thereby fix QueryParser
2008-03-22T09:23:21 <ThomasWaldmann> moin
2008-03-22T09:29:29 <xorAxAx> johill: well, as i said, the old version didnt use parenthesis to group its argument
2008-03-22T10:31:22 <dreimark> moin
2008-03-22T10:32:47 <dreimark> did one have seen a hg diff showing
2008-03-22T10:32:59 <dreimark> abort: unknown revision '#w��J/�Ƥz�/���['!
2008-03-22T10:37:44 <dreimark> first hg pull gives a traceback and another one seems to
2008-03-22T10:37:55 <dreimark> however i do use better a new clone
2008-03-22T10:39:05 <xorAxAx> yes, thats a local corruption
2008-03-22T10:46:00 <dreimark> zlib.error: Error -5 while decompressing data at traceback
2008-03-22T10:46:38 <dreimark> I do file this bug report
2008-03-22T10:51:43 <xorAxAx> ?
2008-03-22T10:52:24 <dreimark> the traceback tells to file it as bugreport
2008-03-22T10:52:33 <xorAxAx> wellll
2008-03-22T10:52:37 <xorAxAx> you have a local corruption
2008-03-22T10:52:41 <dreimark> no
2008-03-22T10:52:47 <xorAxAx> then what did you do?
2008-03-22T10:52:52 <xorAxAx> on the command line
2008-03-22T10:52:56 <dreimark> hg pull
2008-03-22T10:53:06 <dreimark> the second time it vanished
2008-03-22T10:53:17 <xorAxAx> did you try with a fresh clone?
2008-03-22T10:53:17 <dreimark> no all seems ok
2008-03-22T10:53:45 <dreimark> I tried with the "somehow" broken one
2008-03-22T10:53:50 <dreimark> the fresh clone is ok
2008-03-22T10:54:00 <xorAxAx> so its a corruption :)
2008-03-22T10:54:35 <dreimark> I hope it is, but the corruption is gone now
2008-03-22T11:31:19 <dreimark> traceback for Page(request, 'StartSeite').send_page(msg=u'Done',emit_headers=False)
2008-03-22T11:34:12 <johill> xorAxAx: explain?
2008-03-22T11:34:22 <johill> xorAxAx: the help says it uses it to group
2008-03-22T11:34:55 <xorAxAx> johill: it didnt back in the old sane days
2008-03-22T11:35:18 <johill> it has to though if you want any decent syntax
2008-03-22T11:35:20 <xorAxAx> well, my impression is that the parser was severely broken in 2006
2008-03-22T11:35:32 <xorAxAx> johill: ? whats the problem with '"'?
2008-03-22T11:35:38 <johill> ??
2008-03-22T11:35:45 <johill> that's a quote
2008-03-22T11:35:47 <johill> not a group
2008-03-22T11:35:59 <xorAxAx> well, grouping in terms of query syntax, not regex syntax
2008-03-22T11:36:15 <xorAxAx> and you showed an example where it stopped accepting the regex at the end of )
2008-03-22T11:36:18 <xorAxAx> which is broken
2008-03-22T11:37:06 <johill> imho it isn't
2008-03-22T11:37:15 <johill> what it did was brokne
2008-03-22T11:37:38 <johill> but treating () specially is pretty much required if you want to allow grouping
2008-03-22T11:38:33 <xorAxAx> ?
2008-03-22T11:38:37 <xorAxAx> i dont understand :)
2008-03-22T11:38:47 <xorAxAx> hmm
2008-03-22T11:39:11 <johill> a (b or c)
2008-03-22T11:39:19 <xorAxAx> yes
2008-03-22T11:39:37 <johill> making a(b or c) work any differently is magic
2008-03-22T11:39:44 <xorAxAx> well
2008-03-22T11:39:55 <xorAxAx> a here was "regex:" thats not a valid expr
2008-03-22T11:40:03 <johill> or you'd have to require a ( b or c )
2008-03-22T11:40:08 <xorAxAx> no
2008-03-22T11:40:08 <johill> which is crappy syntax
2008-03-22T11:40:14 <xorAxAx> you just need a sane grammar
2008-03-22T11:40:40 <dreimark> bbl
2008-03-22T11:40:46 <johill> no, this is for users
2008-03-22T11:40:46 <xorAxAx> (maybe its time to use a parser generator to get something sane :))
2008-03-22T11:40:50 <xorAxAx> johill: ?
2008-03-22T11:41:02 <johill> imho that's too confusing
2008-03-22T11:41:12 <johill> plus it's not extensible
2008-03-22T11:41:19 <xorAxAx> what is too confusing?
2008-03-22T11:41:33 <johill> having all those "valid" vs "invalid" cases
2008-03-22T11:41:39 <xorAxAx> ?
2008-03-22T11:42:29 <johill> I don't see why a regex prefix should make it be treated any differently
2008-03-22T11:43:36 <johill> and then you treat "j:(a)" as a regular search but when in the future somebody adds a j prefix it's treated specially
2008-03-22T11:45:34 <johill> and why should there be a syntax difference between 'no space before opening' and 'no space before closing' parenthesis?
2008-03-22T11:46:26 <johill> or what if the user enters regex:(a b), do you try to fix up the regex syntax error too?
2008-03-22T11:46:55 <johill> certainly, I will not implement such things. feel free
2008-03-22T11:48:30 <xorAxAx> well, you can require quotes then
2008-03-22T11:49:01 * xorAxAx prepares for the osterlauf
2008-03-22T11:53:02 <xorAxAx> johill: or does that still look ugly?
2008-03-22T11:55:20 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3407:c3aa12fcdf12 1.7/MoinMoin/_tests/test_wikiutil.py: add a parser test about leading None arguments
2008-03-22T11:55:41 <johill> sorry was away for a sec. of course you can require quotes, but imnsho that is the sane thing to do in all cases
2008-03-22T12:31:37 <dreimark> moin grzywacz
2008-03-22T12:34:14 <grzywacz> hi dreimark
2008-03-22T12:43:24 <johill> hi
2008-03-22T12:43:27 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3408:7fd796b2e573 1.7/MoinMoin/search/ (_tests/test_search.py queryparser.py): queryparser: catch exceptions, raise exception on '""'
2008-03-22T12:44:37 <dreimark> how do you find a revision which changed e.g. theme.__init__.searchform
2008-03-22T12:44:42 <dreimark> without reading all?
2008-03-22T12:45:05 <johill> hg annotate?
2008-03-22T12:45:14 <johill> and then hg annotate for the parent etc
2008-03-22T12:45:18 <dreimark> ok
2008-03-22T12:45:32 <johill> or just go to the view in hg web
2008-03-22T12:46:09 <dreimark> I missed that annotate action on the hg web
2008-03-22T12:46:59 <johill> :)
2008-03-22T12:49:21 <dennda> g'morning
2008-03-22T12:49:34 <johill> hi dennda
2008-03-22T12:50:19 <johill> huh
2008-03-22T12:50:21 <dennda> Hi johill
2008-03-22T12:50:26 <dennda> I really wonder when you slept
2008-03-22T12:50:35 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3711:1ae16a7fee86 1.7-storage-hwendel/MoinMoin/ (5 files in 4 dirs): merge parser improvements from main
2008-03-22T12:51:30 <johill> heh
2008-03-22T12:51:33 * dreimark tries to fix searchform
2008-03-22T12:51:40 <johill> irregularly
2008-03-22T12:52:36 <johill> dreimark: what's wrong with it?
2008-03-22T12:53:36 <dennda> ;)
2008-03-22T12:54:59 <kikka> Moinmoin
2008-03-22T12:55:13 <dreimark> 'pagename_quoted': wikiutil.quoteWikinameURL(self.request.page.page_name),
2008-03-22T12:55:47 <dreimark> AttributeError: 'NoneType' object has no attribute 'page_name'
2008-03-22T12:56:19 <johill> hm
2008-03-22T12:56:21 <johill> just remove it all?
2008-03-22T12:56:26 <johill> i.e. use action=""
2008-03-22T12:56:27 <dreimark> that was added have a valid form
2008-03-22T12:56:51 <dreimark> bug fix of MoinMoinBugs/MoveAttachmentNotWorkingWithModPython (thanks to Boleslaw Kulbabinski) (ported from 1.6)
2008-03-22T12:56:55 <johill> action="" isn't valid?
