2008-07-25T00:12:25 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4367:4c8d7ebe807e 1.8-mercurialbackend-ppacana/MoinMoin/ (6 files in 3 dirs): Merge with 1.8-storage-cdenter.
2008-07-25T00:12:25 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4368:3d5d60b133ad 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Fixed backend init bug when creating repository over existing data.
2008-07-25T00:12:26 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4369:afe5bb18a3ff 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/ (hg.py research/repo_force_changes.diff): (hg) Added to repo hg patch for enabling forced commits with no changes in beetween. Docstring update.
2008-07-25T00:12:27 <CIA-53> Pawel Pacana <pawel.pacana@gmail.com> default * 4370:fd567e1d5f8c 1.8-mercurialbackend-ppacana/MoinMoin/storage/backends/hg.py: (hg) Revision meta dictionary keys mangling to avoid name collisions with hg repo keys.
2008-07-25T08:57:47 <ThomasWaldmann> moin
2008-07-25T10:00:18 <CIA-53> MelitaMihaljevic default * 3954:dbbcd2a8f118 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/group_ldap.py: fixed ManageFactory: added ldap configuration options required for Manager and Group
2008-07-25T10:00:20 <CIA-53> MelitaMihaljevic default * 3955:fa64982876e7 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/ (3 files in 2 dirs): datastruct/backend: fixed comments and typos
2008-07-25T10:00:20 <CIA-53> MelitaMihaljevic default * 3956:7a94500262b9 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/group_ldap.py: removed unused code, remove unsane returns and changed names
2008-07-25T10:00:21 <CIA-53> MelitaMihaljevic default * 3957:a937dae909d5 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/test_groupldap.py: test if the ldap backend works with acl code
2008-07-25T10:00:23 <CIA-53> MelitaMihaljevic default * 3958:f6a93b547b9c 1.8-ldapgroups-mmihaljevic/MoinMoin/_tests/ldap_testdata.py: added entries for testing ldap with acl code (group and user are defined in camelcase format) - will be removed later
2008-07-25T10:00:27 <CIA-53> MelitaMihaljevic default * 3959:8ab2dc294f9c 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/ (_tests/test_groupldap.py group_ldap.py): fixed __contain__ method (avoid IndexError)
2008-07-25T10:00:30 <CIA-53> MelitaMihaljevic default * 3960:c7c5a7fc3a94 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/group_wiki.py: fixed wiki backend - added raise KeyError if the page does not exist, it is used in datastruct/__init__.py in Manager __gettitem__
2008-07-25T10:00:34 <CIA-53> MelitaMihaljevic default * 3961:67bdd29c39a0 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/test_groupwiki.py: fixed comments
2008-07-25T10:00:39 <CIA-53> MelitaMihaljevic default * 3962:14a12079b152 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/test_groupmultiple.py: added test for using multiple defined group backends
2008-07-25T10:00:42 <CIA-53> MelitaMihaljevic default * 3963:ad4e41548c44 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/test_groupldap.py: fixed pep8 errors
2008-07-25T10:00:45 <CIA-53> MelitaMihaljevic default * 3964:88c54aa4b57b 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/test_groupmultiple.py: fixed test and added explanation
2008-07-25T10:00:50 <CIA-53> MelitaMihaljevic default * 3965:5ebbd3a029e3 1.8-ldapgroups-mmihaljevic/wiki/config/more_samples/multi_backend_wikiconfig_snippet: added snippet for multiple group backends
2008-07-25T10:00:53 <CIA-53> MelitaMihaljevic default * 3966:4d4e50a6e6e7 1.8-ldapgroups-mmihaljevic/MoinMoin/datastruct/backend/_tests/test_groupldap.py: removed unused comments in test
2008-07-25T10:01:11 <TheSheep> whee, it was a long night
2008-07-25T10:05:34 <gizmach> moin
2008-07-25T10:05:57 <gizmach> dreimark: I sent all code I commited these days
2008-07-25T10:08:02 <dreimark> moin
2008-07-25T10:09:19 <gizmach> hi dreimark
2008-07-25T10:09:56 * gizmach got normal connection for about an hour on linux at mums company
2008-07-25T10:14:29 <gizmach> brb (is reading)
2008-07-25T10:16:04 <dreimark> hi gizmach, will look into your commits later
2008-07-25T10:22:55 <dreimark> http://www.sysadminday.com/
2008-07-25T10:57:14 <gizmach> dreimark: ok, can you please write them somewhere because soon I'll need to disconect
2008-07-25T11:04:04 <gizmach> well I need to go
2008-07-25T12:03:21 <ThomasWaldmann> zenhase: AttributeError: class Config has no attribute 'traceback_show'
2008-07-25T12:04:04 <ThomasWaldmann> can you please fix the stuff ASAP so that one does not get errors when using ./wikiserver.py
2008-07-25T12:23:22 <ThomasWaldmann> zenhase: 12 test failures in your repo (please fix all test_sourcecode failures in a single commit without other changes)
2008-07-25T12:34:19 <PawelPacana> moin
2008-07-25T12:37:51 <ThomasWaldmann> moin pawel
2008-07-25T12:48:06 <waldi> ThomasWaldmann: after reading pep-8 again, i think the check is too sensitive. the check say that there needs to be a space after _each_ comma. the test even violates against the rule that there should be no spaces directly in parentheses in this case: "(bla,)"
2008-07-25T12:51:34 <ThomasWaldmann> waldi: the no spaces in parens like means ( stufflikethis ) or (stuff = likethis), but it still wants (stufflikethis, )
2008-07-25T12:52:16 <waldi> and the later violates against: Avoid extraneous whitespace in the following situations: Immediately inside parentheses, brackets or braces.
