2007-06-05T00:15:31 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2080:937da10cb5b3 1.7-jabber-knowak/MoinMoin/user.py: s/filter/filter_func/ to avoid conflict with builtin name
2007-06-05T00:20:53 <grzywacz> This was a bad day for science... ;)
2007-06-05T00:21:43 <ThomasWaldmann> lanius: logging.log is your friend
2007-06-05T00:25:00 <ThomasWaldmann> lanius: if not self.__metadata: -> if self.__metadata is None:
2007-06-05T00:25:13 <ThomasWaldmann> this also applies for some other similar places
2007-06-05T00:26:14 <ThomasWaldmann> if you just check with "if not", the {} case will also trigger that if-clause
2007-06-05T00:30:46 <grzywacz> Hm, fugly. xmlrpclib lets low-level exceptions (like socket.error) propagate without wrapping them in its own classes, like ProtocolError
2007-06-05T00:32:09 <grzywacz> Looks like I'll have to catch Exception anyway... Sad.
2007-06-05T00:38:07 <xorAxAx> good night
2007-06-05T00:40:59 <dreimark> good night too
2007-06-05T00:41:11 <grzywacz> night xorAxAx
2007-06-05T00:41:13 <grzywacz> and dreimark :)
2007-06-05T00:57:34 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2059:ce9cdc91bc60 1.7-storage-hwendel/MoinMoin/storage/ (external.py storage16.py): small fixes
2007-06-05T00:58:38 <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2060:d6e3f9e892c2 1.7-storage-hwendel/MoinMoin/storage/_tests/test_external.py: another unit test
2007-06-05T00:59:15 <lanius> ThomasWaldmann: generally the user stuff seems to be working
2007-06-05T00:59:34 <lanius> ThomasWaldmann: the integration in user.py is not really nice, but working
2007-06-05T01:16:15 <zenhase> hmm
2007-06-05T01:17:15 <zenhase> short pause nap
2007-06-05T01:26:23 * grzywacz is getting confused
2007-06-05T01:36:27 <grzywacz> Blah, func_name is read-only in python 2.3? Oh the horror...
2007-06-05T02:58:16 <CIA-27> moin: Florian Krupicka <florian.krupicka@googlemail.com> * 2046:2639cfa8b978 1.7-template-fkrupicka/ (2 files in 2 dirs): Starting to build up context dictionaries for genshi & first idea for a base-template
2007-06-05T02:59:20 <zenhase> hmm, should go to sleep now but i will update the wikipages later today with the ideas i got from talking to xorAxAx
2007-06-05T03:00:34 <zenhase> hmm .. i should split the changesets, this looks awful to track the changes ...
2007-06-05T03:00:47 <zenhase> *note down*
2007-06-05T03:00:53 <zenhase> gn8 then
2007-06-05T12:03:30 <ThomasWaldmann> moin
2007-06-05T12:06:28 <grzywacz> moin
2007-06-05T12:17:18 <zenhase> moin
2007-06-05T12:30:47 <ffesti> moin
2007-06-05T12:41:41 <dreimark> moin
2007-06-05T12:43:01 <johill> moin!
2007-06-05T13:16:04 <johill> dreimark: can you merge those {{{}}} fixes to 1.7?
2007-06-05T14:23:20 <dreimark> johill:yes
2007-06-05T14:26:27 <dreimark> and some more e.g. SystemInfo
2007-06-05T15:39:36 <johill> dreimark: thanks
2007-06-05T15:49:09 <zenhase> awful weather ...
2007-06-05T15:49:33 * zenhase is sweating all over
2007-06-05T15:51:06 * ffesti passes some ice cream to zenhase
2007-06-05T15:53:50 * zenhase rubs himself with the icecream
2007-06-05T15:54:07 <zenhase> *pssschhhhhh* *little clouds arise*
2007-06-05T15:54:30 <zenhase> ah, that was good
2007-06-05T15:57:39 <neagulm> hello
2007-06-05T16:05:21 <dreimark> moin neagulm, skip has agreed for talking
2007-06-05T16:05:50 <neagulm> dreimark, when is the meeting scheduled ?
