2007-07-16T00:00:14 <grzywacz> I'm not sure if xmlrpc is surge protected...
2007-07-16T00:00:36 <grzywacz> request/__init__.py:247
2007-07-16T00:01:23 <xorAxAx> possible that it calls it later
2007-07-16T00:02:11 <grzywacz> Well, maybe.
2007-07-16T00:02:48 <xorAxAx> i mean, you are right
2007-07-16T00:03:21 <xorAxAx> grzywacz: your test is bad, it also works if _save_file is removed
2007-07-16T00:03:48 <xorAxAx> you need to check for the page's existence
2007-07-16T00:03:59 <xorAxAx> and delete it beforehand if it exists
2007-07-16T00:07:00 <grzywacz> :S
2007-07-16T00:07:27 <xorAxAx> or you ensure that _save_file is called if you dont abort it
2007-07-16T00:07:34 <xorAxAx> choose the way you prefer
2007-07-16T00:09:14 <grzywacz> The second one. ;)
2007-07-16T00:42:06 <grzywacz> AccessDenied: You are not allowed to delete this page!
2007-07-16T00:42:08 <grzywacz> hah
2007-07-16T00:42:20 <grzywacz> How do I elevate my privileges in tests? :D
2007-07-16T00:42:32 <xorAxAx> see the packages test :)
2007-07-16T00:42:38 <grzywacz> ok
2007-07-16T00:42:49 <xorAxAx> and maybe factor it out
2007-07-16T00:43:30 <grzywacz> Is _test/common.py ok?
2007-07-16T00:43:35 <grzywacz> *_tests
2007-07-16T00:48:19 <xorAxAx> yes
2007-07-16T00:49:46 <grzywacz> Unchanged: You did not change the page content, not saved!
2007-07-16T00:49:49 <grzywacz> hrmpf
2007-07-16T00:49:53 <grzywacz> This page should be deleted...
2007-07-16T00:50:25 <xorAxAx> did you find a bug? :)
2007-07-16T00:50:45 <grzywacz> No idea. I delete a page, then save it with the same contents as before, but it fails.
2007-07-16T00:50:54 <grzywacz> Maybe it's supposed to work like that, dunno. :)
2007-07-16T00:51:10 <xorAxAx> hmm
2007-07-16T00:51:11 <grzywacz> deletePage() is called on it for sure ;)
2007-07-16T00:51:25 <xorAxAx> deleted pages should have the text "deleted" in them
2007-07-16T00:51:33 <grzywacz> hm, true
2007-07-16T00:51:34 <xorAxAx> as the text
2007-07-16T00:52:06 <grzywacz> I use separate PageEditor objects to make sure there's no state shared.
2007-07-16T00:53:48 <xorAxAx> well, have a look at the page directory
2007-07-16T00:54:22 <xorAxAx> maybe its a bug in the itemcache (that will be gone soon)
2007-07-16T00:56:06 <grzywacz> It has "ThisIsSomeStupidTestPageText!" in it.
2007-07-16T00:56:09 <grzywacz> On disk, that is.
2007-07-16T00:56:16 <grzywacz> Doesn't look like "deleted" to me. :X
2007-07-16T00:56:38 <grzywacz> xorAxAx, I hope you've noticed the code I had pasted you
2007-07-16T00:58:11 <xorAxAx> yes of course
2007-07-16T00:58:18 <xorAxAx> it looks correct
2007-07-16T00:58:29 <xorAxAx> so it didnt delete?
2007-07-16T00:58:33 <xorAxAx> thats weird
2007-07-16T00:58:40 <xorAxAx> maybe you can trace it in the debugger
2007-07-16T00:58:59 <grzywacz> Weird, because delete works in this branch for sure. ;)
2007-07-16T01:02:20 <grzywacz> Huh. Is it normal that current points to a revision which has no file in revisions?
2007-07-16T01:05:59 <xorAxAx> yes
2007-07-16T01:06:22 <xorAxAx> hmm
2007-07-16T01:06:24 <xorAxAx> ah, i remember
2007-07-16T01:06:28 <xorAxAx> delete is not written to disk
2007-07-16T01:06:46 <xorAxAx> now i wonder how reverting works if there is such a protection
2007-07-16T01:08:19 <xorAxAx> elif newtext == self.get_raw_body():
2007-07-16T01:08:33 <xorAxAx> i think get_raw_body should return an empty string for a deleted page
2007-07-16T01:08:46 <xorAxAx> indeed, thats very weird
2007-07-16T01:08:51 <xorAxAx> try to find out, why it doesnt :)
2007-07-16T01:09:00 <xorAxAx> (in the saveText method)
2007-07-16T01:32:08 <grzywacz> http://it.slashdot.org/it/07/07/14/2229250.shtml
2007-07-16T01:32:09 <grzywacz> ...