2008-03-22T12:57:10 <dreimark> empty string is not valid
2008-03-22T12:57:50 <johill> we have that in another place too
2008-03-22T12:58:05 <johill> at the drop-down action form
2008-03-22T12:58:34 <dreimark> hmm, should have been fixed everywhere by bollek
2008-03-22T12:58:56 <johill> hm
2008-03-22T12:59:03 <johill> quoteWikiNameURL seems wrong here
2008-03-22T12:59:17 <johill> are you sure that works *at*all*?
2008-03-22T12:59:57 <johill> quoteWikinameURL gives a URL-quoted string
2008-03-22T13:00:04 <johill> but that should, in this case, be teh responsibility of the browser
2008-03-22T13:00:09 <johill> so that should just be wikiutil.escape()
2008-03-22T13:00:37 <johill> afaict quoteWikinameURL is only valid when we create a redirect
2008-03-22T13:02:53 <dreimark> yeah,( two bugs )
2008-03-22T13:03:11 <johill> it seems to work which confuses me a bit
2008-03-22T13:03:25 <johill> ah
2008-03-22T13:03:26 <johill> no
2008-03-22T13:03:27 <johill> it doesn't
2008-03-22T13:03:45 <johill> http://localhost:8080/%C3%B6
2008-03-22T13:03:58 <johill> (/ö)
2008-03-22T13:04:02 <johill> click on search you get
2008-03-22T13:04:51 <johill> hrm
2008-03-22T13:04:55 <dreimark> argh
2008-03-22T13:05:07 <dreimark> Nameerror
2008-03-22T13:05:10 <dreimark> global name '_' is not defined
2008-03-22T13:05:34 <johill> where's that now?
2008-03-22T13:05:53 <dreimark> thas was me
2008-03-22T13:05:55 <johill> does gecko do some magic here?
2008-03-22T13:05:56 <dreimark> soory
2008-03-22T13:06:08 <johill> /%C3%B6?action=fullsearch&...
2008-03-22T13:06:49 <johill> it seems right
2008-03-22T13:06:53 <johill> but conceptually I don't think it is
2008-03-22T13:08:04 <johill> dreimark: I suppose we'll have to use '' if no page object, though I wonder when you have no page object
2008-03-22T13:08:08 <johill> can you show a backtrace?
2008-03-22T13:10:24 <dreimark> the example is quite easy
2008-03-22T13:10:33 <dreimark> from MoinMoin.request import request_cli
2008-03-22T13:10:39 <dreimark> request = request_cli.Request()
2008-03-22T13:10:44 <dreimark> rom MoinMoin.Page import Page
2008-03-22T13:10:48 <dreimark> pagename = u'StartSeite'
2008-03-22T13:10:57 <dreimark> Page(request, pagename).send_page(msg=u'Done', emit_headers=False)
2008-03-22T13:11:03 <dreimark> (besides the typos)
2008-03-22T13:12:14 <johill> oh
2008-03-22T13:12:16 <johill> I was wrong
2008-03-22T13:12:25 <johill> uri attributes in html want to be url-encoded
2008-03-22T13:13:38 <johill> so the code is correct except for that bug
2008-03-22T13:13:50 <dreimark> ok, then we should check if wikiutil.escape was used where we need quoteWikinameURL
2008-03-22T13:15:24 <dreimark> http://moinmo.in/ReimarBauer?action=AttachFile&do=view&target=traceback.txt
2008-03-22T13:15:38 <dreimark> johill:
2008-03-22T13:16:40 <johill> yaeh makes sense with request_cli
2008-03-22T13:17:19 <johill> but that is a bit abusive
2008-03-22T13:18:14 <johill> I think you should do something like r = reqcli.Request(page='somepage'), r.run()
2008-03-22T13:18:39 <johill> but then again somebody figured with cli we don't use run?
2008-03-22T13:18:57 <johill> which is fairly dumb given the huge amount of request code...
2008-03-22T13:24:22 <johill> maybe we should do the wsgi thing instead
2008-03-22T13:25:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3409:1c5bb02539cf 1.7/MoinMoin/theme/__init__.py: fix actionsmenu form's action attribute
2008-03-22T13:34:25 <johill> oi
2008-03-22T13:34:33 <johill> millions of 'You cannot call add_msg() after send_title()'
2008-03-22T13:38:40 <dreimark> send_title knows the pagename for cli too
2008-03-22T13:42:35 <dreimark> johill: we have only to use d['page'].page_name
2008-03-22T13:43:15 <johill> heh ok
2008-03-22T13:43:43 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3410:81b8a5bb8711 1.7/MoinMoin/Page.py: send_page sucks. fix bug with invalid highlight REs
2008-03-22T13:50:39 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3411:b9b958b46ed1 1.7/MoinMoin/theme/__init__.py: theme.__init__: bugfix for searchform and actionsMenu
2008-03-22T13:55:16 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3412:91d38a790d8e 1.7/MoinMoin/request/ (8 files): request framework: handle connection aborts gracefully
2008-03-22T13:57:46 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3714:b27e85740010 1.7-storage-hwendel/MoinMoin/ (Page.py theme/__init__.py): merge main
2008-03-22T13:57:47 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3717:9d5e7e1af13a 1.7-storage-hwendel/MoinMoin/ (9 files in 2 dirs): merge fixes from main
2008-03-22T14:08:08 <dennda> um
2008-03-22T14:08:11 <dennda> did I get that right:
2008-03-22T14:08:42 <dennda> moinmoin determines groups by parsing special wiki-pages?
2008-03-22T14:08:49 <dennda> sounds *quite* unusual
2008-03-22T14:14:00 <johill> yes, it does
2008-03-22T14:16:09 <dennda> ok
2008-03-22T14:21:54 <johill> quite a cool concept
2008-03-22T14:22:00 <johill> cf. EditingOnMoinMaster
2008-03-22T14:22:13 <johill> there is a feature request for allowing other groups
2008-03-22T14:22:18 <johill> e.g. from ldap
2008-03-22T14:25:33 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3413:fc0f007111e9 1.7/MoinMoin/_tests/test_Page.py: test_Page: testSendPage added
2008-03-22T14:27:59 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 2594:50d308749a18 1.6/MoinMoin/theme/__init__.py: heme.__init__: bugfix for searchform and actionsMenu (backport from 1.7)
2008-03-22T14:32:01 <johill> which, I think, the best way to implement would be to make the 'is user in group' function a function of the user stoage backend
2008-03-22T14:33:22 <johill> in fact, that could be done with arbitrary metadata
2008-03-22T14:44:10 <johill> and should, much more efficient
2008-03-22T14:57:48 * dennda likes the createTicket() & checkTicket() thing
2008-03-22T14:59:08 * dreimark spends next time more time on the converter
2008-03-22T14:59:18 <dreimark> hopefully it is not soon
2008-03-22T14:59:57 <dennda> dreimark: how recent is the UnderConstruction-Text?
2008-03-22T15:00:37 <dreimark> quite recent
2008-03-22T15:01:15 <dennda> cool
2008-03-22T15:01:28 <dennda> I'm just a bit sad that the storage thing isn't explained :)
2008-03-22T15:01:37 <dreimark> currently I do test the plugin examples with 1.7
2008-03-22T15:02:21 <dreimark> storage isn't published yet.
2008-03-22T15:02:35 <johill> dreimark: "ist. Zur Drucklegung dieses Buches befindet sich die Version 1.7 kurz vor der Veröffentlichung und an der Version 1.7 wird bereits gearbeitet "
2008-03-22T15:03:04 <dreimark> missed to change the second on to .18
2008-03-22T15:03:17 <dreimark> 1.8 first was 1.6
2008-03-22T15:03:53 <dennda> johill: noticed that too :)
2008-03-22T15:05:19 <dreimark> well, it doesn't currently look like it will be published on dead trees
2008-03-22T15:05:54 <johill> I assume otherwise you wouldn't have published it there :)
2008-03-22T15:06:36 <dreimark> before I would have had to wait untill it was published
2008-03-22T15:07:30 <johill> right
2008-03-22T15:07:41 <dreimark> I have had a contract to use it on MM too. (by adding a note describing the book)
2008-03-22T15:07:49 <johill> dreimark: imho you should remove Page.getPagePath()
2008-03-22T15:08:06 <johill> we know it won't be availabe much longer :)
2008-03-22T15:10:42 <dreimark> done (commented off)
2008-03-22T15:11:33 <johill> dreimark: the example for logfile.eventlog is weird
2008-03-22T15:11:39 <johill> what's the may.read doing there?