2007-06-05T16:07:31 <dreimark> neagulm: classification project: #moin-dev, Tuesday, 05. Juni: 19:00 UTC (22:00 EET, 21:00 CET, 14:00 CDT)
2007-06-05T16:07:41 <dreimark> its our normal time
2007-06-05T16:08:00 <neagulm> perfect
2007-06-05T16:08:18 <neagulm> I'm working for the attachment part...
2007-06-05T16:08:25 <dreimark> fine
2007-06-05T16:08:48 <dreimark> I am back in something around 3h
2007-06-05T16:08:55 <neagulm> do you have some more suggestions related to attachment's or UI ?
2007-06-05T16:08:56 <neagulm> ok
2007-06-05T18:21:55 <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2018:1ecfbf05930a 1.7-classify-mneagul/MoinMoin/ (5 files in 4 dirs): Add support for page attachments
2007-06-05T18:54:35 <ThomasWaldmann> neagulm: read pep8 again
2007-06-05T18:56:22 <neagulm> ThomasWaldmann, spaces again ?
2007-06-05T18:56:32 <grzywacz> xorAxAx, are you there?
2007-06-05T18:56:52 <grzywacz> xorAxAx, I'd like to discuss the issue of ACLs in regard to notifications / event handling.
2007-06-05T19:00:34 <xorAxAx> grzywacz: yes, i am here :)
2007-06-05T19:03:27 <tw> fsck, my irssi on thinkmo hangs
2007-06-05T19:03:45 <grzywacz> xorAxAx, oh, good.
2007-06-05T19:04:14 <xorAxAx> tw: the screen or the irssi?
2007-06-05T19:04:23 <xorAxAx> tw: PON/POFF? :)
2007-06-05T19:04:33 <tw> irssi
2007-06-05T19:04:58 <tw> and I guess you mean XON/XOFF?
2007-06-05T19:05:03 <xorAxAx> yes %-)
2007-06-05T19:05:08 <xorAxAx> pon was ppp
2007-06-05T19:05:15 <grzywacz> xorAxAx, to begin with, I was wondering whether I should try to reuse the existing ACL concept, or create some new syntax to manage permissions.
2007-06-05T19:05:28 * xorAxAx never seriously used software flow control, my cables had enough wires
2007-06-05T19:05:51 <xorAxAx> grzywacz: why do you need new syntax?
2007-06-05T19:05:55 <xorAxAx> is there any new requirement?
2007-06-05T19:06:21 <grzywacz> I expect the number of "actions" defined in std ACLs may grow a little bit too large and harder to understand.
2007-06-05T19:06:32 <grzywacz> Right now there's not much that you can set access to.
2007-06-05T19:07:38 <grzywacz> But if you were to write/maintain an acl list of mixed page permissions and notification filters, it might get fugly.
2007-06-05T19:08:08 <xorAxAx> grzywacz: well, can you put these new requirements onto the wiki page?
2007-06-05T19:08:40 <grzywacz> xorAxAx, sure, will do it right now,
2007-06-05T19:08:57 <grzywacz> Let me get some tea first, though. :D
2007-06-05T19:09:12 <neagulm> ThomasWaldmann, in what file do I have problems with PEP8 ?
2007-06-05T19:14:13 <ThomasWaldmann> last diff, no blanks around operators and assignment
2007-06-05T19:25:27 <neagulm> ok, i'm going home. bbl
2007-06-05T19:27:15 <lanius> ThomasWaldmann: did you take a look at the user.py integration
2007-06-05T19:28:53 <ThomasWaldmann> i did a short test and it worked
2007-06-05T19:30:12 <xorAxAx> grzywacz: have you edited already?
2007-06-05T19:30:19 <grzywacz> Looking at preview.