2007-07-16T01:33:33 <grzywacz> http://it.slashdot.org/comments.pl?sid=250341&cid=19864117 :D
2007-07-16T01:34:16 <xorAxAx> gah, stupid article
2007-07-16T01:34:22 <xorAxAx> outlook 2003 doenst convert automatically
2007-07-16T01:34:30 <xorAxAx> it asks you
2007-07-16T01:34:44 <xorAxAx> nice comment :)
2007-07-16T01:35:11 <xorAxAx> grzywacz: did you find out why get_raw_body returns something wrong?
2007-07-16T01:35:31 <grzywacz> No, I ended up reading /. >_>
2007-07-16T01:35:48 <xorAxAx> i will ask again tomorrow, good night :)
2007-07-16T01:35:55 <xorAxAx> well, i mean today
2007-07-16T01:36:06 <grzywacz> Night, I'm staying a few more hours, so I'll take a look for sure.
2007-07-16T01:44:42 <grzywacz> xorAxAx, btw. is this a normal task, or a test? ;-)
2007-07-16T01:49:05 <xorAxAx> ???
2007-07-16T01:49:23 <xorAxAx> well, it looks like a bug in moin so we need to find where it is
2007-07-16T01:50:01 <grzywacz> ack, sleep well ;-)
2007-07-16T01:50:16 <grzywacz> I've just found a silly bug in my code. ;)
2007-07-16T01:59:20 <grzywacz> -_-;
2007-07-16T01:59:46 <grzywacz> deletePage() saves text "deleted" with kw argument deleted=True, which means: "@keyword deleted: if True, then don't save page content (used by DeletePage, default: False)"
2007-07-16T02:00:03 <xorAxAx> 01:06:28 < xorAxAx> delete is not written to disk
2007-07-16T02:00:18 <xorAxAx> 01:08:19 < xorAxAx> elif newtext == self.get_raw_body():
2007-07-16T02:00:18 <xorAxAx> 01:08:32 < xorAxAx> i think get_raw_body should return an empty string for a deleted page
2007-07-16T02:00:35 <grzywacz> I know, I'm commenting on general weirdness.
2007-07-16T02:00:37 <grzywacz> ;)
2007-07-16T02:09:20 <grzywacz> maybe Page.py:379
2007-07-16T02:09:23 <grzywacz> Not tested, though.
2007-07-16T02:09:37 <grzywacz> But I reuse the same request in my test case.
2007-07-16T02:27:57 <grzywacz> xorAxAx, it may be ItemCache after all, as you have already guessed.
2007-07-16T02:29:55 <grzywacz> xorAxAx, call chain seems to reach Page's get_rev, which asks the cache for data
2007-07-16T02:30:11 <grzywacz> xorAxAx, the body is of course None, which causes get_body() to try to load the page from disk
2007-07-16T02:31:01 <grzywacz> Debug print indicates that cache is being used for return value.
2007-07-16T08:27:59 <ThomasWaldmann> moin
2007-07-16T09:09:16 <vpv> ThomasWaldmann: moin. any idea why the wiki xmlrpc is taking so long?
2007-07-16T09:10:05 <vpv> not sure if you saw our discussion yesterday, but importing the man pages from Fedora's updates-testing took almost 80 minutes
2007-07-16T10:13:28 <ThomasWaldmann> vpv: how many items have that been?
2007-07-16T10:14:06 <vpv> I guess something around 150
2007-07-16T10:14:32 <vpv> but the weird thing is, you only need to have maybe 2 or 3 pages in the multicall and it's still slow
2007-07-16T10:14:43 <ThomasWaldmann> server method and machine type?
2007-07-16T10:15:47 <vpv> I have moin running as standalone on i686 (if that's what you meant)
2007-07-16T10:16:14 <vpv> the xmlrpc calls are made on this same machine, so I don't have a separate server, that might be one reason
2007-07-16T10:18:21 <ThomasWaldmann> a single manpage takes about 30s?
2007-07-16T10:19:24 <vpv> I should actually test that again, but there's a weird time when it "stalls" by looking at the cpu usage in every xmlrpc call
2007-07-16T10:21:09 <ThomasWaldmann> you could write some timestamps to the logging
2007-07-16T10:26:07 <zenhase> moin
2007-07-16T10:31:54 <vpv> ok, one man page takes about 40 seconds, of which maybe 35 is executing the multicall
2007-07-16T10:45:27 <vpv> now I added timing, it actually takes 27 seconds to execute the multicall which has applytoken and one putpage. most of that is just cpu idle time
2007-07-16T11:24:49 <xorAxAx> ThomasWaldmann: have you seen the abovely described item cache bug?