2008-03-22T15:11:45 <johill> should it be in the loop?
2008-03-22T15:12:14 <johill> oh n/m I think I understand somewhat
2008-03-22T15:12:32 <johill> mind you, ?action=info works on pages you can't read I think
2008-03-22T15:13:11 <johill> oh, no, it doesn't
2008-03-22T15:15:26 <dreimark> that example counts only one page and the user needs access to that page
2008-03-22T15:16:47 <dreimark> behaviour of action=info is correct. otherwise it can expose acl protected pages
2008-03-22T15:17:53 <johill> yeah you're right
2008-03-22T15:17:57 <johill> I was confused for a minute
2008-03-22T15:20:38 <johill> "Wiki Seiten von MoinMoin werden in form von HTML " -> Form
2008-03-22T15:21:09 <dennda> Under which license is the text released?
2008-03-22T15:21:45 <dreimark> dennda: what do you think?
2008-03-22T15:22:22 <dennda> no idea, creative commons?
2008-03-22T15:22:23 <dreimark> there is only one license possible :)
2008-03-22T15:22:30 <dennda> gpl fdl?
2008-03-22T15:22:36 <dreimark> gpl
2008-03-22T15:22:41 <dennda> ok
2008-03-22T15:23:25 <dreimark> you are right I do add later on a note on top,
2008-03-22T15:24:51 <dennda> Good. Wanted to change a few things myself
2008-03-22T15:25:18 <dreimark> fine, no edit conflicts please :)
2008-03-22T15:25:36 <dennda> Isn't there some kind of lock?
2008-03-22T15:25:42 <dreimark> fine -> ok
2008-03-22T15:26:02 <dreimark> we have "warn" on MM
2008-03-22T15:26:48 <dreimark> one has to recognize. (I missed it sometimes too)
2008-03-22T15:27:42 <dennda> hehe :)
2008-03-22T15:28:15 <dreimark> it makes sense to work on it logged in
2008-03-22T15:31:14 <dennda> user css
2008-03-22T15:31:16 <dennda> cool
2008-03-22T15:34:01 <dreimark> :)
2008-03-22T15:39:30 <johill> in the storage branch
2008-03-22T15:39:39 <johill> should we write a 'flat filesystem backend' that doesn't do revisions?
2008-03-22T15:39:56 <johill> then we could ship help/system pages in the sources rather than having moinmaster
2008-03-22T15:40:05 <johill> otoh, moinmaster is good for translators, hm
2008-03-22T15:41:06 <xorAxAx> sigh
2008-03-22T15:41:30 <xorAxAx> vcs integration doesnt mean to abandon all wiki ideas
2008-03-22T15:41:48 <xorAxAx> also note that nobody stops you to use a real editor
2008-03-22T15:44:27 <johill> you don't get the point
2008-03-22T15:44:50 <johill> the point is that it's terrible to make docs actually be in lock-step with the code
2008-03-22T15:45:09 <johill> we *try* to at least do it on major revisions but that's not too great
2008-03-22T15:45:36 <xorAxAx> umm
2008-03-22T15:45:54 <xorAxAx> the major problem is that people cant document the 1.7 branch, right?
2008-03-22T15:46:03 <xorAxAx> i sketched useful solutions 2 years ago
2008-03-22T15:46:15 <xorAxAx> that do not abandon the wiki principle and work out pretty fine for other projects
2008-03-22T15:46:16 <johill> that's a side effect of the problem
2008-03-22T15:46:24 <xorAxAx> which other problems are there?
2008-03-22T15:46:31 <johill> the main problem is that you cannot commit code and doc changes at the same time
2008-03-22T15:46:53 <xorAxAx> yes, true. but even then, nobody commits unittests and code at the same time
2008-03-22T15:46:54 <johill> e.g. the storage branch can't be documented properly either
2008-03-22T15:47:02 <xorAxAx> i dont see why
2008-03-22T15:47:11 <xorAxAx> you can start with textfiles on the branch
2008-03-22T15:47:21 <xorAxAx> if you really want the changeset coupling
2008-03-22T15:47:28 <johill> xorAxAx: you're saying "because people make mistakes we don't them to do allow the right thing"
2008-03-22T15:47:29 <xorAxAx> otherwise extend the existing wiki pages
2008-03-22T15:47:31 <johill> ?!
2008-03-22T15:47:46 <johill> s/allow //
2008-03-22T15:47:48 <xorAxAx> johill: no, i am not. just that that kind of discipline is lacking as well
2008-03-22T15:48:02 <johill> yes
2008-03-22T15:48:09 <johill> but you're generalising that to: it doens't matter
2008-03-22T15:48:17 <xorAxAx> no, i see the usefulness
2008-03-22T15:48:57 <johill> if I do textfiles then I can't use wiki markup (unless I'm perfect and don't make mistakes in it), and I have another extra step to take to make them into real docs later either way
2008-03-22T15:49:04 <johill> of course it's possible
2008-03-22T15:49:09 <xorAxAx> but dropping a wiki for that goal is not really nice
2008-03-22T15:49:13 <johill> but development tools should be about making things easy
2008-03-22T15:49:48 <johill> well personally, I would start by separating all translations into their own wiki
2008-03-22T15:49:57 <xorAxAx> how about this approach:
2008-03-22T15:50:00 <johill> and make the english one be part of the source code
2008-03-22T15:50:10 <xorAxAx> moinmaster is kept in the repo and merged like code
2008-03-22T15:50:31 <xorAxAx> and then you have some backend that retains some canonical form to ease merging and committing
2008-03-22T15:51:00 <xorAxAx> hmm
2008-03-22T15:51:07 <xorAxAx> but the history property is lacking
2008-03-22T15:51:10 <johill> the question isn't much how to get wiki stuff into the hg repo but rather the other way around
2008-03-22T15:51:27 <xorAxAx> johill: the users only get a flat snapshot
2008-03-22T15:51:32 <johill> yes later
2008-03-22T15:51:35 <johill> but for moinmaster I mean
2008-03-22T15:51:45 <xorAxAx> hmm, i guess an hg backend is the only viable solution i currently see
2008-03-22T15:51:46 <johill> if I make a change and commit it to the hg repo
2008-03-22T15:51:49 <johill> it has to show up there
2008-03-22T15:52:32 <johill> but if moinmaster runs on the live source code we get into trouble with merging often
2008-03-22T15:52:41 <johill> if another user just changes a typo I need to merge etc
2008-03-22T15:53:04 <xorAxAx> hmm, where is trouble?
2008-03-22T15:53:20 <xorAxAx> well, the problem is that the merging needs to be data aware
2008-03-22T15:53:21 <johill> well already if three of us are hacking we merge all the time
2008-03-22T15:53:30 <johill> no I mean just the hg merge step
2008-03-22T15:53:35 <xorAxAx> you cant use your usual diff3/patch stack to merge the data
2008-03-22T15:53:42 <xorAxAx> yes, the hg merge step
2008-03-22T15:53:44 <johill> sure you can
2008-03-22T15:53:50 <xorAxAx> describe how
2008-03-22T15:53:56 <johill> you just use a backend that doesn't save a file for each revision
2008-03-22T15:54:04 <xorAxAx> hmm, yes, the hg backend
2008-03-22T15:54:22 <johill> it doesn't even need to be one
2008-03-22T15:54:28 <johill> a flat file backend would be sufficient
2008-03-22T15:54:34 <johill> you'd use that for within the source code
2008-03-22T15:54:38 <johill> and you'd use the regular backend on moinmaster
2008-03-22T15:54:56 <johill> periodically you can convert moinmaster from regular to flat and merge
2008-03-22T15:55:23 <johill> although of course an hg backend would be nicer
2008-03-22T15:56:26 <xorAxAx> hmm, the problem of an hg backend is that the history of hg is not linear
2008-03-22T15:57:09 <johill> or you could formulate that as ".. is that the history in moin is linear" ;)
2008-03-22T15:57:57 <xorAxAx> yes, the merging approach might work out - to have a flatbackend and some merging scripts
2008-03-22T15:58:31 <johill> it may mean that the full history isn't always visible on moinmaster
2008-03-22T15:58:37 <johill> but imho that doesn't matter too much in this case
2008-03-22T15:59:26 <xorAxAx> you can use the comment field to attribute namens
2008-03-22T15:59:32 <xorAxAx> s/namens/names/
2008-03-22T15:59:53 <xorAxAx> hmm, this would even end up to have something that git has and hg not - the distinction of committer and author
2008-03-22T16:00:14 <johill> heh
2008-03-22T16:00:37 <johill> not necessarily
2008-03-22T16:01:18 <xorAxAx> hm?