2008-07-25T12:54:44 <ThomasWaldmann> well, the spaces after comma has higher prio :)
2008-07-25T12:55:19 <TheSheep> ThomasWaldmann: I haven't seen anything like that anywhere
2008-07-25T12:55:56 <TheSheep> sheep@ghostwheel:/usr/lib/python2.5$ grep ',)' *.py | wc -l
2008-07-25T12:55:56 <TheSheep> 156
2008-07-25T12:56:04 <TheSheep> sheep@ghostwheel:/usr/lib/python2.5$ grep ', )' *.py | wc -l
2008-07-25T12:56:05 <TheSheep> 1
2008-07-25T12:56:39 <TheSheep> that's Cookie.py and it has spacing wrong on that line entirely
2008-07-25T12:56:41 <ThomasWaldmann> TheSheep: how many other pep8 errors does this code have? X)
2008-07-25T12:57:03 <TheSheep> 156:1 is a pretty strong bias, I think
2008-07-25T12:57:36 <TheSheep> alsmost seems like intentional
2008-07-25T12:57:39 <TheSheep> almost
2008-07-25T12:58:08 <xorAxAx> waldi: lets go by (my) aesthetics and not by the rules - i like (foo, ) better :)
2008-07-25T12:59:07 <waldi> ThomasWaldmann: the space after comma rule is not even mentioned in the document
2008-07-25T12:59:13 <TheSheep> 1946:496 for grep -R
2008-07-25T12:59:13 <xorAxAx> hmm
2008-07-25T12:59:29 <xorAxAx> note that stdlib is pretty old code often
2008-07-25T12:59:39 <TheSheep> pep8 is old too
2008-07-25T12:59:45 <waldi> pep8 is older
2008-07-25T12:59:49 <TheSheep> waldi: is not
2008-07-25T13:00:30 <xorAxAx> yeah, its old but wasnt often considered when writing the code. but if they like (f,) better for the one tuple pattern, its fine. we should just have a unique style probably
2008-07-25T13:01:15 <ThomasWaldmann> our style is (foo, )
2008-07-25T13:01:32 <ThomasWaldmann> (and we won't change that now)
2008-07-25T13:01:44 <TheSheep> ok :0
2008-07-25T13:01:45 <TheSheep> :)
2008-07-25T13:02:23 <TheSheep> pep8 also says that consistency with the rest of the project is more important than the rules
2008-07-25T13:02:30 <ThomasWaldmann> (I think all students and mentors currently have better things to do than to change the pep8 checks and all affected src)
2008-07-25T13:10:20 <zenhase> ThomasWaldmann: uhm .. i thought i fixed everything for wikiserver.py
2008-07-25T13:10:25 <zenhase> sorry that
2008-07-25T13:11:11 <zenhase> and regarding test_sourcecode.py: it didn't show me any errors, therefore i missed them
2008-07-25T13:11:41 <zenhase> perhaps the xattr-based caching of sourcecode-tests doesn't work as expected
2008-07-25T13:14:40 <zenhase> this is really strange ... i ran again with every source-file touched, so they get reevaluated
2008-07-25T13:14:54 <zenhase> and still no errors show up
2008-07-25T13:15:00 <zenhase> can you paste me a log for this test?
2008-07-25T13:15:06 <gizmach> moin
2008-07-25T13:18:11 <ThomasWaldmann> zenhase: do you have xattr module?
2008-07-25T13:20:03 <zenhase> yes i have
2008-07-25T13:20:26 <ThomasWaldmann> ok, then just add a raise ImportError in test_sourcecode
2008-07-25T13:20:37 <ThomasWaldmann> then you'll see if it is related to that
2008-07-25T13:23:55 <zenhase> hmm no, doesn't seem to be related
2008-07-25T13:24:29 <zenhase> yields 457 tests here and every single one passes
2008-07-25T13:24:45 <CIA-53> Florian Krupicka <florian.krupicka@googlemail.com> default * 3937:f35fdfdc1995 1.8-wsgi-fkrupicka/MoinMoin/ (script/server/standalone.py wsgiapp.py): Fix missing attributes for standalone server config, remove reloader (crashes currently)
2008-07-25T13:25:05 <mmihaljevic> moin again
2008-07-25T13:25:42 <TheSheep> wb mmihaljevic
2008-07-25T13:25:57 <TheSheep> guys, do you remember when j ohill is back?
2008-07-25T13:34:34 <ThomasWaldmann> mmihaljevic: if you #comment, please do it like this: # comment
2008-07-25T13:35:04 <xorAxAx> TheSheep: this weekend
2008-07-25T13:35:10 <xorAxAx> he went there for a week
2008-07-25T13:35:27 <ThomasWaldmann> mmihaljevic: and maybe rather use ldap_conn consistently (not "lo")
2008-07-25T13:35:54 <TheSheep> xorAxAx: thanks
2008-07-25T13:36:42 <zenhase> ThomasWaldmann: i really don't know why these tests are not showing up :o
2008-07-25T13:36:52 <zenhase> ThomasWaldmann: i mean the errors in some of them
2008-07-25T13:36:55 <mmihaljevic> ThomasWaldmann: ah yes pardon I will fix it
2008-07-25T13:37:03 <xorAxAx> zenhase: uncommitted changes probably?
2008-07-25T13:37:05 <mmihaljevic> I forget on it completly
2008-07-25T13:37:13 <xorAxAx> zenhase: or multiple heads?
2008-07-25T13:37:13 <zenhase> xorAxAx: no
2008-07-25T13:37:32 <zenhase> also no multiple heads
2008-07-25T13:37:37 <xorAxAx> hmm, weird
2008-07-25T13:38:11 <mmihaljevic> ThomasWaldmann: I'm writing a list of todos now and some things that are not completly clear to me, will send them so if you, dreimark or anybody could comment them.
2008-07-25T13:38:19 <ThomasWaldmann> mmihaljevic: also, maybe give reasonable default values to scope and timeout params of ldap group manager backend
2008-07-25T13:38:19 <zenhase> the only uncommited modification right now is the raise ImportError in test_sourcecode to work around xattr being used
2008-07-25T13:39:03 <mmihaljevic> ThomasWaldmann: ok
2008-07-25T13:39:08 <zenhase> the tests are not always stable it seems, already had issues in the past, where tests where failing until i cleaned out all .pyc files
2008-07-25T13:39:28 <xorAxAx> well, thats a python problem, not a moin one :)
2008-07-25T13:39:28 <ThomasWaldmann> mmihaljevic: give ldap_conn as first argument
2008-07-25T13:40:26 <ThomasWaldmann> zenhase: you pushed everything except that?