2007-07-16T11:25:21 <xorAxAx> vpv: well, you need to profile to see where it stalls
2007-07-16T12:21:44 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2116:8b21c6c44de8 1.6/wiki/server/moin.fcg: moin.fcg: add example for running under / url
2007-07-16T12:29:12 <xorAxAx> it should be fixed in 1.6 but not in 1.7 IMHO
2007-07-16T12:30:31 <vpv> xorAxAx: you mean the stalling or what?
2007-07-16T12:30:43 <vpv> or Thomas' commit?
2007-07-16T12:30:59 <xorAxAx> the item cache bug
2007-07-16T12:31:07 <xorAxAx> so neither of both
2007-07-16T12:31:24 <vpv> oh :D
2007-07-16T12:31:47 <vpv> yeah, I should have read backlog better
2007-07-16T12:34:50 <ThomasWaldmann> xorAxAx: yes, but I didnt debug it yet
2007-07-16T12:35:24 <xorAxAx> well, we have a test case
2007-07-16T12:35:34 <xorAxAx> grzywacz: is it checked in somewhere?
2007-07-16T12:37:24 <grzywacz> xorAxAx, no, wait
2007-07-16T12:38:01 <grzywacz> I'll commit it with debugging prints left, maybe will be helpful for now.
2007-07-16T12:38:59 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2393:d1b05e0eb6ea 1.7-jabber-knowak/MoinMoin/_tests/test_sourcecode.py: test_sourcecode: save some ms by some more excludes
2007-07-16T12:39:01 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2394:656685f94105 1.7-jabber-knowak/MoinMoin/_tests/test_sourcecode.py: test_sourcecode: only test recently changed stuff, use a single os.stat call's results
2007-07-16T12:39:04 <CIA-27> moin: Johannes Berg <johannes AT sipsolutions DOT net> * 2395:a7f4b02f5fdb 1.7-jabber-knowak/MoinMoin/ (5 files in 5 dirs):
2007-07-16T12:39:07 <CIA-27> moin: fix the quoting mess
2007-07-16T12:39:09 <CIA-27> moin: * double double-quotes (") to get a double quote inside quotes
2007-07-16T12:39:11 <CIA-27> moin: * quote things correctly when needed (not just "%s")
2007-07-16T12:39:13 <CIA-27> moin: * add and fix unit tests for this quoting
2007-07-16T12:39:15 <CIA-27> moin: * formatter must quote the name before it can pass it to resolve_wiki
2007-07-16T12:39:26 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2396:0c56b1468e78 1.7-jabber-knowak/MoinMoin/ (6 files in 5 dirs): Merge main.
2007-07-16T12:39:33 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2397:4bc95d828197 1.7-jabber-knowak/MoinMoin/ (PageEditor.py _tests/test_PageEditor.py): Whitespace fixes.
2007-07-16T12:39:38 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2398:9c0c2b782737 1.7-jabber-knowak/jabberbot/xmlrpcbot.py: Typo fix.
2007-07-16T12:39:45 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2399:d915ef70ebc9 1.7-jabber-knowak/ (27 files in 17 dirs): Merge main.
2007-07-16T12:39:50 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2400:d2487ea5bcf2 1.7-jabber-knowak/jabberbot/xmlrpcbot.py: Merge devel.
2007-07-16T12:39:58 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2401:0813af0f2425 1.7-jabber-knowak/MoinMoin/_tests/common.py: Refactor some common code into a seaparate file.
2007-07-16T12:40:04 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2402:212cd3474046 1.7-jabber-knowak/MoinMoin/PageEditor.py: Don't send PageDeletedEvent twice, silly...
2007-07-16T12:40:04 <xorAxAx> zenhase: so
2007-07-16T12:40:12 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2403:ca02fd95fe59 1.7-jabber-knowak/MoinMoin/_tests/test_PageEditor.py: "Fixed" PageEditor test.
2007-07-16T12:40:14 <xorAxAx> zenhase: have you thought about the discussion we had on saturday?
2007-07-16T12:40:20 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2404:c3d3d7b7a048 1.7-jabber-knowak/MoinMoin/ (PageEditor.py _tests/test_PageEditor.py _tests/common.py): Merge devel.
2007-07-16T12:40:31 <grzywacz> ThomasWaldmann, that would be _tests/test_PageEditor.py in my branch now.
2007-07-16T12:44:21 <ThomasWaldmann> grzywacz: tests finished: 6 passed
2007-07-16T12:44:41 <ThomasWaldmann> (btw, test_sourcecode fails :)
2007-07-16T12:55:28 <ThomasWaldmann> http://moinmo.in/FeatureRequests/SubpagesInheritDifferentAclFromParent looks like this use case was overlooked
2007-07-16T12:56:26 <xorAxAx> well
2007-07-16T12:56:36 <xorAxAx> IMHO its ugly by design
2007-07-16T12:56:54 <xorAxAx> other hierarchic ACL systems dont know this distinction either
2007-07-16T12:58:33 <ThomasWaldmann> the problem is that once hierarchic acl support is switched on, you get it (even if you dont want it)
2007-07-16T12:58:57 <ThomasWaldmann> (see that protected parent page with unprotected child page example)
2007-07-16T12:59:11 <xorAxAx> ?