2008-03-22T16:01:20 <johill> I wouldn't even bother storing the author in a flat backend
2008-03-22T16:01:33 <xorAxAx> the author is encoded in the hg history
2008-03-22T16:01:54 <xorAxAx> you only need some mergepoint info which is easy to generate when doing the moinmaster export
2008-03-22T16:02:09 <johill> not sure we're talking about the same thing
2008-03-22T16:02:32 <xorAxAx> well, the idea is to have moinmaster in an hg repo, one file per page
2008-03-22T16:02:35 <johill> I would run moinmaster as the regular current stuff and periodically call a script that pulls it into flat storage in the source code hg tree
2008-03-22T16:02:44 <xorAxAx> yes
2008-03-22T16:02:53 <johill> ah I see
2008-03-22T16:03:00 <johill> no, I wouldn't bother pulling each revision
2008-03-22T16:03:03 <johill> I would just pull a snapshot
2008-03-22T16:03:18 <johill> and then commit the snapshot
2008-03-22T16:03:37 <johill> merge with whatever else changed, and reset moinmaster
2008-03-22T16:03:38 <xorAxAx> hmm?
2008-03-22T16:03:42 <xorAxAx> "reset"?
2008-03-22T16:03:59 <johill> though I guess pulling each revision would make that history trackable
2008-03-22T16:04:07 <xorAxAx> you need merge point info in the flat snapshot to merge it backe
2008-03-22T16:04:09 <xorAxAx> s/e$//
2008-03-22T16:04:21 <xorAxAx> (i.e. page revision numbers on the mm side)
2008-03-22T16:04:31 <xorAxAx> you can also use wikisync for that :)
2008-03-22T16:04:46 <xorAxAx> s/on the/from the/
2008-03-22T16:05:09 <johill> well what I'm saying is that I wouldn't bother doing it that way
2008-03-22T16:05:29 <johill> I would just take the last moinmaster revision, merge with the current file, commit, and put that to moinmaster as the new revision
2008-03-22T16:05:50 <xorAxAx> how do you want to "merge" here?
2008-03-22T16:05:52 <johill> (if changed)
2008-03-22T16:05:55 <johill> just text based merge
2008-03-22T16:05:59 <xorAxAx> that wouldnt be a diff3 :)
2008-03-22T16:06:13 <xorAxAx> so you would end up generating context diffs and edit them
2008-03-22T16:06:13 <dennda> dreimark: you wrote that book / chapter?
2008-03-22T16:06:53 <johill> xorAxAx: ah I see what you mean, yes, my original idea was to wipe moinmaster every time and then you'd have info on that implicitly (oldest and newest version on master)
2008-03-22T16:07:22 <xorAxAx> hmm
2008-03-22T16:07:34 <xorAxAx> that would make it non-distributed
2008-03-22T16:07:44 <xorAxAx> so its a bottleneck to do the merging
2008-03-22T16:07:47 <johill> yeah
2008-03-22T16:08:00 <johill> storing a list of revision numbers for merging is fine I guess
2008-03-22T16:08:05 <johill> makes it more efficient too
2008-03-22T16:08:14 <johill> no need to pull the text, just check "has this changed"
2008-03-22T16:09:26 <dreimark> dennda: I wrote that chapter
2008-03-22T16:09:31 <dennda> ok
2008-03-22T16:09:43 <dreimark> and I do write a book with some members of pyCologne
2008-03-22T16:11:00 <dreimark> http://www.pycologne.de/
2008-03-22T16:11:45 * dennda wants pyKoblenz ;)
2008-03-22T16:19:00 <johill> I would still separate out the languages so they can be distributed separatedly
2008-03-22T16:19:48 <johill> xorAxAx: what does wikisync require from the local wiki/storage?
2008-03-22T16:20:08 <johill> i.e. target
2008-03-22T16:20:14 <xorAxAx> ?
2008-03-22T16:20:20 <johill> oh wait it's bidi
2008-03-22T16:20:23 <xorAxAx> you mean its data structures?
2008-03-22T16:20:32 <johill> well
2008-03-22T16:20:49 <johill> could I use wikisync with a flat file storage that doesn't present revisions?
2008-03-22T16:21:08 <xorAxAx> no
2008-03-22T16:21:36 <johill> what we just outlined works though, I'm trying to figure out where the difference is
2008-03-22T16:21:36 <xorAxAx> it always attempts to do a 3-way merge
2008-03-22T16:21:51 <xorAxAx> and transmits only the necessary delta to reconstruct the same information on both sides
2008-03-22T16:22:40 <johill> ah
2008-03-22T16:23:08 <johill> I see the difference
2008-03-22T16:23:51 <dreimark> he, moin-dev log double lines are fixed
2008-03-22T16:24:08 <johill> dreimark: yeah when the bot's twin was killed ;)
2008-03-22T16:25:02 <dennda> boy
2008-03-22T16:25:04 <dennda> call me sherlock
2008-03-22T16:28:47 <johill> xorAxAx: but a history-hg-backend would work, right?
2008-03-22T16:29:00 <johill> i.e. one that stores flat files but reads hg history for a file when asked
2008-03-22T16:29:32 <xorAxAx> sure
2008-03-22T16:29:35 <johill> or does wikisync more than one change and assume it'll be present?
2008-03-22T16:29:45 <xorAxAx> wikisync changes metadata, yes
2008-03-22T16:29:46 <johill> s/more/do more/
2008-03-22T16:29:54 <johill> yeah but more than once?
2008-03-22T16:29:58 <xorAxAx> yes
2008-03-22T16:30:04 <xorAxAx> for every sync
2008-03-22T16:30:10 <xorAxAx> (potentially)
2008-03-22T16:30:22 <johill> well I really mean more than once per page I mean
2008-03-22T16:30:48 <xorAxAx> yes, of course (not in the same sync run, though)
2008-03-22T16:30:57 <johill> yeah in one sync run
2008-03-22T16:31:05 <johill> so we just make the script check that the hg working dir is clean
2008-03-22T16:31:32 <johill> and possibly make the backend verify that
2008-03-22T16:31:41 <johill> then you run sync and commit yourself
2008-03-22T16:31:53 <johill> that'd work, no?
2008-03-22T16:32:41 <johill> a hg-history-read backend should be fairly simple but I don't want to go into one that commits too
2008-03-22T16:33:15 <xorAxAx> sync from where to where?
2008-03-22T16:33:40 <johill> master/local
2008-03-22T16:33:49 <johill> I thought sync is bidirectional
2008-03-22T16:40:11 <johill> sync is borked
2008-03-22T16:42:01 <johill> funky. my gnome-termianl process is so large it can no longer fork
2008-03-22T16:46:07 <johill> xorAxAx: what was that about most people and tests? I hear there's a saying about glas houses...
2008-03-22T16:46:41 <xorAxAx> johill: yes, the wikisync stuff isnt well tested :
2008-03-22T16:49:36 <xorAxAx> but i didnt need to skip any unittests to make wikisync work :)
2008-03-22T16:50:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3414:898a544b8f74 1.7/MoinMoin/action/SyncPages.py: make syncpages not crash on remote errors
2008-03-22T16:50:23 <johill> you could easily have written a lot of tests then
2008-03-22T16:50:26 <johill> test():
2008-03-22T16:50:29 <johill> py.test.skip()
2008-03-22T16:50:31 <johill> assert False
2008-03-22T16:52:04 <xorAxAx> hmm?
2008-03-22T16:53:05 <johill> are there tests at all?
2008-03-22T16:53:20 <johill> "Synchronisation finished. Look below for the status messages."
2008-03-22T16:53:29 <johill> oww. I synced master, that'll be a long page
2008-03-22T16:53:57 <xorAxAx> :)
2008-03-22T16:54:33 <xorAxAx> btw, there is still no positive evidence that the cache bugs have been fixed
2008-03-22T16:54:48 <xorAxAx> but i guess you will soon find out :)
2008-03-22T16:54:59 <johill> I'm not going to sync towards master
2008-03-22T16:55:06 <johill> I think ThomasWaldmann would kill us both if that went wrong
2008-03-22T16:55:20 <xorAxAx> well, he wrote the buggy code :)
2008-03-22T16:55:28 <xorAxAx> and he has backups
2008-03-22T16:55:30 <johill> or I could create a new username
2008-03-22T16:55:35 <johill> so it's easy to mass-revert
2008-03-22T16:55:35 <xorAxAx> ?