2008-07-25T13:42:46 <zenhase> yes
2008-07-25T13:42:58 <ThomasWaldmann> ok, i rerun tests now
2008-07-25T13:44:22 <zenhase> should be only 2 failing tests (which fail because they raise NotImplementedError)
2008-07-25T13:45:12 <ThomasWaldmann> mmihaljevic: as your stuff seems to work now, we should do that simplification change now
2008-07-25T13:46:18 <mmihaljevic> ThomasWaldmann: hm.. ok
2008-07-25T13:46:59 <ThomasWaldmann> i.e. instead of using a GMF and GMBFs, you just use cfg.gmf = lambda request: GM(GMB1(request, ...), GMB2(request, ...))
2008-07-25T13:47:45 <ThomasWaldmann> mmihaljevic: is it clear what I mean and why we do this?
2008-07-25T13:49:23 <mmihaljevic> ok we do that to get rid f the GMF and GMFB and to simplify admin to define only GM and GMB which want to use
2008-07-25T13:49:28 <ThomasWaldmann> (and please do everything related to this change in a single and separate changeset)
2008-07-25T13:50:01 <ThomasWaldmann> yes
2008-07-25T13:50:16 <ThomasWaldmann> you can tell how it works after that change?
2008-07-25T13:51:59 <mmihaljevic> the admin only need to define gm = GM (GMB....), and upon request the gmb is called with the right request
2008-07-25T13:52:50 <ThomasWaldmann> no, admin needs to define the lambda as seen above
2008-07-25T13:52:54 <xorAxAx> hmm, that scheme looks pretty expensive if the backends need to reconnect on every request
2008-07-25T13:52:54 <zenhase> ThomasWaldmann: how was the outcome of your rerun?
2008-07-25T13:53:16 <ThomasWaldmann> zenhase: it is still running
2008-07-25T13:53:27 <xorAxAx> also the admin shouldnt be bothered with such details probably (that should be hidden in the multiconfig class setup)
2008-07-25T13:53:31 <ThomasWaldmann> tests in your repo seem to be slower somehow
2008-07-25T13:53:58 <zenhase> uh ... ok :o
2008-07-25T13:54:18 <ThomasWaldmann> xorAxAx: yes, we can refine that later
2008-07-25T13:54:40 <xorAxAx> probably
2008-07-25T13:55:18 <mmihaljevic> ThomasWaldmann: ok than I don't understand how the wikiconfig_local should look like (got lost a bit)
2008-07-25T13:55:56 <ThomasWaldmann> mmihaljevic: you know what f = lamba x: x*x means?
2008-07-25T13:56:21 <mmihaljevic> yes
2008-07-25T13:56:38 <ThomasWaldmann> what does it mean?
2008-07-25T13:56:40 <mmihaljevic> it defines f(x) = x*x
2008-07-25T13:56:52 <mmihaljevic> but does not give pa parm
2008-07-25T13:56:52 <ThomasWaldmann> ok
2008-07-25T13:57:56 <ThomasWaldmann> what i told at 13:46 is just that cfg.gmf is a function that gets request as single param and then creates all the instances, giving request there as needed
2008-07-25T13:58:28 <mmihaljevic> ok
2008-07-25T13:58:40 <ThomasWaldmann> that way you can get rid of all factory code
2008-07-25T13:59:05 <mmihaljevic> ok that part is clear to me
2008-07-25T13:59:54 <mmihaljevic> but the part that is not clear to me is where do I need to do that?
2008-07-25T14:00:25 <ThomasWaldmann> the lambda thing is done in wikiconfig(_local) [for now]
2008-07-25T14:00:44 <mmihaljevic> ThomasWaldmann: ok
2008-07-25T14:00:52 <ThomasWaldmann> and you call that function at the same place where you use your factory now
2008-07-25T14:01:54 <mmihaljevic> ok that makse sense
2008-07-25T14:05:20 <ThomasWaldmann> zenhase: http://static.moinmo.in/temp/wsgitests
2008-07-25T14:05:57 <ThomasWaldmann> (sorry, false alarm: all pep8 failures are in werkzeug package (i just symlinked it), but there are some other failures also)
2008-07-25T14:05:58 <zenhase> ThomasWaldmann: uhm
2008-07-25T14:06:09 <zenhase> yeah :)
2008-07-25T14:07:11 <ThomasWaldmann> if you want to execute the ldap tests, you just need slapd and python-ldap installed (and disable apparmor if you have it running)
2008-07-25T14:07:12 <zenhase> perhaps mitsuhiko would like some test_sourcecode in his repo too :)
2008-07-25T14:07:36 <zenhase> oO
2008-07-25T14:08:12 <zenhase> when there are no failing tests left anymore i will merge with 1.8 now
2008-07-25T14:08:17 <zenhase> s/when/since
2008-07-25T14:08:24 <ThomasWaldmann> hm?
2008-07-25T14:08:33 <ThomasWaldmann> look at that file
2008-07-25T14:10:09 <ThomasWaldmann> ldap_login tests fail (in your repo only)
2008-07-25T14:11:31 <zenhase> ah, sorry, didn't realized you were still referring to me
2008-07-25T14:11:56 <mmihaljevic> bbl (programming)
2008-07-25T14:12:27 <zenhase> ok, will fix that
2008-07-25T14:12:39 <zenhase> but i need an ldap-server
2008-07-25T14:12:46 <ThomasWaldmann> dennda: any progress on action=info traceback and saves not working?
2008-07-25T14:12:46 <zenhase> *install install*
2008-07-25T14:13:42 <ThomasWaldmann> dennda: when creating a new page, first save seems to work. but if you change it again, you don't see that change afterwards.
2008-07-25T14:20:46 <ThomasWaldmann> byeongweon: table caption roundtrip is not working
2008-07-25T14:25:22 <ThomasWaldmann> it doesn't make sense to give table height in %, right?