2007-07-16T12:59:31 <xorAxAx> it only changes the ACL of the child page and not of the parent page
2007-07-16T12:59:43 <xorAxAx> (the effective ACL)
2007-07-16T12:59:57 <ThomasWaldmann> yeah, that's the problem
2007-07-16T12:59:58 <xorAxAx> yes, that will change semantics of course so you might need to reorder the page
2007-07-16T13:00:01 <xorAxAx> s
2007-07-16T13:00:25 <xorAxAx> i dont see a problem here, its a one time task and backwards compatible because its disabled by default
2007-07-16T13:00:56 <ThomasWaldmann> the problem is when you want to have ProtMainPage/NotProtSubPage
2007-07-16T13:01:45 <xorAxAx> well, multiple unprotected pages, yes
2007-07-16T13:01:52 <ThomasWaldmann> (for some single subpage, you can of course just put the default acl onto it, but if you have many of those subpages, this gets annoying)
2007-07-16T13:02:12 <xorAxAx> hm
2007-07-16T13:02:51 <xorAxAx> so you would have 2 acls on every page?
2007-07-16T13:03:08 <xorAxAx> and the recursive check would check only the subacl one?
2007-07-16T13:03:26 <xorAxAx> what if there is an acl and some effective subacls from the parent page?
2007-07-16T13:03:51 <xorAxAx> which evaluation order guarantees that you dont lose any features
2007-07-16T13:04:14 <ThomasWaldmann> if you already have an acl, you dont look any further, do you?
2007-07-16T13:04:44 <xorAxAx> only if i have a match
2007-07-16T13:05:08 * ThomasWaldmann didnt use it yet
2007-07-16T13:05:18 <xorAxAx> so its basically chaining all ACLs
2007-07-16T13:06:08 <ThomasWaldmann> i guess it could just use the subacl if it is there, otherwise use the acl
2007-07-16T13:06:46 * ThomasWaldmann isnt too happy with this getting more and more complex, but well, this seems to be the price of having hacls
2007-07-16T13:08:00 <ThomasWaldmann> (and the subacl only counts for sub-pages, not for the page it is on)
2007-07-16T13:08:37 <xorAxAx> the current system isnt very complex
2007-07-16T13:08:49 <xorAxAx> but the proposed one would basically touch the whole ACL chain
2007-07-16T13:08:56 <xorAxAx> because currently only one ACL is known
2007-07-16T13:09:05 <xorAxAx> and thats why i dislike it :)
2007-07-16T13:09:09 <ThomasWaldmann> (that way, you can also protect all subpages by a subacl and leave the main page on default acl by not using a acl)
2007-07-16T13:09:11 <xorAxAx> windows users dont need the feature either
2007-07-16T13:10:05 <ThomasWaldmann> they don't have comment pages :D
2007-07-16T13:11:17 <xorAxAx> comment pages are orthogonal and dont benefit from such a change
2007-07-16T13:11:24 <xorAxAx> because they are one subpage per n parent pages
2007-07-16T13:11:32 <xorAxAx> which would mean setting n acls anyway
2007-07-16T13:12:25 <ThomasWaldmann> well, if you want to leave it as is, maybe document this limitation so we won't get bug reports / feature requests about it from everybody
2007-07-16T13:13:20 <xorAxAx> ok
2007-07-16T13:13:38 <xorAxAx> well, if somebody presents a well-tested patch
2007-07-16T13:15:28 <ThomasWaldmann> "if you use this, child pages don't get the configured default acl, so if you protect the parent page, the child pages will get same protection as long as you don't put another acl on them"
2007-07-16T13:16:36 <ThomasWaldmann> assert find(somebody) != []
2007-07-16T13:17:24 <zenhase> xorAxAx: yeah, and i also talked with my mentor
2007-07-16T13:17:52 <xorAxAx> zenhase: is the result docuemnted somewhere?
2007-07-16T13:19:52 <zenhase> xorAxAx: not really
2007-07-16T13:20:00 <xorAxAx> sigh
2007-07-16T13:20:19 <xorAxAx> i mean, as long as it is not personal or related to internal goal settings, it should be made public ...
2007-07-16T13:20:30 <zenhase> well, i agreed on picking up on where i left off two weeks ago
2007-07-16T13:20:44 <xorAxAx> and whats the next step?
2007-07-16T13:22:20 <zenhase> implement 'show' and then 'edit' seperate from .send_page and try to get 2-3 small functions in the process that are tailored towards replacing send_page for those actions
2007-07-16T13:22:43 <zenhase> then augment bit by bit for the other actions
2007-07-16T13:23:57 <xorAxAx> so what about making parsers return strings?