2007-06-05T19:30:23 <grzywacz> one second
2007-06-05T19:30:34 <xorAxAx> ok
2007-06-05T19:30:45 <ThomasWaldmann> lanius: btw, storage/storage16 is somewhat redundant. how about storage/fs_moin16
2007-06-05T19:31:50 <grzywacz> xorAxAx, http://moinmoin.wikiwikiweb.de/JabberSupport#head-c1157d6dba96fec9a86d6a2d71550d5175813c5a
2007-06-05T19:31:54 <grzywacz> I hope it's understandable. ;)
2007-06-05T19:32:49 <ThomasWaldmann> lanius: 118 TODO: caching <- I guess we don't need to tune the 1.6 storage
2007-06-05T19:33:09 <ThomasWaldmann> neither with caches, nor with indexes
2007-06-05T19:34:43 <ThomasWaldmann> lanius: you can remove the todo from user.py header
2007-06-05T19:34:59 <ThomasWaldmann> and add your (C)
2007-06-05T19:35:45 <grzywacz> xorAxAx, it's just that I'd prefer not to add more than just a few "actions" to known ACLs
2007-06-05T19:36:10 <grzywacz> xorAxAx, how about the idea of controlling notifications based on "event classes"?
2007-06-05T19:36:14 <grzywacz> Not per-event-type?
2007-06-05T19:37:10 <lanius> ThomasWaldmann: ok
2007-06-05T19:38:06 <xorAxAx> grzywacz: i think for the first steps, checking for super user rights is enough
2007-06-05T19:38:32 <xorAxAx> grzywacz: in order to make this more flexible, one would need to introduce ACLs that are usable wiki-wide
2007-06-05T19:38:35 <xorAxAx> grzywacz: moin currently doesnt have that
2007-06-05T19:38:38 <grzywacz> Yes.
2007-06-05T19:38:45 <grzywacz> Hm, not?
2007-06-05T19:38:54 <grzywacz> Ah.
2007-06-05T19:39:00 <grzywacz> Right. Per-page. My bad.
2007-06-05T19:39:11 <lanius> ThomasWaldmann: the problem with the metadata is that even for writing you first need to read the metadata again
2007-06-05T19:39:20 <grzywacz> Well, if time permits, I might give it ago at the end of summer.
2007-06-05T19:39:38 <xorAxAx> grzywacz: yeah
2007-06-05T19:39:47 <xorAxAx> grzywacz: put it onto the todo list :)
2007-06-05T19:39:50 <grzywacz> xorAxAx, ok, I'll add a super-user check for event that shouldn't be sent to normal users
2007-06-05T19:40:01 <grzywacz> xorAxAx, also, I have a python-related question.
2007-06-05T19:41:08 <dreimark> someone hates helping on current state of Gallery2
2007-06-05T19:41:13 <grzywacz> xorAxAx, in xmlrpcbot.py, every function that handles xml rpc request must check the secret submitted with the request
2007-06-05T19:41:21 <xorAxAx> dreimark: hmm?
2007-06-05T19:41:25 <grzywacz> xorAxAx, adding the check to each of them seems to be a no-no to me.
2007-06-05T19:41:34 <xorAxAx> grzywacz: you can use the decorator pattern
2007-06-05T19:41:39 <grzywacz> Yes, wait.
2007-06-05T19:41:42 <dreimark> Oh, and only a inept dork puts the delay box in the sideshow viewer to Milliseconds, if you think such a 'feature' in a user interface is acceptable I suggest you should stop contributing to Open Source projects, please.
2007-06-05T19:41:47 <grzywacz> I tried this, but there's a problem, it seems.
2007-06-05T19:42:17 <xorAxAx> btw
2007-06-05T19:42:20 <grzywacz> Hand-made decoration doesn't seem to work, because SimpleXMLRPCServer registers functions using their names.
2007-06-05T19:42:32 <grzywacz> There are changed if you return a wrapped function.
2007-06-05T19:42:34 <xorAxAx> grzywacz: you just have to decorate your decorator
2007-06-05T19:42:42 <xorAxAx> with a decorator-decorator
2007-06-05T19:42:46 <grzywacz> Huh? ;)
2007-06-05T19:43:27 <xorAxAx> a decorator-decorator is used to persist the name and parameters
2007-06-05T19:43:47 <xorAxAx> the shortest decorator decorator is 3 lines, see the decorator syntax PEP, there should be one in there
2007-06-05T19:43:59 <grzywacz> Wait. This is supposed to work with python 2.3, no? ;)
2007-06-05T19:44:10 <xorAxAx> yes
2007-06-05T19:44:12 <xorAxAx> so? :)
2007-06-05T19:44:22 <xorAxAx> you dont need to use the decorator syntax to use decorators
2007-06-05T19:44:27 <grzywacz> Sure.