2008-07-25T14:25:53 <xorAxAx> well, it certainly has consequences
2008-07-25T14:26:03 <xorAxAx> byeongweon: ping
2008-07-25T14:28:35 <ThomasWaldmann> byeongweon: image roundtrips, but in a bit strange way (if you insert a non-existing image, you first see some "broken image" graphics, then switch to text mode and back to gui mode, then it will just display the attachment name)
2008-07-25T14:29:04 <ThomasWaldmann> maybe it should rather give "broken image" graphics after roundtrip also
2008-07-25T14:29:46 <xorAxAx> ThomasWaldmann: can you amend the wiki page?
2008-07-25T14:30:37 <ThomasWaldmann> yes
2008-07-25T14:44:24 <byeongweon> ThomasWaldmann: ok. I will check that.
2008-07-25T14:45:02 * ThomasWaldmann still editing your page. thanks.
2008-07-25T14:46:26 <xorAxAx> byeongweon: how is the bug fixing going? any blockers?
2008-07-25T14:46:49 <byeongweon> xorAxAx: not yet any blokers.
2008-07-25T14:48:36 <dreimark> mmihaljevic: backend works with acs (in LDAPstBase there are defined setup class and teardown class - can I use them
2008-07-25T14:49:49 <dreimark> s/acs/acls/
2008-07-25T14:50:03 <dreimark> and is that at all a question ?
2008-07-25T14:52:24 <dreimark> # JoeDoe has right to read FirstGroup - return True in testConfigBackendAcl(self)
2008-07-25T14:53:45 <dreimark> that might be right but that is not what the test promises
2008-07-25T14:54:05 <dreimark> or did by allow = acl.may(self.request, name, dowhat)
2008-07-25T14:58:05 <ThomasWaldmann> dreimark: she's gone
2008-07-25T14:58:32 <dreimark> I know that she reads it on the MM log page
2008-07-25T15:00:18 <dennda> ThomasWaldmann: Actually, don't you notice something else going wrong when changing the page? :)
2008-07-25T15:06:39 <ThomasWaldmann> sure i also noticed the cosmetical problem
2008-07-25T15:14:59 <dennda> You think it's only cosmetical?
2008-07-25T15:15:19 <dennda> It seems to retreive the wrong information from the cache
2008-07-25T15:16:01 <xorAxAx> which cache, dennda?
2008-07-25T15:18:56 <dennda> Ok that was the wrong word
2008-07-25T15:18:57 <dennda> huh
2008-07-25T15:19:00 <dennda> wait a second
2008-07-25T15:20:17 <xorAxAx> currently there should be no caches in place besides the formatter/send_page related one
2008-07-25T15:24:47 <dennda> xorAxAx: I meant the place where pagelinks and text_html resides
2008-07-25T15:24:53 <dennda> (which is /tmp, in my branch)
2008-07-25T15:25:48 <dennda> If you delete those two files while the wikiserver is running, the cosmetical problem is gone
2008-07-25T15:26:52 <xorAxAx> dennda: hmm, then are the cache files wrong or is it reading the wrong files?
2008-07-25T15:27:18 <dennda> I am not sure, but I think the latter applies
2008-07-25T15:27:42 <dennda> I would love to solve that but it is quite difficult for me without the whole understanding of how moin works here
2008-07-25T15:27:52 <dennda> Maybe you could provide a pointer?
2008-07-25T15:29:30 <xorAxAx> NULL
2008-07-25T15:29:46 <xorAxAx> or better, (void*) NULL
2008-07-25T15:29:53 <dennda> Not that kind of pointer
2008-07-25T15:29:54 <dennda> :)
2008-07-25T15:30:15 <dennda> Because that inevidently leads to me throwing a NullPointerException back at you
2008-07-25T15:31:03 <xorAxAx> well, first find out which files are created during page rendering
2008-07-25T15:32:41 <dennda> Sounds easy, is hard
2008-07-25T15:34:56 <xorAxAx> um
2008-07-25T15:35:03 <xorAxAx> ?
2008-07-25T15:35:11 <xorAxAx> dennda: what is the problem?
2008-07-25T15:35:21 * xorAxAx throws ls at dennda
2008-07-25T15:39:42 <ThomasWaldmann> dennda: if you are stuck somewhere when trying to find some bug, using a debugger sometimes gives some hint rather fast
2008-07-25T15:56:29 <ThomasWaldmann> waldi: fuer bilder bitte <img>, nicht <object>
2008-07-25T15:58:30 <ThomasWaldmann> waldi: ein (optionaler) html pretty printed output waere auch nett (uebersichtlicher als wenn alles am stueck ist)
2008-07-25T15:58:55 <ThomasWaldmann> koennte man solange es nicht released ist auch als default so einstellen
2008-07-25T15:59:31 <waldi> ThomasWaldmann: hmm, needs some hacks on the serializer
2008-07-25T15:59:43 <waldi> and a proper heuristic for images?
2008-07-25T16:00:05 <ThomasWaldmann> wikiutil.isImage or similar
2008-07-25T16:02:35 <ThomasWaldmann> waldi: isPicture
2008-07-25T16:05:33 <ThomasWaldmann> (of course this is not foolproof, but the problem with object is that stuff like adblock displays it in a weird way)
2008-07-25T16:12:20 <TheSheep> for local pages one could also use imghdr
2008-07-25T16:12:30 <TheSheep> but there is the mime type
2008-07-25T16:12:35 <TheSheep> ok, that was silly
2008-07-25T16:15:37 <ThomasWaldmann> waldi: if i preview the creole test page:
2008-07-25T16:15:40 <ThomasWaldmann> File "/home/tw/eclipse-ws/moin-1.8-dom/MoinMoin/request/__init__.py", line 1527, in end_include
2008-07-25T16:15:43 <ThomasWaldmann> self.include_id, pids = self._include_stack.pop()
2008-07-25T16:15:46 <ThomasWaldmann> IndexError: pop from empty list
2008-07-25T16:17:00 <waldi> complete traceback?