2007-07-16T13:24:06 <xorAxAx> (and macros)
2007-07-16T13:24:13 <xorAxAx> that would simplify your steps
2007-07-16T13:24:28 <zenhase> erm
2007-07-16T13:24:54 <zenhase> i thought that was considered too much change
2007-07-16T13:24:58 <xorAxAx> no
2007-07-16T13:25:04 <xorAxAx> read the logs ...
2007-07-16T13:25:31 <xorAxAx> its a rather small change and completly compatible
2007-07-16T13:25:57 * xorAxAx sighs
2007-07-16T13:26:12 <xorAxAx> please ask if you unsure if you understood me/somebody else
2007-07-16T13:28:19 <zenhase> i am reading saturdays log again
2007-07-16T13:32:16 <zenhase> ok
2007-07-16T13:33:22 <zenhase> it must've slipped on me somehow
2007-07-16T13:33:44 <zenhase> well, ok. it's a wrapper for the parser then?
2007-07-16T13:34:04 <xorAxAx> yes, all parsers would be called through simple wrappers
2007-07-16T13:35:12 <zenhase> and then capture output and watch if the capture-buffer fills or there is anything returned from the parser, then pass it along to the next layer
2007-07-16T13:35:28 <xorAxAx> yes
2007-07-16T13:35:38 <xorAxAx> well, i would simply check for the return value
2007-07-16T13:35:41 <xorAxAx> being not None
2007-07-16T13:35:57 <zenhase> makes sense
2007-07-16T13:36:38 <xorAxAx> you will have to fix ~ 3 macros as well
2007-07-16T13:36:50 <zenhase> and the caching-formatter?
2007-07-16T13:36:57 <xorAxAx> hmm
2007-07-16T13:36:59 <xorAxAx> indeed
2007-07-16T13:37:08 <xorAxAx> well
2007-07-16T13:37:25 <xorAxAx> it should work fine if you keep the order
2007-07-16T13:37:46 <xorAxAx> and write everything to the template that you got
2007-07-16T13:38:02 <xorAxAx> but maybe ask ffesti if this assumption is correct, he wrote that formatter
2007-07-16T13:38:24 <zenhase> will do
2007-07-16T13:38:33 <ffesti> hmm, let me think
2007-07-16T13:38:41 <zenhase> ah :>
2007-07-16T13:38:52 * xorAxAx is gone shopping
2007-07-16T13:39:58 <ffesti> caching should be encapsulated into send_page content properly
2007-07-16T13:40:13 <ffesti> but you will have to use output capuring as it does request.write
2007-07-16T13:41:01 <ffesti> the only way to ge around that would be to modify the whole caching process to return a string
2007-07-16T13:41:19 <zenhase> hmm
2007-07-16T13:41:53 <zenhase> what about string-generators? i think many of those parser/formatter stuff could be more easily converted to a string-yielding scheme
2007-07-16T13:42:03 <ffesti> zenhase, we can just start with output captureing and then see how far we can push it down until its no longer needed
2007-07-16T13:42:16 <zenhase> ok
2007-07-16T13:42:23 <zenhase> sounds reasonable
2007-07-16T13:42:39 <ffesti> I think we should start with the send_page refactoring
2007-07-16T13:42:57 <ffesti> this is closer and more valuable for the Themimg stuff
2007-07-16T13:43:23 <ffesti> if this works we can try to get rid of the outputcapturing
2007-07-16T13:46:14 <zenhase> ok
2007-07-16T14:28:15 <dreimark> hi marian
2007-07-16T14:28:20 <dreimark> hi neagulm
2007-07-16T14:28:30 <neagulm> hello dreimark
2007-07-16T14:32:10 <dreimark> neagulm: the doc needs a bit mor improved, e.g. where to get spambayes and config/wikiconfig.py should get some hints about the vars to configure too
2007-07-16T14:41:49 <grzywacz> ThomasWaldmann, fails in one of my clones, works in another one.:D
2007-07-16T14:42:04 <grzywacz> ThomasWaldmann, maybe I should start with a fresh underlay ;S
2007-07-16T14:42:32 <grzywacz> s/underlay/pages/
2007-07-16T14:43:44 <ThomasWaldmann> grzywacz: if you expect page X there so you can delete it, you should create it before
2007-07-16T14:51:09 <dreimark> neagulm: I do agree in adding a multi categorisation classifier as your next step. but I do want documentations too. While completing/fixing the documentation of the spampbayes integretaion you could add at the same time the proposal for the multi categorisation tool on MoinMoin. The last time unicode support was an open request for orange. What is the status now? Ideas to ommit that problem?
2007-07-16T14:57:48 <dreimark> brb
2007-07-16T15:04:19 <xorAxAx> grzywacz: http://hg.alexanderweb.de/moin-1.7-jabber-knowak/rev/0813af0f2425 -- i dont see an edit of the packages test module ....