2008-03-22T16:56:06 <xorAxAx> as long as the code writes out sane stuff, yes
2008-03-22T16:56:13 <johill> uh huh
2008-03-22T16:56:18 <johill> Got a list of 759 local and 2000 remote pages. This results in 2000 different pages over-all.
2008-03-22T16:56:25 <johill> that doesn't seem right
2008-03-22T16:56:32 <johill> in a second sync run
2008-03-22T16:56:35 <xorAxAx> did you just push your local pages? :)
2008-03-22T16:57:00 <johill> no, I did 'down'
2008-03-22T16:57:13 <xorAxAx> which part is not right, the 2000 i presume?
2008-03-22T16:57:14 <johill> (and I don't have that many local pages anyway)
2008-03-22T16:57:20 <johill> no the 2000 is the number of pages on master
2008-03-22T16:57:24 <johill> it should be 2000 local too
2008-03-22T16:57:29 <xorAxAx> ah
2008-03-22T16:57:29 <johill> since I just synced
2008-03-22T16:57:35 <johill> before it says
2008-03-22T16:57:36 <johill> said
2008-03-22T16:57:37 <johill> Got a list of 0 local and 2000 remote pages. This results in 2000 different pages over-all.
2008-03-22T16:57:48 <johill> yes, I can see that :)
2008-03-22T16:58:14 <xorAxAx> well, check the log. maybe it gives you a hint where the remining ones are :)
2008-03-22T16:58:18 <dreimark> you won't get attachments
2008-03-22T16:59:13 <xorAxAx> attachments arent counted
2008-03-22T16:59:28 <xorAxAx> except if somebody changed the pagelist to return attachments :)
2008-03-22T16:59:44 <johill> well
2008-03-22T16:59:45 <johill> $ grep 'Synchronising page' /tmp/SyncTest | wc -l
2008-03-22T16:59:45 <johill> 1903
2008-03-22T16:59:54 <johill> that was the original
2008-03-22T17:00:00 <johill> not 2000 as it said
2008-03-22T17:00:41 <xorAxAx> also note that the counts might be different than the no. of pages that will be synced afterwards
2008-03-22T17:01:01 <johill> still doesn't explain why the second time it syncs 700 pages
2008-03-22T17:01:59 <johill> it tried 1619 in the second run
2008-03-22T17:02:10 <johill> (couldn't because of ACLs, all of them failed)
2008-03-22T17:02:13 <xorAxAx> johill: are you logged in?
2008-03-22T17:02:28 <xorAxAx> because it behaves differently if the pages are not writable
2008-03-22T17:02:37 <johill> locally I'm not logged in
2008-03-22T17:02:43 <johill> but I can write
2008-03-22T17:02:54 <johill> except I can't update the pages after sync because they have ACLs forbidding it
2008-03-22T17:03:04 <xorAxAx> so you should see exceptions?
2008-03-22T17:03:04 <johill> remmeber, it did sync 1903 pages
2008-03-22T17:03:07 <johill> they're on my filesystem
2008-03-22T17:03:15 <johill> no
2008-03-22T17:03:18 <xorAxAx> hmm
2008-03-22T17:03:20 <johill> no exceptions
2008-03-22T17:03:22 <johill> it's just borked
2008-03-22T17:03:27 <johill> without an explanation :)
2008-03-22T17:03:36 <xorAxAx> can you check your data dir?
2008-03-22T17:03:43 <johill> what do you want to know?
2008-03-22T17:04:09 <johill> I used master/ as prefix
2008-03-22T17:04:10 <johill> $ ls -d wiki/data/pages/master\(2f\)*|wc -l
2008-03-22T17:04:10 <johill> 1575
2008-03-22T17:04:29 <johill> so it says it synced 1903 but did 1575
2008-03-22T17:04:32 <xorAxAx> well, check if all of those pages have one rev
2008-03-22T17:04:42 <johill> sorry no
2008-03-22T17:04:46 <johill> it synced 2000
2008-03-22T17:05:04 <johill> (some the ending \) was not good, some pages start with a quoted letter)
2008-03-22T17:05:19 <johill> $ ls wiki/data/pages/master\(2f*/current|wc -l
2008-03-22T17:05:19 <johill> 1903
2008-03-22T17:05:24 <johill> aeh??
2008-03-22T17:05:47 <johill> johannes:~/Projects/moin/moin-1.7$ cat wiki/data/pages/master\(2f*/current | grep -v 00000001
2008-03-22T17:05:50 <johill> johannes:~/Projects/moin/moin-1.7$
2008-03-22T17:05:59 <johill> so they all have one revision
2008-03-22T17:06:03 <johill> all 1903
2008-03-22T17:06:16 <xorAxAx> yes, as expected
2008-03-22T17:07:05 <xorAxAx> hmm, the 2000 vs 1903 is a bit weird, the code really tries to operate on the pages it counted (so it should be 2000)
2008-03-22T17:07:21 <johill> want a list of the missing ones?
2008-03-22T17:07:23 <xorAxAx> ah, btw
2008-03-22T17:07:29 <johill> some have dirs but no current file
2008-03-22T17:07:30 <xorAxAx> grep for warnings
2008-03-22T17:07:33 <xorAxAx> in the logs
2008-03-22T17:07:46 <xorAxAx> it should give you a reason in every case it failed
2008-03-22T17:07:52 <xorAxAx> or even errors
2008-03-22T17:08:07 <johill> johannes:~$ grep '!' /tmp/SyncTest
2008-03-22T17:08:07 <johill> johannes:~$
2008-03-22T17:08:39 <xorAxAx> hmm
2008-03-22T17:08:42 <johill> oh
2008-03-22T17:08:47 <johill> those pages were deleted on master
2008-03-22T17:08:52 <johill> e.g. 'xegon10'
2008-03-22T17:08:59 <xorAxAx> ah
2008-03-22T17:09:19 <johill> so that explains the missing 97 pages
2008-03-22T17:09:22 <xorAxAx> yes, it needs a list of all pages, even deleted ones
2008-03-22T17:09:32 <xorAxAx> not for the down sync, but in the normal case
2008-03-22T17:09:42 <johill> this is the down sync, but yeah
2008-03-22T17:09:46 <johill> still doesn't explain the second run
2008-03-22T17:10:22 <johill> it needs deelted for down-sync too
2008-03-22T17:10:31 <johill> to say "whoops, you edited a page that upstream deleted"
2008-03-22T17:10:44 <xorAxAx> self.log_status(ActionClass.WARN, _("The page %s was deleted remotely but changed locally."), (sp.name, ))
2008-03-22T17:10:47 <xorAxAx> yes
2008-03-22T17:11:29 <johill> the log really needs to be less verbose
2008-03-22T17:11:35 <johill> or sorted so that warnings are more apparent
2008-03-22T17:11:47 <johill> right now you get two entries for each successful page!
2008-03-22T17:11:59 <ThomasWaldmann> re
2008-03-22T17:12:10 <xorAxAx> johill: hehe, yes
2008-03-22T17:12:38 <ThomasWaldmann> iirc, it is allowed to just use & for the query args separator, even for the first (instead of ?), right?
2008-03-22T17:12:58 <xorAxAx> ThomasWaldmann: no
2008-03-22T17:12:58 <johill> ThomasWaldmann: I don't think so
2008-03-22T17:13:21 <johill> ThomasWaldmann: the bug is that it does ?a=b&c=d?e=f
2008-03-22T17:13:21 <ThomasWaldmann> (or was it vice versa?)
2008-03-22T17:13:28 <johill> and that's not correct either
2008-03-22T17:13:30 <ThomasWaldmann> johill: sure
2008-03-22T17:13:38 <johill> xorAxAx: it insists ..
2008-03-22T17:13:41 <johill> third run says "Got a list of 759 local and 2000 remote pages. This results in 2000 different pages over-all."
2008-03-22T17:13:55 <ThomasWaldmann> ok, i'll have a look at it
2008-03-22T17:14:06 <xorAxAx> johill: well, the 759 comes from the local pagelist function
2008-03-22T17:14:12 <xorAxAx> johill: did you disable your local cache?