2008-07-25T16:17:19 <ThomasWaldmann> you can't reproduce?
2008-07-25T16:19:26 <xorAxAx> please write a failing test :)
2008-07-25T16:19:35 <xorAxAx> there is a "test page" and no unittest? *sigh*
2008-07-25T16:19:55 <waldi> ThomasWaldmann: no, i can't
2008-07-25T16:20:19 <waldi> there is not much left which calls request.end_include
2008-07-25T16:20:40 <xorAxAx> did you modify request, waldi?
2008-07-25T16:20:46 <waldi> to be exact the html formatter, which is only used to expand macros
2008-07-25T16:20:47 <waldi> no
2008-07-25T16:21:01 <xorAxAx> so your stuff all fits into the old apis?
2008-07-25T16:21:16 <xorAxAx> (otherwise i dont see how you could get it running without modifying request :))
2008-07-25T16:21:57 <waldi> ThomasWaldmann: okay, so it is the editor code
2008-07-25T16:22:18 <waldi> xorAxAx: yep, no modification to request
2008-07-25T16:23:11 <waldi> but I can't reproduce it
2008-07-25T16:45:06 <dreimark> emeraldtree ?
2008-07-25T16:48:08 <dreimark> waldi: ThomasWaldmann where do I get emeraldtree ?
2008-07-25T16:48:29 <waldi> dreimark: http://hg.moinmo.in/EmeraldTree
2008-07-25T16:49:26 <xorAxAx> waldi: does it also have a different api?
2008-07-25T16:49:34 <xorAxAx> because there are etree impls that offer more features as well
2008-07-25T16:49:45 <xorAxAx> (e.g. lxml can do validation against xsds)
2008-07-25T16:51:10 <waldi> xorAxAx: the whole fork is about api incompatibilities
2008-07-25T16:51:40 <xorAxAx> waldi: hmm
2008-07-25T16:51:56 <xorAxAx> btw, why doesnt it show up here?
2008-07-25T16:51:59 <xorAxAx> via cia
2008-07-25T16:52:37 <dreimark> waldi: if I do enter only #format creole on a page
2008-07-25T16:53:02 <dreimark> then the result looks different to http://moinmo.in/ReimarBauer/creoletest
2008-07-25T16:53:13 <waldi> xorAxAx: ask thomas, i don't know if he setup cia for it
2008-07-25T16:53:34 <dreimark> there is no empty line and the second lower editbar
2008-07-25T16:53:41 <dreimark> is shifted to the right
2008-07-25T16:55:22 <xorAxAx> ThomasWaldmann: can we have emeraldtree cia notices here?
2008-07-25T16:55:49 <ThomasWaldmann> i didnt register it
2008-07-25T16:56:17 <xorAxAx> i think you dont need to
2008-07-25T16:56:23 <waldi> dreimark: yeah, the page content is serialized as xml, but the complete page is sent as html. xml with a html parser produces such problems
2008-07-25T16:56:44 <xorAxAx> just run that via the moin cia project and use a different branch name for example
2008-07-25T17:00:59 <ThomasWaldmann> ok, we'll see when next push to ET happens
2008-07-25T17:01:16 <dreimark> <div xmlns="http://www.w3.org/1999/xhtml" /> the /> is not HTML4.0.1
2008-07-25T17:01:49 <ThomasWaldmann> dreimark: from the dom repo code?
2008-07-25T17:02:07 <xorAxAx> well, genshi has usable serializers
2008-07-25T17:02:10 <xorAxAx> thats not a large issue
2008-07-25T17:02:20 <ThomasWaldmann> (that is expected to give xhtml)
2008-07-25T17:02:41 <dreimark> ThomasWaldmann: page source of the #format only rendered page
2008-07-25T17:03:03 <dreimark> waldi: repo yes
2008-07-25T17:03:05 <ThomasWaldmann> (and we'll change the theme and other code to be also xhtml after we merge that)
2008-07-25T17:03:26 <dreimark> ok
2008-07-25T17:05:04 <waldi> gecko even missparses them as xhtml
2008-07-25T17:05:07 * dreimark wondered only because if you add a line with one sign then the line is changed
2008-07-25T17:05:16 <dreimark> to <div xmlns="http://www.w3.org/1999/xhtml"><p> k</p></div>
2008-07-25T17:06:27 <waldi> <div /> is a shorter version of <div></div> in xml
2008-07-25T17:08:26 <dreimark> that's ok. I'd not expected that it changes it if there is <p> following
2008-07-25T17:08:37 <dreimark> I found an other problem
2008-07-25T17:08:51 <dreimark> entering <<br>> shows a traceback
2008-07-25T17:08:57 <dreimark> global name 'elem' is not defined
2008-07-25T17:09:28 <waldi> are macro names case insensitive?
2008-07-25T17:09:31 <waldi> please show
2008-07-25T17:09:52 <dreimark> yes they are
2008-07-25T17:09:59 <dreimark> and in uppercase it gives an other one
2008-07-25T17:10:09 <dreimark> TypeError
2008-07-25T17:10:27 * dreimark prepares
2008-07-25T17:15:13 <dreimark> waldi: http://moinmo.in/4ct10n/AttachFile/BastianBlank/TreeOutputFormatter?action=AttachFile
2008-07-25T17:16:15 <waldi> the second is the error that it can't load the macro "br"
2008-07-25T17:16:56 <waldi> dreimark: even in my 1.8 copy macro names are case sensitive
2008-07-25T17:17:44 <dreimark> waldi: they are case sensitive but it does not give a traceback
2008-07-25T17:17:59 <dreimark> MM tells "Cannot load macro br"
2008-07-25T17:18:19 <waldi> 17:09 < waldi> are macro names case insensitive?