2007-07-16T15:04:52 <dreimark> bb
2007-07-16T15:12:42 <neagulm> dreimark, we could work around unicode support... :) while still maintaining the functionality
2007-07-16T15:14:06 <dreimark> neagulm: we need to be sure about that it goes well with unicode pages
2007-07-16T15:14:25 <dreimark> text e.g. chinese
2007-07-16T15:25:08 <grzywacz> xorAxAx, ok
2007-07-16T15:30:08 <xorAxAx> grzywacz: did you answer ThomasWaldmann's question?
2007-07-16T15:30:47 <grzywacz> erm
2007-07-16T15:31:09 <grzywacz> ThomasWaldmann, isn't it possible that it will be there?
2007-07-16T15:31:19 <grzywacz> ThomasWaldmann, like, left from a failed test run or something?
2007-07-16T16:18:00 <ThomasWaldmann> grzywacz: hm?
2007-07-16T16:19:12 <ThomasWaldmann> what I mean is that if the point of your test is testing DELETE, you should create it before or it won't be able to delete something
2007-07-16T16:28:33 <grzywacz> ThomasWaldmann, the point of the test is save :)
2007-07-16T16:29:40 <xorAxAx> grzywacz: well, it has to fail, does it fail?
2007-07-16T16:33:35 <grzywacz> xorAxAx, fail? It has to pass if handler returns Abort and save doesn't work. And the injected handler returns Abort.
2007-07-16T16:34:11 <xorAxAx> grzywacz: well, you had an itemcache related failure that ThomasWaldmann wanted to debug, where is it?
2007-07-16T16:35:42 <grzywacz> xorAxAx, turnos out it doesn't occur normally, only in my dev branch, probably because of broken pages/ dir
2007-07-16T16:36:08 <xorAxAx> hmm, what is broken about it?
2007-07-16T16:36:37 <grzywacz> The page with the same text is already there.
2007-07-16T16:36:44 <grzywacz> Thomas doesn't have this page, so it doesn't occur. ;)
2007-07-16T16:36:53 <xorAxAx> ah, well, then extend the test :)
2007-07-16T16:37:02 <xorAxAx> so it is reproducible
2007-07-16T16:38:58 <grzywacz> Isn't ItemCache going to vanish soon anyway?
2007-07-16T16:39:09 <xorAxAx> not in 1.6
2007-07-16T16:39:15 <grzywacz> ah, ok
2007-07-16T16:39:16 <xorAxAx> thats why he wants to debug it
2007-07-16T16:42:37 <grzywacz> ThomasWaldmann, please remove "return Abort" from this test for one run, then put it back
2007-07-16T16:42:42 <grzywacz> ThomasWaldmann, that's enough to reproduce
2007-07-16T16:45:11 <xorAxAx> and he needs to port it to 1.6
2007-07-16T16:46:07 <grzywacz> This triggers a bug. If he can find it in 1.7, then it should be enough, no?
2007-07-16T16:47:39 <xorAxAx> ?
2007-07-16T16:47:55 <xorAxAx> no, the unittest should be ported and then he should fixing the bug in 1.6
2007-07-16T16:48:07 <xorAxAx> otherwise you lose 50% of the usefulness of unittests
2007-07-16T16:48:29 <xorAxAx> (of course he can look at the issue in the 1.7 branch as well, but it should be reproducible using the test in 1.6)
2007-07-16T17:16:12 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2464:fe06a7d12b22 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Fix/add command line options, add timing to xmlrpc calls
2007-07-16T17:16:16 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2465:d1b05e0eb6ea 1.7-maninfo-vpv/MoinMoin/_tests/test_sourcecode.py: test_sourcecode: save some ms by some more excludes
2007-07-16T17:16:17 <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2466:656685f94105 1.7-maninfo-vpv/MoinMoin/_tests/test_sourcecode.py: test_sourcecode: only test recently changed stuff, use a single os.stat call's results
2007-07-16T17:16:25 <CIA-27> moin: Johannes Berg <johannes AT sipsolutions DOT net> * 2467:a7f4b02f5fdb 1.7-maninfo-vpv/MoinMoin/ (5 files in 5 dirs):
2007-07-16T17:16:28 <CIA-27> moin: fix the quoting mess
2007-07-16T17:16:33 <CIA-27> moin: * double double-quotes (") to get a double quote inside quotes
2007-07-16T17:16:37 <CIA-27> moin: * quote things correctly when needed (not just "%s")
2007-07-16T17:16:41 <CIA-27> moin: * add and fix unit tests for this quoting
2007-07-16T17:16:45 <CIA-27> moin: * formatter must quote the name before it can pass it to resolve_wiki