2008-03-22T17:14:33 <johill> I just did ./moin.py after adding an interwiki name (on which it insists even for down)
2008-03-22T17:14:43 <xorAxAx> so you likely see broken cache behaviour
2008-03-22T17:14:59 <xorAxAx> johill: did you restart moin.py?
2008-03-22T17:15:00 <johill> good thing you're working so actively on fixing it
2008-03-22T17:15:12 <xorAxAx> johill: i would rip it out if i was working on it
2008-03-22T17:15:13 <johill> not so far
2008-03-22T17:15:35 <johill> xorAxAx: it's not impossible to implement caching. your whining doesn't help
2008-03-22T17:16:04 <xorAxAx> i am not whining, just pointing out that this issue is somewhere else
2008-03-22T17:16:11 <xorAxAx> likely
2008-03-22T17:16:27 <xorAxAx> and i really dislike the ignorance of this lowerlevel issue
2008-03-22T17:16:29 <johill> you've been whining about the cache for a hugely long time
2008-03-22T17:16:37 <johill> btw
2008-03-22T17:16:39 <johill> "Got a list of 759 local and 2000 remote pages. This results in 2000 different pages over-all."
2008-03-22T17:16:41 <xorAxAx> in this context
2008-03-22T17:16:42 <johill> after a restart
2008-03-22T17:16:47 <xorAxAx> well, that rules it out
2008-03-22T17:17:54 <xorAxAx> johill: you could check what the diff between l_pages and r_pages is
2008-03-22T17:18:06 <ThomasWaldmann> dreimark: what exactly is this theme fix fixing?
2008-03-22T17:18:09 <johill> xorAxAx: caching the page list is useful and necessary because getdir() is not cached on any other level. your complaining about it really is effectively just whining
2008-03-22T17:18:19 <ThomasWaldmann> (would be nice to have THAT in the comment)
2008-03-22T17:18:22 <johill> same thing after deleting on-disk caches too
2008-03-22T17:18:31 <xorAxAx> johill: i am not talking about the pagelist cache
2008-03-22T17:18:36 <xorAxAx> that was added by nir years ago
2008-03-22T17:18:42 <xorAxAx> but about the new page cache
2008-03-22T17:19:11 <johill> I assume the diff between r_pages and l_pages is what it tries to sync
2008-03-22T17:19:20 <johill> doesn't seem to have any pattern
2008-03-22T17:19:31 <johill> but obviously all the local pages exist since it says
2008-03-22T17:19:36 <johill> "Skipped page master/PreferensiSintax because of no write access to local page."
2008-03-22T17:19:54 <xorAxAx> whats the len of l_pages before it does the filteirng?
2008-03-22T17:20:55 * johill checks
2008-03-22T17:21:02 <johill> 759
2008-03-22T17:21:15 <johill> $ ls wiki/data/pages/|wc -l
2008-03-22T17:21:15 <johill> 2047
2008-03-22T17:22:18 <xorAxAx> then have a look at wikisync like 329
2008-03-22T17:22:20 <xorAxAx> 349
2008-03-22T17:22:28 <xorAxAx> it only really calls getPageList
2008-03-22T17:22:34 <xorAxAx> with a custom filter
2008-03-22T17:23:02 <xorAxAx> ah, if you may not write, it doesnt add the page
2008-03-22T17:23:08 <xorAxAx> so it makes some sense :)
2008-03-22T17:23:22 <xorAxAx> so there are 759 pages where the acls re missing
2008-03-22T17:23:34 <xorAxAx> the bug here is that it writes acl lines without anything complaining
2008-03-22T17:24:47 <xorAxAx> which is a way to elevate your permissions from write to admin, effectively
2008-03-22T17:25:10 <xorAxAx> (not very severe imho :))
2008-03-22T17:26:11 <johill> ??
2008-03-22T17:26:39 <johill> for one, I cannot see any write check
2008-03-22T17:26:48 <xorAxAx> line 324
2008-03-22T17:26:53 <johill> secondly, it tells me that I cannot write on 759 pages
2008-03-22T17:27:04 <xorAxAx> hmm
2008-03-22T17:27:20 <xorAxAx> oh oh
2008-03-22T17:27:26 <xorAxAx> it uses the wrong name for the acl check
2008-03-22T17:27:42 <xorAxAx> substitute normalised_name by page_name
2008-03-22T17:27:44 <johill> no wait it tells me I cannot write 1619
2008-03-22T17:27:46 <xorAxAx> and try again :)
2008-03-22T17:27:49 <johill> pages
2008-03-22T17:29:52 <johill> Got a list of 284 local and 2000 remote pages. This results in 2000 different pages over-all.
2008-03-22T17:30:18 <xorAxAx> so, better.
2008-03-22T17:30:27 <johill> 1619 I cannot write to
2008-03-22T17:30:43 <johill> that makes 1903
2008-03-22T17:30:47 <johill> that makes some sense
2008-03-22T17:31:02 <johill> but the message is dumb
2008-03-22T17:31:09 <johill> there aren't 2000 different pages
2008-03-22T17:31:16 <xorAxAx> indeed
2008-03-22T17:31:19 <johill> just because you did the acl filtering at that level doesn't mean that the page doesn't exist
2008-03-22T17:32:45 <xorAxAx> and the 7xx - 284 pages is the no. of pages that your local underlay differs from moinmaster
2008-03-22T17:32:58 <starshine> actually, I was part of a project once where if people had no rights to a section the preferred response was to not count them
2008-03-22T17:33:01 <johill> no
2008-03-22T17:33:08 <starshine> they weren't supposed to know they exist :)
2008-03-22T17:33:21 <johill> that can't be true
2008-03-22T17:33:27 <johill> oh underlay, hm, maybe
2008-03-22T17:33:30 * johill removes underlay
2008-03-22T17:33:56 <starshine> except for LocalBadContent fairly few people modify underlay on their own...
2008-03-22T17:34:09 <starshine> but I bet some don't bother to unpack a new one when they upgrade
2008-03-22T17:34:21 <johill> starshine: yeah we've seen that
2008-03-22T17:34:30 <johill> starshine: I use underlay differently or not at all ;)
2008-03-22T17:34:49 <johill> xorAxAx: even w/o underlay it says "Got a list of 284 local and 2000 remote pages. This results in 2000 different pages over-all."
2008-03-22T17:34:59 <xorAxAx> johill: you need to revert the code fix
2008-03-22T17:35:10 <xorAxAx> do you want to check it in or should i?
2008-03-22T17:35:17 <johill> I only have 47 local pages
2008-03-22T17:35:26 <johill> revert?
2008-03-22T17:35:38 <xorAxAx> you modified one line of code, didnt you?
2008-03-22T17:35:41 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2595:490f961b3552 1.6/ (MoinMoin/formatter/text_html.py docs/CHANGES): fix interwiki links using correct query arg separator for the case that there are query args in the interwiki map entry and additionally given in the link markup
2008-03-22T17:35:48 <johill> yes why should I revert that?
2008-03-22T17:36:04 <xorAxAx> to see a 47 or 0 or whatever instead of 7xx
2008-03-22T17:36:37 <johill> that would be fine but 284 doesn't make sense
2008-03-22T17:36:48 <xorAxAx> what did you just remove?
2008-03-22T17:36:52 <johill> unless some of the pages I synced from master are actually writable
2008-03-22T17:36:58 <johill> I only set data-underlay_dir = None
2008-03-22T17:36:58 <xorAxAx> of course, 284
2008-03-22T17:37:08 <xorAxAx> thats a pretty fine number
2008-03-22T17:37:09 <johill> more likley 284-47
2008-03-22T17:37:14 <xorAxAx> or that :)
2008-03-22T17:37:24 <xorAxAx> same magnitude still
2008-03-22T17:37:28 <xorAxAx> s/fine/expected/
2008-03-22T17:37:33 <johill> 10%
2008-03-22T17:37:51 <johill> ok still
2008-03-22T17:37:59 <xorAxAx> ok, i commit the fix
2008-03-22T17:38:00 <johill> can you fix it to not report those pages that are identical
2008-03-22T17:38:11 <johill> which I can't write
2008-03-22T17:38:14 <johill> right now it logs them all
2008-03-22T17:38:14 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3415:cb22f4c4a0fd 1.7/MoinMoin/formatter/text_html.py: fix interwiki links using correct query arg separator for the case that there are query args in the interwiki map entry and additionally given in the link markup
2008-03-22T17:38:41 <xorAxAx> johill: now it shouldnt log or count any nonwritable pages anymore
2008-03-22T17:38:48 <xorAxAx> after this fix
2008-03-22T17:39:08 <johill> it does
2008-03-22T17:39:15 <johill> "Skipped page master/HilfeZumEditieren because of no write access to local page."