2008-07-25T17:18:20 <waldi> 17:09 < dreimark> yes they are
2008-07-25T17:18:40 <dreimark> oh read wed wrong
2008-07-25T17:18:44 <dreimark> they are sensitive
2008-07-25T17:19:07 <xorAxAx> well, serializing to xhtml has known disadvantage
2008-07-25T17:19:08 <xorAxAx> s
2008-07-25T17:20:10 <waldi> dreimark: please provide the traceback of the typeerror
2008-07-25T17:20:28 <dreimark> that is the first one starting with BR
2008-07-25T17:20:46 <dreimark> http://moinmo.in/BastianBlank/TreeOutputFormatter?action=AttachFile&do=view&target=BR_traceback.html
2008-07-25T17:25:25 <waldi> dreimark: found it
2008-07-25T17:27:20 <waldi> fixed, you may pull
2008-07-25T17:28:20 <dreimark> :)
2008-07-25T17:33:24 <dreimark> waldi: another one http://moinmo.in/BastianBlank/TreeOutputFormatter?action=AttachFile&do=get&target=Assertion_traceback.html
2008-07-25T17:33:42 <dreimark> <<WikiConfig>> on the page
2008-07-25T17:33:55 <dreimark> end tag mismatch (expected {http://www.w3.org/1999/xhtml}p, got {http://www.w3.org/1999/xhtml}div)
2008-07-25T17:36:32 <waldi> dreimark: this is one of the problems of the handling of old macros. if they create html, it needs to be more correct
2008-07-25T17:37:10 <waldi> there is not much i can do about that
2008-07-25T17:38:30 <waldi> but why does WikiConfig return plain html?
2008-07-25T17:38:45 <xorAxAx> waldi: because the formatter doesnt have form operations?
2008-07-25T17:39:13 <xorAxAx> i fear some "solution" is needed here
2008-07-25T17:39:14 <waldi> dreimark: please provide the log
2008-07-25T17:39:15 <waldi> xorAxAx: yes
2008-07-25T17:39:35 <waldi> dreimark: it should show at least a warning
2008-07-25T17:39:45 <xorAxAx> and it doesnt sound too hard - in this case its an open p tag
2008-07-25T17:40:11 <dreimark> waldi: have not set any log rules, just tried your repo
2008-07-25T17:40:14 <xorAxAx> (which one could fix by seeing that its still open and closing it - of course it might not validate against the dtd but thats a bigger problem)
2008-07-25T17:41:02 <waldi> yeah, paragraph have an optional endtag
2008-07-25T17:41:06 * waldi hats sgml
2008-07-25T17:41:42 <waldi> s/hats/hates/
2008-07-25T17:41:44 <dreimark> xorAxAx: resizing of images by fckeditor is broken
2008-07-25T17:42:11 <waldi> xorAxAx: in this case extending the compatibility formatter is easier
2008-07-25T17:42:12 <dreimark> you can resize but it is not saved
2008-07-25T17:42:37 <dreimark> it does write {{attachment:abc.jpg||style="width: 484px; height: 484px;"}}
2008-07-25T17:44:48 <xorAxAx> waldi: yes, i meant that
2008-07-25T17:45:03 <xorAxAx> dreimark: add it to the wiki page
2008-07-25T17:45:09 <dreimark> doing
2008-07-25T17:46:45 <waldi> xorAxAx: oh, found the bug. macro.request.formatter instead of macro.formatter
2008-07-25T17:47:11 <xorAxAx> hrm
2008-07-25T17:47:18 <waldi> so the compatibility formatter is not used
2008-07-25T17:47:20 <xorAxAx> yes
2008-07-25T17:47:56 <xorAxAx> i think a 5 liner request proxy might compensate that
2008-07-25T17:48:16 <xorAxAx> because this is a pretty ... obscure bug if you request people to port to that interface
2008-07-25T17:48:21 <waldi> yeah. i want to mask out write anyway
2008-07-25T17:48:29 <xorAxAx> (or request.formatter should be removed completly)
2008-07-25T17:48:31 <xorAxAx> ah
2008-07-25T17:48:32 <xorAxAx> ok
2008-07-25T17:53:08 <ThomasWaldmann> waldi: http://hg.moinmo.in/EmeraldTree/rev/7e448d707f62
2008-07-25T17:54:14 <waldi> okay
2008-07-25T17:54:15 <xorAxAx> hmm, can one write a proxy that lets people use other etree impls, waldi?
2008-07-25T17:54:36 <xorAxAx> because its really a pity that moin needs to deviate here (even for good, api-wise)
2008-07-25T17:54:46 <waldi> no
2008-07-25T17:54:57 <xorAxAx> why not?
2008-07-25T17:55:02 <ThomasWaldmann> strange, cia should have announced this
2008-07-25T17:55:11 <waldi> ThomasWaldmann: cia seems to be broken
2008-07-25T17:55:41 <waldi> xorAxAx: because of the changes in the text nodes handling
2008-07-25T17:55:49 <xorAxAx> waldi: well, the proxy would translate that
2008-07-25T17:58:11 <dreimark> xorAxAx: indenting is also broken. It looks like we can get new problems / bugs.
2008-07-25T17:58:29 <xorAxAx> dreimark: yeah, the whole concept is a bit fragile
2008-07-25T17:58:38 <xorAxAx> dreimark: just write it down on a wiki page
2008-07-25T17:59:07 <waldi> xorAxAx: the question is, can a proxy translate this
2008-07-25T17:59:24 <dreimark> I've added it to his page. I think because we aren't almost the people using the gui editor, we should have a testwiki from his project
2008-07-25T17:59:51 <dreimark> so he can get more response from people who like this editing concept
2008-07-25T18:02:02 <dreimark> bbl
2008-07-25T18:22:05 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 42:7e448d707f62 EmeraldTree/README: README: add a minimal header about the EmeraldTree changes
2008-07-25T18:57:36 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4032:edb32c2db024 1.8-dom-bblank/MoinMoin/formatter/compatibility.py: Compatibility formatter - Support definition lists
2008-07-25T18:57:42 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4033:f43f2ff2d87e 1.8-dom-bblank/MoinMoin/converter2/macro.py:
2008-07-25T18:57:42 <CIA-53> Macro converter - Proxy the request object
2008-07-25T18:57:42 <CIA-53> Add a proxy for the request object to catch direct writes and access to the
2008-07-25T18:57:42 <CIA-53> formatter object. The formatter object is localy overriden.