2007-07-16T17:16:51 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2468:71abd6011e9b 1.7-maninfo-vpv/MoinMoin/ (6 files in 5 dirs): merge from main
2007-07-16T17:16:56 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2469:82b2f4a794a7 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: cleanup imports
2007-07-16T17:17:04 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2470:3f722749018f 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: better comments
2007-07-16T17:17:12 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2471:c5f39c5e5a99 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: add some close()s and the evrdict
2007-07-16T17:17:20 <CIA-27> moin: Ville-Pekka Vainio <vpivaini AT cs DOT helsinki DOT fi> * 2472:01168acc421f 1.7-maninfo-vpv/MoinMoin/script/xmlrpc/manimport.py: Basics of marshalling / update handling, in a non-working state right now, will fix later
2007-07-16T17:18:01 <grzywacz> xorAxAx, if one removes the part with testing abort, I think this test should also work in 1.6
2007-07-16T17:18:46 <grzywacz> Lines 184-185 and 176-178
2007-07-16T18:02:45 <grzywacz> xorAxAx, looks like I can't really use xml:lang to detect language reliably
2007-07-16T18:02:59 <grzywacz> xorAxAx, psi doesn't even send it, and I see no option to change it
2007-07-16T18:06:37 <xorAxAx> grzywacz: then you have to use the language from the user profile
2007-07-16T18:06:54 <grzywacz> xorAxAx, yes, xml:lang is good in theory, but in practice doesn't seem to be used much... :S
2007-07-16T18:07:29 <xorAxAx> hehe
2007-07-16T18:07:46 <grzywacz> At least psi and pidging don't send it. I'm doing a little bit of testing with my friends. ;)
2007-07-16T18:07:52 <grzywacz> *pidgin
2007-07-16T18:08:16 <grzywacz> btw. is it also that hot in Germany? 35 degrees here all the time. :S
2007-07-16T18:08:27 <xorAxAx> 33° here outside
2007-07-16T18:09:32 <grzywacz> Almost as bad as here, then...
2007-07-16T18:10:07 <xorAxAx> but it should be colder tomorrow
2007-07-16T18:10:20 <xorAxAx> AC is running here
2007-07-16T18:10:45 <grzywacz> No AC here. ;)
2007-07-16T18:10:55 <zenhase> the weather is killing me ... every 2-3 days it's something different at the moment
2007-07-16T18:10:58 <xorAxAx> bought it last summer :)
2007-07-16T18:11:49 <grzywacz> xorAxAx, I have a fan, but it's loud and rapid air flow is rather disturbing ;>
2007-07-16T18:12:13 <zenhase> grzywacz: a fan is how i catched my current cold i think
2007-07-16T18:12:15 <xorAxAx> then increase the distance between you and the fan
2007-07-16T18:12:28 <xorAxAx> and slow it down
2007-07-16T18:12:57 <zenhase> you sit in the cooling refreshing stream all day ... and end up in bed, nose running
2007-07-16T18:13:21 <grzywacz> Hm, could happen. :S
2007-07-16T18:13:50 <grzywacz> Hm, maybe I could put it under my desk to cool my feet? :D
2007-07-16T18:17:57 <dreimark> grzywacz: put your feet in a bucket full of cold water
2007-07-16T18:20:51 <grzywacz> dreimark, will do so if temperature rises further ;)
2007-07-16T18:21:24 <grzywacz> Heh, I've just found the worst xmpp client ever written.
2007-07-16T18:23:10 <xorAxAx> which one? :)
2007-07-16T18:23:39 <dreimark> bbl
2007-07-16T18:23:59 <grzywacz> No idea, my friend uses it on his mobile. You'd expect it to conserve bandwidth, which is rather expensive, but no.
2007-07-16T18:24:10 <grzywacz> It sends 100+ bytes of obscure events per character he writes. ;)
2007-07-16T18:24:24 <grzywacz> And has just spammed me with ~50 presence "online" messages when he connected.
2007-07-16T18:24:26 <grzywacz> O_o;;;
2007-07-16T18:25:01 <grzywacz> Ah, it's called "im+". "+" probably stands for induced costs. ;>
2007-07-16T18:25:08 <xorAxAx> hah
2007-07-16T20:52:26 <dreimark> action=RenamePage is broken in 1.7
2007-07-16T21:38:19 <ThomasWaldmann> re
2007-07-16T21:54:23 <grzywacz> By me?
2007-07-16T22:07:07 <grzywacz> Works for me.