2007-06-05T19:44:50 <dreimark> bbl, 15mins
2007-06-05T19:55:32 <ThomasWaldmann> lanius: did you try [[SystemAdmin]] with your code?
2007-06-05T19:57:46 <ThomasWaldmann> or the notification code?
2007-06-05T19:58:49 <lanius> no, i tried login, preferences and registration
2007-06-05T20:12:33 <ThomasWaldmann> lanius: btw, your tests are under MoinMoin/user/_tests/*, shouldn't they be under MoinMoin/_tests/* ?
2007-06-05T20:12:47 <xorAxAx> ThomasWaldmann: no
2007-06-05T20:13:09 <xorAxAx> but maybe we should remove the underscore :)
2007-06-05T20:13:36 <xorAxAx> the idea is that the test is near to the covered code module
2007-06-05T20:14:03 <xorAxAx> so that you can find the particular test much easiert
2007-06-05T20:14:06 <xorAxAx> s/t$//
2007-06-05T20:14:41 <dreimark> bb
2007-06-05T20:14:56 <ThomasWaldmann> we have to handle exclusion for packaging and running them all in one go
2007-06-05T20:15:17 <xorAxAx> yes, the packaging has to filter them
2007-06-05T20:15:30 <xorAxAx> what do you mean by the latter?
2007-06-05T20:15:50 <xorAxAx> hmm, it doesnt /have/ to filter them (as we dont currently either) but it /could/ filter
2007-06-05T20:17:10 <ThomasWaldmann> does make test run them all if they are not at one place?
2007-06-05T20:17:33 <xorAxAx> yes
2007-06-05T20:17:45 <xorAxAx> btw, no
2007-06-05T20:17:53 <xorAxAx> make test doesnt run any test :)
2007-06-05T20:17:53 <xorAxAx> try it
2007-06-05T20:18:07 <xorAxAx> py.test is used
2007-06-05T20:18:15 <xorAxAx> as a command
2007-06-05T20:18:52 <ThomasWaldmann> ? that comment is not helpful
2007-06-05T20:19:07 <ThomasWaldmann> make test used to run all tests
2007-06-05T20:19:35 <xorAxAx> well, after you have read its output, feel free to put "py.test" into the makefile
2007-06-05T20:19:43 <xorAxAx> but i dont see the point
2007-06-05T20:19:58 <ThomasWaldmann> xorAxAx: backwards compatibility :)
2007-06-05T20:20:26 <xorAxAx> life long learning :-p
2007-06-05T20:26:08 <ThomasWaldmann> lanius: the module docstring in external.py is not sufficient
2007-06-05T20:26:55 <ThomasWaldmann> same for interfaces.py
2007-06-05T21:05:06 <dreimark> moin neagulm
2007-06-05T21:06:16 <neagulm> hello dreimark
2007-06-05T21:08:15 <dreimark> so we are waiting for skip
2007-06-05T21:09:32 <neagulm> ok
2007-06-05T21:10:44 <ne> neagulm: skip has mailed me that he has his xmlrpc interface partly working
2007-06-05T21:15:06 <dreimark> neagulm: and he plans to have the full api working in the next days
2007-06-05T21:15:21 <neagulm> dreimark, that is perfect... on more step in creating a modular interface
2007-06-05T21:21:20 <dreimark> neagulm: if we use sb for attachments
2007-06-05T21:22:19 <dreimark> we can only use it for documents of type text, or ?
2007-06-05T21:22:45 <dreimark> e.g. exif output is that usefull too?