2008-07-25T18:57:44 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4034:2ad7f2b3455f 1.8-dom-bblank/MoinMoin/formatter/text_html.py: HTML formatter - Drop auto close magic
2008-07-25T19:22:31 <waldi> xorAxAx: the element tree version in python 2.5 is not even able to write XHTML
2008-07-25T19:34:44 <xorAxAx> waldi: so?
2008-07-25T19:34:57 <xorAxAx> thats not a reason to fork the fundamental datatype
2008-07-25T19:40:17 <ThomasWaldmann> xorAxAx: that tail stuff was disliked by quite some people from the beginnings of ET
2008-07-25T19:40:45 <xorAxAx> well, its still proprietary - with all disadvantages
2008-07-25T19:41:09 <xorAxAx> at least you can easily transform it if you need to connect to another system. but reusing e.g. lxml isnt possible anymore
2008-07-25T19:41:30 <xorAxAx> i dont say it was a bad design decision to "move" that text stuff in the api, its just incompatible
2008-07-25T19:42:09 <ThomasWaldmann> well, the discussion about that was 2 months ago
2008-07-25T19:42:24 <xorAxAx> and nobody had this idea? :)
2008-07-25T19:42:52 <ThomasWaldmann> what idea?
2008-07-25T19:43:13 <xorAxAx> that forking a commonly used datatype and large stdlib package to nicen the api is bad
2008-07-25T19:44:04 <ThomasWaldmann> we were aware of that a fork means additional work, but still is better than to work with a bad api
2008-07-25T19:47:14 <xorAxAx> ThomasWaldmann: just for the record, please point me to the discussion in the logs
2008-07-25T19:47:26 <ThomasWaldmann> search it yourself :)
2008-07-25T19:47:33 <xorAxAx> i found that the name was chosen on 30th of may
2008-07-25T19:47:40 <xorAxAx> but i cannot find any forking discussion
2008-07-25T19:47:54 <xorAxAx> it seems to me that this may happened at some unnamed lug meeting
2008-07-25T19:48:00 <xorAxAx> maybe dennda can remember
2008-07-25T19:48:38 <xorAxAx> (just pointing out: it doesnt make sense to revert or change anything regarting emeraldtree, people should just be aware that design wasnt very open at the beginning, from my and johills point of view)
2008-07-25T19:49:09 <ThomasWaldmann> iirc there was quite some discussion about choosing the right xml/tree lib
2008-07-25T19:49:31 <ThomasWaldmann> finding that about every lib has its problems
2008-07-25T19:50:06 <ThomasWaldmann> and you and johill didn't provide any alternative / better idea back then either, so what?
2008-07-25T19:50:31 <xorAxAx> of course, we had a few alternatives
2008-07-25T19:50:46 <xorAxAx> but our suggestions werent very fruitful (this already started in the soc webapp)
2008-07-25T19:50:59 <xorAxAx> i dont want to say that the choice is inherently bad ...
2008-07-25T19:51:10 <xorAxAx> or should be made undone
2008-07-25T19:51:32 <ThomasWaldmann> it's pretty pointless to repeat the discussion now, except if you have more to say now than you had 2 months ago
2008-07-25T19:52:05 <waldi> oh no ...
2008-07-25T19:52:05 <xorAxAx> i dont want to continue the discussion
2008-07-25T19:52:06 <ThomasWaldmann> and for that i suggest a wiki page, not irc (logs)
2008-07-25T19:52:13 <xorAxAx> waldi decided on 03-31th publicly that he needs some modifications
2008-07-25T19:52:32 <xorAxAx> sigh, you dont understand my point
2008-07-25T19:53:08 <ThomasWaldmann> get more concrete :) on that page.
2008-07-25T19:53:58 <xorAxAx> my point is that there was no discussion about the issues of forking (or i cant find it) -- so i get the impression that people are not completly aware of the burden ...
2008-07-25T19:54:35 <xorAxAx> ah, found it
2008-07-25T19:56:38 <xorAxAx> ~ 30 minutes of waldi and zenhase talking :)
2008-07-25T19:57:11 <xorAxAx> ok, anyway
2008-07-25T19:57:27 <xorAxAx> waldi: do you know how fast the c-version of etree is compared to the pure python one?
2008-07-25T19:57:39 <waldi> xorAxAx: nope
2008-07-25T19:57:59 <xorAxAx> because they didnt write parsing in python anyway ... so i wonder if there is a large factor speedup at all
2008-07-25T19:58:37 * ThomasWaldmann guesses it is faster and has more bugs :)
2008-07-25T19:58:38 <waldi> they use pyexpat
2008-07-25T19:59:20 <waldi> what needs to be done to support pickling of objects implemented in c?
2008-07-25T20:00:15 <xorAxAx> waldi: something along the lines of getstate/setstate
2008-07-25T20:00:40 <ThomasWaldmann> btw, looking at the recent benchmarks of waldi, it doesn't look like we have a speed problem
2008-07-25T20:00:55 <xorAxAx> no, the maintenance and interop ones are bigger
2008-07-25T20:01:02 <xorAxAx> with the first one being the biggest
2008-07-25T20:01:25 <ThomasWaldmann> you mean merging ET changesets?
2008-07-25T20:01:35 <xorAxAx> yeah
2008-07-25T20:02:24 <ThomasWaldmann> we'll see. as it is also in stdlib there maybe won't be very big changes.
2008-07-25T20:04:48 <waldi> ThomasWaldmann: no changes since the last tag
2008-07-25T20:05:19 <xorAxAx> sometimes one has the impression that effbot learned programming before docstrings started to exist :)
2008-07-25T20:05:50 <ThomasWaldmann> he's using a different doc tool
2008-07-25T20:06:21 <waldi> he is using doxygen which supports this form of docs for all elements including variables
2008-07-25T20:07:12 <xorAxAx> waldi: yeah, its probably more sound to use only one way but maybe merging would be more useful (e.g. you can use help() o nthe prompt)
2008-07-25T20:07:25 <xorAxAx> s/merging/ both ways of comments and docstrings/
2008-07-25T20:08:05 <waldi> it is pssible to convert them to docstrings
2008-07-25T20:08:15 <waldi> the tool will still find them
2008-07-25T20:09:17 <xorAxAx> one reason for an upcoming issue - imagine somebody writes to write a trafo that uses xpath or xslt
2008-07-25T20:09:31 <xorAxAx> then moin needs to convert the trees internally and use emeraldtrees and etrees:
2008-07-25T20:09:34 <xorAxAx> "There’s also an independent implementation, lxml.etree, based on the well-known libxml2/libxslt libraries. This adds full support for XSLT, XPath, and more."