2007-07-16T22:14:59 <grzywacz> hi neagulm
2007-07-16T22:15:08 <neagulm> hello grzywacz
2007-07-16T22:20:05 <grzywacz> neagulm, it is now possible to abort savePage if you create an event handler for PagePreSaveEvent and return MoinMoin.events.Abort from it
2007-07-16T22:20:34 <grzywacz> Well, it is possible in my branch, feel fee to pull it if you're adventurous. ;)
2007-07-16T22:22:10 <neagulm> grzywacz, ok... :) This way I can replace the SecurityPolicy
2007-07-16T22:24:27 <neagulm> grzywacz, Currently I'm searching for ways to isolate the changes to MM.action.edit and MM.action and move them in a separate location, it would be nice if could do everything from MoinMoin.events.classifier
2007-07-16T22:25:12 <grzywacz> neagulm, the event interface is really simple, so you can easily add events/return types by yourself or let me know what to add if you don't want to bother with this code
2007-07-16T22:26:22 <grzywacz> xorAxAx, as for i18n of the bot, is it ok to create dicts (1 per language) of translated strings, with original strings being keys?
2007-07-16T22:26:31 <xorAxAx> grzywacz: yes, makes sense
2007-07-16T22:27:48 <grzywacz> Ok, that's even easy to send with xmlrpc.
2007-07-16T22:36:48 <grzywacz> xorAxAx, ok, so basically I have to somehow mark strings in jabber bot as exportable to .po files; add an xmlrpc function returning dict for a specified language; make sure that the dict gets requested and stored in bot when appropriate; use it to get string prior to sending?
2007-07-16T22:37:34 <xorAxAx> grzywacz: i suggest that you simply transfer all strings
2007-07-16T22:38:13 <grzywacz> xorAxAx, well, ok. Makes little difference.
2007-07-16T22:38:35 <xorAxAx> you need to define a _ function in the bot
2007-07-16T22:38:40 <grzywacz> Sure.
2007-07-16T22:38:50 <xorAxAx> that fetches the dict lazily
2007-07-16T22:38:51 <grzywacz> To get string exported, is it enough to put it in _()?
2007-07-16T22:38:55 <xorAxAx> yes
2007-07-16T22:39:09 <grzywacz> Should I add jabberbot/ dir to some script which does that?
2007-07-16T22:39:23 <xorAxAx> somewhere, yes
2007-07-16T22:40:48 <grzywacz> ThomasWaldmann, :)
2007-07-16T22:45:04 <ThomasWaldmann> mk_POTFILES.py should find it
2007-07-16T22:48:16 <grzywacz> Ah, looks ok.
2007-07-16T22:49:33 <grzywacz> Hm.
2007-07-16T22:49:56 <grzywacz> xorAxAx, maybe a separate gettext domain would make sense, though. Otherwise I'll be fetching all Moin translations... -_-;
2007-07-16T22:50:11 <xorAxAx> yes, where is the problem?
2007-07-16T22:50:18 <xorAxAx> thats 64 kB per language
2007-07-16T22:50:27 <grzywacz> How many languages are there?
2007-07-16T22:50:34 <xorAxAx> 33
2007-07-16T22:50:49 <grzywacz> That's over 1MB. ;-)
2007-07-16T22:50:59 <grzywacz> (of pure data)
2007-07-16T22:50:59 <xorAxAx> the connectivity between the jabber bot and the wiki is assumed to be good if not perfect (localhost)
2007-07-16T22:51:19 <grzywacz> I still see this as a useless waste of resources.
2007-07-16T22:51:21 <grzywacz> *an
2007-07-16T22:52:55 <xorAxAx> well
2007-07-16T22:53:07 <xorAxAx> how do you plan to share the code? copy&paste?
2007-07-16T22:53:33 <grzywacz> Erm, no. I'd rather make it possible to query moin for a different batch of strings.
2007-07-16T22:54:23 <xorAxAx> ah
2007-07-16T22:54:25 <xorAxAx> well
2007-07-16T22:54:33 <grzywacz> There's a different problem as well.
2007-07-16T22:54:45 <xorAxAx> can you roundtrip any kind of label through gettext tools?
2007-07-16T22:54:45 <grzywacz> I must be able to get all strings and their translations at once.
2007-07-16T22:54:57 <xorAxAx> yes, that shoul be easy
2007-07-16T22:55:04 <xorAxAx> see the i18n code
2007-07-16T22:55:05 <grzywacz> There are domains, globally. And comments per string.
2007-07-16T22:55:21 <grzywacz> At least that what I know.
2007-07-16T22:55:36 <xorAxAx> yes, and moin is already using a comment label besides fuzzy
2007-07-16T22:55:37 <grzywacz> *that's
2007-07-16T22:55:51 <xorAxAx> so maybe this is easy, yes
2007-07-16T22:56:17 <grzywacz> Ok, I'll take an another look at i18n code and maybe read a little bit more about gettext.
2007-07-16T23:03:55 <grzywacz> oops, found some rotten comments in i18n/check_i18n.py 8)
2007-07-16T23:59:51 * neagulm needs to learn a lot more about mercurial :( MoinMoin: MoinMoinChat/Logs/moin-dev/2007-07-16 (last edited 2007-10-29 19:21:21 by localhost)