2008-03-22T17:39:22 <johill> because you remove it from the local list right away
2008-03-22T17:39:25 <johill> so it thinks that it changed
2008-03-22T17:39:32 <johill> I mean, thinks that it's completely new
2008-03-22T17:39:35 <dreimark> it is snowing
2008-03-22T17:39:49 <johill> and then tries to sync it into the same page and gets upset about it
2008-03-22T17:39:56 <starshine> hi dreimark :)
2008-03-22T17:40:13 <johill> xorAxAx: filtering permissions there seems to be wrong
2008-03-22T17:40:27 <xorAxAx> johill: ok, then i remove the check :)
2008-03-22T17:40:47 <johill> or you keep them on a separate list
2008-03-22T17:41:04 <johill> but that's not necessary since you already check permissions a level higher
2008-03-22T17:41:08 <xorAxAx> and i wont remove the log msg
2008-03-22T17:41:26 <xorAxAx> becuase otherwise debugging will be hard
2008-03-22T17:41:47 <johill> well yes
2008-03-22T17:41:47 <xorAxAx> and you should fix your filters instead of relying on that write-permission problem ignorance :)
2008-03-22T17:42:24 <johill> but without the initial acl filter it won't try to sync pages that are identical even if I cannot write to them
2008-03-22T17:42:37 <dreimark> hi starshine
2008-03-22T17:42:59 <johill> although removing the test doesn't seem to help
2008-03-22T17:43:41 <johill> actually it does
2008-03-22T17:43:42 <johill> "Got a list of 1903 local and 2000 remote pages. This results in 2000 different pages over-all."
2008-03-22T17:43:42 <CIA-39> Alexander Schremmer <alex AT alexanderweb DOT de> default * 3416:5aec9a2ff763 1.7/MoinMoin/wikisync.py: Remove ACL check from the wikisync code - it will be checked again later on. This check might have avoided syncing certain pages even though it would have been ok permission-wise. Thats why a (backport needed).
2008-03-22T17:43:43 <CIA-39> Alexander Schremmer <alex AT alexanderweb DOT de> default * 3417:425d072840ab 1.7/MoinMoin/formatter/text_html.py: Merge main.
2008-03-22T17:43:54 <johill> except the message doesn't make sense
2008-03-22T17:44:06 <xorAxAx> why doesnt it seem to help?
2008-03-22T17:44:19 <johill> it does, I read the wrong log line
2008-03-22T17:44:20 <xorAxAx> looks fine to me
2008-03-22T17:44:26 <johill> well it doesn't seem to list local deleted pages
2008-03-22T17:44:32 <johill> while it does list remoet deleted pages
2008-03-22T17:44:43 <xorAxAx> yes
2008-03-22T17:44:46 <johill> I guess that's only because of 'down'
2008-03-22T17:44:48 <xorAxAx> because you are in down mode
2008-03-22T17:44:51 <johill> but
2008-03-22T17:44:58 <johill> there still aren't 2000 different pages
2008-03-22T17:45:09 <xorAxAx> i guess different is a bit wrong here
2008-03-22T17:45:22 <xorAxAx> "when merging these page sets, you get %i pages"
2008-03-22T17:45:24 <xorAxAx> is the meaning :)
2008-03-22T17:46:13 <johill> yes I suppose it should say
2008-03-22T17:46:24 <ThomasWaldmann> johill: send_page sucks. fix bug with invalid highlight REs - what exactly was the bug?
2008-03-22T17:46:25 <johill> Got a list of 1903 local and 2000 remote pages. This results in 2000 pages to check.
2008-03-22T17:46:37 <xorAxAx> good idea, johill
2008-03-22T17:46:52 <xorAxAx> johill: i change it
2008-03-22T17:47:04 <johill> ThomasWaldmann: http://test17.moinmo.in/RecentChanges?highlight=(.*
2008-03-22T17:47:53 <johill> ThomasWaldmann: because RandomQuote uses send_page which directly looks up the request.form....
2008-03-22T17:48:46 <johill> the problem is old, the fact that it warns you new
2008-03-22T17:49:37 <ThomasWaldmann> ah, ok
2008-03-22T17:50:04 <johill> the root cause is send_page
2008-03-22T17:50:22 <johill> or more to the point, the fact that send_page invokes the theme rather than the other way around
2008-03-22T17:50:34 <johill> (or both invoked from request)
2008-03-22T17:53:28 <johill> xorAxAx: do you already know how many deleted pages remote has at that point? might be worth adding that to the message too
2008-03-22T17:53:31 <johill> i.e.
2008-03-22T17:53:40 <xorAxAx> johill: no, i dont
2008-03-22T17:53:43 <johill> Got a list of 1903 local, 1903 remote and 97 remotely deleted pages
2008-03-22T17:54:07 <johill> ok, too bad
2008-03-22T17:54:44 <CIA-39> Alexander Schremmer <alex AT alexanderweb DOT de> default * 3418:60e0a3979520 1.7/MoinMoin/action/SyncPages.py: Changed status message to "This results in %s pages to process.". It is not necessary to backport this changeset because it changes a translated string.
2008-03-22T17:56:09 <johill> gtg, will be back in about two hours
2008-03-22T17:56:24 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2596:8159d9a04b0c 1.6/MoinMoin/Page.py: send_page sucks. fix bug with invalid highlight REs (backport from 1.7)
2008-03-22T17:58:55 <johill> lanius: had another question yesterday, maybe you can think on it for a bit... the distinction between revisioned and unrevisioned metadata is rather weird, what's up with that?
2008-03-22T18:31:16 <dreimark> johill: this queryparser fix does this help to fix the Hompepage regex
2008-03-22T18:31:19 <dreimark> <<FullSearchCached(regex:(----(-*)(\r)?\n)(.*)CategoryHomepage\b)>>
2008-03-22T18:47:14 <dreimark> bbl
2008-03-22T19:58:36 <johill> dreimark: no. I disagree with that being a valid expression
2008-03-22T20:22:34 <johill> augh
2008-03-22T20:22:46 <johill> why is it that metadata is an object all the way through but items aren't?
2008-03-22T20:24:31 <johill> is metadata revisioned or not?
2008-03-22T20:24:38 <johill> it seems to be half yes half no
2008-03-22T20:25:58 <johill> and doing wikiutil.clean_input in the storage code is just wrong
2008-03-22T20:27:41 <johill> just like accessing self.item._request.remote_addr
2008-03-22T20:29:12 <ThomasWaldmann> johill: the meta data relating to the data revisions is revisioned
2008-03-22T20:29:38 <ThomasWaldmann> the meta data containing e.g. the revision number is not revisioned
2008-03-22T20:30:05 <johill> right
2008-03-22T20:30:12 <johill> it's not exactly clear which is which though
2008-03-22T20:32:08 <johill> is there any metadata other than the revision number that isn't revisioned?
2008-03-22T20:35:02 <ThomasWaldmann> i dont remember exactly. maybe some "deleted" flag.
2008-03-22T21:35:00 <johill> that should be revisioned though, no?
2008-03-22T21:56:33 <johill> this is all so inconsistent
2008-03-22T21:57:01 <johill> EDIT_LOG_ACTION is a metadata entry
2008-03-22T21:57:12 <johill> yet there's .news()
2008-03-22T21:57:18 <johill> rather than going through the metadata for revisions
2008-03-22T22:06:13 <xorAxAx> why should be deleted revisioned?
2008-03-22T22:06:41 <johill> so you can see when in the past the page was deleted and restored
2008-03-22T22:06:52 <johill> ok this is really crappy. LocalEditLog is used in code other than the backends?!
2008-03-22T22:07:04 <johill> that cannot work
2008-03-22T22:19:03 <johill> ahrg
2008-03-22T22:19:07 <johill> why was .news invented?
2008-03-22T22:20:24 <xorAxAx> see the discussions :)
2008-03-22T22:22:46 <johill> no, look, it just doesn't make sense
2008-03-22T22:23:00 <johill> you have evertyhing be metadata of an item
2008-03-22T22:23:05 <johill> yet you have .news *and* .list_revisions
2008-03-22T22:23:24 <xorAxAx> i dont see the duplicity
2008-03-22T22:24:16 <johill> what does news give you?