2007-06-05T21:23:06 <dreimark> pdf could extract text and so on
2007-06-05T21:25:53 <neagulm> dreimark, not only for text attachment (or any type of supported files) but for any type of file: image/exe/ etc. We achieve this by using the first few bytes as a token, like in the current tokenizer
2007-06-05T21:26:38 <neagulm> dreimark, I do this only for unsupported attachments. For pdf/odt I will extract the data and send it to the classifier
2007-06-05T21:51:37 <neagulm> dreimark, ping
2007-06-05T21:59:55 <dreimark> ping neagulm
2007-06-05T22:00:14 <skipm> Neagulm, dreimark: my apologies for forgetting about the meetup
2007-06-05T22:01:30 <dreimark> hi skipm
2007-06-05T22:02:46 <neagulm> hello skip
2007-06-05T22:02:59 <dreimark> phonecall moment
2007-06-05T22:04:18 <neagulm> ok,
2007-06-05T22:05:18 <dreimark> bb
2007-06-05T22:05:52 <dreimark> skipm: did you tried neagulm code already or peeked into it?
2007-06-05T22:09:48 <neagulm> skipm, dreimark: I don't know how will SB be affected by MM attachments, they are currently not versioned
2007-06-05T22:10:51 <neagulm> skipm, dreimark: there is no direct dependency between Page revisions and page attachments
2007-06-05T22:12:53 <skipm> another phonecall moment... just a moment..
2007-06-05T22:15:48 <dreimark> neagulm: skipm: currently the only dependency is an ATT entry in the log file and that the name of the file could only once used on a page
2007-06-05T22:16:15 <skipm> dreimark: i have neither peeked at the code nor tried it
2007-06-05T22:17:13 <skipm> neagulm: wouldn't attachments be accepted or rejected before overwriting the preexisting attachment?
2007-06-05T22:18:11 <neagulm> skipm, I suggested to use SB only to classify pages and not to enforce a policy, just warn the user...
2007-06-05T22:18:37 <skipm> dreimark, neagulm: i have a mostly working standalone SB server which speaks XML-RPC. I've got more work to do, but hope to merge the branch I;m working on to CVS HEAD in the next few days
2007-06-05T22:19:13 <neagulm> skipm, enforcing a policy would be somehow in contradiction with the wiki philosophy
2007-06-05T22:19:17 <skipm> SB won't enforce anything, just score the submissions. You can choose to accept everything even if it tells you everything is spam...
2007-06-05T22:19:29 <neagulm> skipm,
2007-06-05T22:20:00 <neagulm> skipm, this is my current approach.
2007-06-05T22:20:49 <neagulm> skipm, please drop me a mail after committing into CVS HEAD, so that i can integrate it in MM
2007-06-05T22:21:31 <skipm> dreimark, neagulm: have you looked at this page? http://moinmoin.wikiwikiweb.de/SkipMontanaro/CoreServer
2007-06-05T22:22:02 <skipm> dreimark, neagulm: that is approximately what i've implemented where the only protocol plugin so far is XML-RPC
2007-06-05T22:22:54 <dreimark> did you have done some speed tests ?
2007-06-05T22:24:22 <neagulm> would it be problematic with attachments (supported ones: odt, pdf) because of their size ?
2007-06-05T22:24:25 <ThomasWaldmann> neagulm: if lanius succeeds, we will have revisioned "attachments" (file items) afterwards
2007-06-05T22:24:44 <skipm> dreimark: nope, nothing yet.
2007-06-05T22:26:11 <neagulm> ThomasWaldmann, will we have a dependency between a page revision and an attachment revision ?
2007-06-05T22:26:26 <ThomasWaldmann> no
2007-06-05T22:27:00 <dreimark> skipm: neagulm: if the server is running on the same system as the wiki(farm) maybe a direct pythonic connection would be fine too
2007-06-05T22:27:09 <ThomasWaldmann> the only relation is subitem for converted attachments of a page
2007-06-05T22:28:02 <neagulm> ThomasWaldmann, ok
2007-06-05T22:28:15 <skipm> dreimark: what's a "direct pythonic connection"? you mean embed the SB core server into the wiki?
2007-06-05T22:28:38 <ThomasWaldmann> lanius: as you seem finished with users, are there open questions about 1.6 pages?