2008-07-25T20:09:49 <xorAxAx> but i guess the conversion isnt too bald
2008-07-25T20:09:50 <xorAxAx> bad
2008-07-25T20:11:55 <waldi> i doubt that lxml could do xslt on a tree provided by xml.etree
2008-07-25T20:14:20 <xorAxAx> why not? you just need a copy constrcutor for that that implements the protocol
2008-07-25T20:17:27 <waldi> yeah, it needs to copy the complete tree
2008-07-25T20:18:21 <waldi> okay, lxml is written in c and this weird python/c mix
2008-07-25T20:20:50 <xorAxAx> cython, yeah
2008-07-25T22:16:26 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4035:e5a109d82e36 1.8-dom-bblank/MoinMoin/converter2/ (include.py _tests/test_include.py): Include converter - Add XPointer parser
2008-07-25T22:16:26 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4036:7b08b753c014 1.8-dom-bblank/MoinMoin/converter2/_tests/test_creole_in.py: Tests for Creole input converter - Add new macro tests
2008-07-25T22:16:27 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4037:9c7c484a136f 1.8-dom-bblank/MoinMoin/converter2/include.py: Include converter - Parse XPointer data, support page regex and limits
2008-07-25T22:22:57 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3776:5b6ba9b8158d 1.7/MoinMoin/parser/text_moin_wiki.py: wiki parser: avoid IndexError for empty shebang line
2008-07-25T22:24:49 <CIA-53> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2669:91f4c54406f1 1.6/MoinMoin/parser/text_moin_wiki.py: wiki parser: avoid IndexError for empty shebang line
2008-07-25T22:51:27 <CIA-53> Florian Krupicka <florian.krupicka@googlemail.com> default * 3938:b8ea98d74d8d 1.8-wsgi-fkrupicka/MoinMoin/ (3 files in 2 dirs): Fixed LDAP auth: Use of MoinMoin.auth.handle_login instead of request.handle_auth / SLAPD without cleartext-support breaks tests
2008-07-25T22:51:30 <zenhase> *phew*
2008-07-25T22:51:36 <zenhase> this was a long bug hunt
2008-07-25T22:51:57 <zenhase> my slapd wasn't very keen on plaintext-passwords
2008-07-25T22:52:23 <zenhase> now the ldap-tests run fine too
2008-07-25T23:03:36 <ThomasWaldmann> great :)
2008-07-25T23:05:03 <zenhase> i wonder why this plaintext password issue came up
2008-07-25T23:05:37 <ThomasWaldmann> zenhase: can't you just configure the slapd so it is happy with plaintext for the tests?
2008-07-25T23:05:54 <zenhase> ThomasWaldmann: i am not configuring the slapd, the tests do ;)
2008-07-25T23:06:04 <zenhase> but i don't know if there is an option
2008-07-25T23:06:37 <zenhase> i wonder why noone else came across this before
2008-07-25T23:10:36 <ThomasWaldmann> my slapd works with the tests
2008-07-25T23:10:44 <zenhase> what version?
2008-07-25T23:11:00 <zenhase> hmm
2008-07-25T23:11:07 <zenhase> could also be the osx slapd
2008-07-25T23:11:09 <ThomasWaldmann> 2.4.9 (ubuntu)
2008-07-25T23:11:14 <zenhase> i guess it's heavily patched
2008-07-25T23:11:18 <zenhase> 2.3.xx here
2008-07-25T23:11:28 <zenhase> 2.3.27
2008-07-25T23:14:30 <zenhase> it's strange, despite the man page saying that rootpw in slapd.conf can be a plain cleartext password, it bailed out on me on that. invalid user for root when the data gets populated from the ldif
2008-07-25T23:21:13 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4038:27e00bf23e7f 1.8-dom-bblank/MoinMoin/converter2/link.py: Link converter - Support complete wiki links
2008-07-25T23:21:14 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4039:43b5800df96e 1.8-dom-bblank/MoinMoin/converter2/_wiki_macro.py: Pseudo-macro handling - Add some more Include arguments
2008-07-25T23:21:14 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4040:eeff4a781e6f 1.8-dom-bblank/MoinMoin/converter2/include.py: Include converter - Support heading
2008-07-25T23:21:15 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4041:d98dd3f34168 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Fix TOC maximum level
2008-07-25T23:29:18 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4042:b54e64016c91 1.8-dom-bblank/MoinMoin/converter2/link.py: Link converter - Support mailto links
2008-07-25T23:35:04 <CIA-53> Bastian Blank <bblank@thinkmo.de> default * 4043:9822f5dcba6e 1.8-dom-bblank/MoinMoin/converter2/html_out.py: HTML output converter - Create img-tag for possible images
2008-07-25T23:39:22 <waldi> zenhase: which error?
2008-07-25T23:39:25 <ThomasWaldmann> waldi: better :)
2008-07-25T23:39:54 <waldi> i don't know a way to disable the CLEARTEXT pw schema
2008-07-25T23:40:35 <waldi> half of the include features works now ...
2008-07-25T23:46:11 <ThomasWaldmann> {{drawing: emits something although drawing: is not supported yet
2008-07-25T23:51:46 <ThomasWaldmann> waldi: a recursive include just vanishes
2008-07-25T23:52:42 <waldi> ThomasWaldmann: yep, its documented MoinMoin: MoinMoinChat/Logs/moin-dev/2008-07-25 (last edited 2008-07-24 22:15:02 by IrcLogImporter)