2008-03-22T22:24:52 <johill> it gives you (item, revno, mtime)
2008-03-22T22:24:59 <xorAxAx> yes
2008-03-22T22:25:14 <xorAxAx> but starting at a certain mtime
2008-03-22T22:26:00 <johill> yeah but that's equivalent to list_items(filter=mtime>start)
2008-03-22T22:26:02 <johill> well
2008-03-22T22:26:04 <johill> it *should* be
2008-03-22T22:26:09 <johill> it in fact is not
2008-03-22T22:26:42 <johill> mind you, the docstring for .news is wrong too
2008-03-22T22:27:06 <xorAxAx> johill: well, no
2008-03-22T22:27:13 <xorAxAx> johill: list_items is much more heavy weight
2008-03-22T22:27:20 <xorAxAx> the idea here is that news is cheaper to implement
2008-03-22T22:27:53 <johill> well no
2008-03-22T22:28:01 <johill> if it was done right then the optimisation were just done within list_items
2008-03-22T22:28:44 <xorAxAx> list_items needs to return full fledged objects, no?
2008-03-22T22:28:53 <xorAxAx> so it should return shadow objects instead?
2008-03-22T22:28:53 <johill> no, it returns just names
2008-03-22T22:29:08 <johill> at the backend level anyway
2008-03-22T22:29:13 <xorAxAx> yes, then you need to look them up etc.
2008-03-22T22:29:23 <xorAxAx> how do you want to optimise that?
2008-03-22T22:29:33 <johill> eh, I mean, revisino numbers, not names
2008-03-22T22:29:45 <johill> and list_revisions
2008-03-22T22:30:14 <xorAxAx> and list_revisions works for multiple items?
2008-03-22T22:32:19 <johill> I'm confused now
2008-03-22T22:33:21 <johill> I think I actually didn't want to complain about .news but about the LocalEditLog use
2008-03-22T22:33:42 <johill> but for .news we also store data twice which is rather dangerous
2008-03-22T22:34:26 <xorAxAx> is the editlog stuff meant at all to be persisted?
2008-03-22T22:34:31 <xorAxAx> or is it only transitional?
2008-03-22T22:34:37 <johill> hm?
2008-03-22T22:35:02 <johill> as in stored to disk or as in 'the code is not done yet'?
2008-03-22T22:35:15 <xorAxAx> the latter
2008-03-22T22:35:41 <johill> probably transitional
2008-03-22T22:35:45 <johill> but the code had to be changed
2008-03-22T22:35:52 <johill> so why it wasn't done right is beyond me
2008-03-22T22:36:09 <johill> and why doesn't .news return deleted items?
2008-03-22T22:36:14 <johill> you want to show something is deleted in RC
2008-03-22T22:37:06 <johill> it does
2008-03-22T22:37:11 <johill> the docstring is bogus I guess
2008-03-22T22:39:37 <johill> and moin16's implementation of list_revisions is bogus too
2008-03-22T22:39:41 <johill> it doesn't list deleted revisions
2008-03-22T22:52:07 <johill> ok this looks saner than I thought
2008-03-22T22:52:20 <johill> the edit log thing just has code to add to it for attachmetns
2008-03-22T22:53:23 <johill> damn inefficient though
2008-03-22T22:53:54 <xorAxAx> the backend?
2008-03-22T22:53:57 <xorAxAx> well, kind of expected
2008-03-22T22:54:10 <johill> not necessary though
2008-03-22T22:54:34 <johill> and I still think list_revisions() is wrong
2008-03-22T22:54:38 <johill> not listing deleted revisions
2008-03-22T22:54:43 <johill> while you can access them
2008-03-22T22:54:51 <johill> i.e. if you create a page, delete it, create again
2008-03-22T22:54:58 <johill> you'll be able to access revisions 1, 2, 3
2008-03-22T22:55:05 <johill> but list revisions will tell you [1, 3]
2008-03-22T22:55:36 <xorAxAx> well, thats maybe a bug
2008-03-22T22:56:03 <johill> also, edit-log assumes that all revisions exist
2008-03-22T22:56:05 <johill> is that true?
2008-03-22T22:56:12 <johill> I mean, the editlog code, not the edit-log file
2008-03-22T22:56:32 <johill> can you create revision 1 and then 5?
2008-03-22T22:57:34 <johill> it seems you could
2008-03-22T22:58:14 <johill> I have nothing against disallowing it though
2008-03-22T22:58:38 <johill> otoh, you can apparently remove_revision() an older one
2008-03-22T22:59:04 <johill> which, tbh, I don't see much sense in
2008-03-22T23:00:05 <johill> what do you think?
2008-03-22T23:02:16 <xorAxAx> well, thats a very strong assumption - whether the number space is contigunous or not
2008-03-22T23:02:32 <xorAxAx> is a bit related to the fact whether revnos are primary handles and how you iterate over them
2008-03-22T23:02:48 <xorAxAx> but indeed, it should be possible to work with non continugous revnos
2008-03-22T23:02:54 <johill> right
2008-03-22T23:02:57 <johill> I agree
2008-03-22T23:03:01 <johill> that makes the moin16 code ugly
2008-03-22T23:03:03 <johill> :)
2008-03-22T23:03:10 <johill> because it'll actually have to delete lines from the edit-log
2008-03-22T23:03:27 <xorAxAx> ?
2008-03-22T23:03:41 <xorAxAx> because delete means prune on that level?
2008-03-22T23:03:46 <johill> yes
2008-03-22T23:05:21 <johill> and list_revisions has to read the edit log
2008-03-22T23:05:42 <johill> not assume that all revisions exist that have a data file
2008-03-22T23:05:47 <johill> (which is wrong for deleted revisions)
2008-03-22T23:26:53 <johill> hrm
2008-03-22T23:27:01 <johill> how can I lock the edit-log for that line removal?
2008-03-22T23:28:43 <xorAxAx> the edit logs dont have locks
2008-03-22T23:28:49 <johill> exactly
2008-03-22T23:28:54 <xorAxAx> IMHO that op shouldnt be supported
2008-03-22T23:28:57 <johill> but if I'm going to remove a line from it, how can I do that?
2008-03-22T23:29:01 <xorAxAx> for that bckend
2008-03-22T23:29:04 <johill> then you can't have holes
2008-03-22T23:29:06 <xorAxAx> you shouldnt :)
2008-03-22T23:29:13 <xorAxAx> well, nobody cares with moin16 :)
2008-03-22T23:29:25 <johill> true
2008-03-22T23:29:36 <johill> so remove_revision just isn't supported?
2008-03-22T23:29:57 <xorAxAx> not in moin16
2008-03-22T23:30:06 <xorAxAx> but in any other backend :-)
2008-03-22T23:30:20 <johill> right
2008-03-22T23:31:17 <johill> ThomasWaldmann: would you mind if I deleted moin17 backend?
2008-03-22T23:31:30 <johill> stupid tests
2008-03-22T23:31:35 <johill> always forget to unlock
2008-03-22T23:31:47 <johill> maybe we should make lock() not allow lock()/unlock() but rather
2008-03-22T23:31:50 <johill> do_locked(function, args)
2008-03-22T23:47:29 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3718:20edb34abd1c 1.7-storage-hwendel/MoinMoin/ (3 files in 3 dirs): remove non-working moin17 backend for now
2008-03-22T23:50:31 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2597:c2981c625591 1.6/ (3 files in 3 dirs): show 'Comments' toggling link in edit bar only if the page really contains comments
2008-03-22T23:51:19 <ThomasWaldmann> johill: it partially worked, so please restore it
2008-03-22T23:53:49 <johill> hm, ok
2008-03-22T23:53:55 <johill> I'll make it skip the tests then
2008-03-22T23:56:53 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3719:175cba04261a 1.7-storage-hwendel/MoinMoin/ (3 files in 3 dirs): revert 20edb34abd1c (moin17 backend removal)
2008-03-22T23:57:59 <johill> huh
2008-03-22T23:58:07 <johill> MoinMoin/storage/backends/common.py
2008-03-22T23:58:12 <dreimark> ThomasWaldmann: can the parser be used to get if a page has used comments and can that be indicated in the theme
2008-03-22T23:58:14 <johill> _handle_error()
2008-03-22T23:58:23 <johill> why can the backend not raise the proper error up front? MoinMoin: MoinMoinChat/Logs/moin-dev/2008-03-22 (last edited 2008-03-21 23:15:02 by IrcLogImporter)