2007-06-05T22:28:53 <dreimark> not using xmlrpc, calling it as module
2007-06-05T22:29:45 <dreimark> e.g. currently I have three farming server
2007-06-05T22:30:07 <skipm> dreimark: that would work in theory as long as the wiki farm was a single long-lived process
2007-06-05T22:31:09 <skipm> dreimark: if it was running as a cgi or multiple processes you'd have database locking/consistency issues
2007-06-05T22:33:24 <dreimark> so the usage of xml-rpc will serialize the requests ? or do I need for each wiki a sb server?
2007-06-05T22:34:23 <skipm> dreimark: depends on how you want to manage things. if the wikis are related and the admins cooperative, probably one would do
2007-06-05T22:34:39 <lanius> ThomasWaldmann: locking is still not exactly clear, theoretically users need that as well
2007-06-05T22:34:57 <skipm> dreimark: on the other hand, if one wiki was about python programming and the other about sex toys you might have problems agreeing on what was spam. ;-)
2007-06-05T22:35:08 <dreimark> hehe
2007-06-05T22:36:14 <ThomasWaldmann> lanius: what kind of locking?
2007-06-05T22:36:21 <lanius> any kind
2007-06-05T22:36:29 <ThomasWaldmann> X)
2007-06-05T22:40:32 <ThomasWaldmann> lanius: the file backend can avoid read locks on the data and meta files, if the writers write to a tmpfile and then rename it to the target filename
2007-06-05T22:40:52 <dreimark> skipm, neagulm, xorAxAx I am not quite familiar with xml-rpc, are there limits which could be exceeded on farm systems
2007-06-05T22:41:03 <ThomasWaldmann> this is true for any posix system, but (of course) not on win32, there you'll need locking
2007-06-05T22:41:12 <ThomasWaldmann> see how the current backend does it
2007-06-05T22:41:14 <xorAxAx> dreimark: what kind of limits?
2007-06-05T22:42:01 <skipm> dreimark: xmlr-rpc normally makes a tcp connection per request so you don't have lots of open connections hanging around
2007-06-05T22:42:06 <xorAxAx> ThomasWaldmann: windows can do file deletions if the file is opened using some special cals
2007-06-05T22:42:55 <dreimark> skipm: ok, the answer before I did the quest. How about concurrent access to different sb servers from different instances in a farm
2007-06-05T22:43:27 <neagulm> dreimark, sharing the same db ?
2007-06-05T22:43:54 <dreimark> not implicitly
2007-06-05T22:44:06 <ThomasWaldmann> xorAxAx: so rename will delete the target file, leaving the inode alive?
2007-06-05T22:44:33 <dreimark> neagulm: optional it should be able to share the same database
2007-06-05T22:45:04 <skipm> dreimark: i'm not familiar with wiki farms - is that one process managing multiple wikis? shouldn't be a problem in concept you just need to tell each wiki where to connect (host/port/path)
2007-06-05T22:46:05 <dreimark> yeah, a new config var will do that
2007-06-05T22:46:07 <xorAxAx> ThomasWaldmann: yes
2007-06-05T22:46:21 <xorAxAx> ThomasWaldmann: but unfortunately, this needs win32* libs
2007-06-05T22:46:42 <xorAxAx> or ctypes and hackery
2007-06-05T22:46:45 <neagulm> dreimark, you meant to do load-balancing between SB servers ?
2007-06-05T22:46:54 <xorAxAx> (the code is implemented in hg)
2007-06-05T22:48:11 <dreimark> neagulm: thats another option
2007-06-05T22:48:12 <ThomasWaldmann> lanius: aside from locking (I guess we'll see it better while implementing the backend), do you have plans where to start?
2007-06-05T22:48:13 <skipm> dreimark, neagulm: note that python's xml-rpc server doesn't do any sort of authentication i don't think - listening on localhost is probably the only viable way to avoid eventual abuse
2007-06-05T22:49:17 <lanius> ThomasWaldmann: i think i go the same way as with the user, start with the basic backend stuff compatible with 1.6 and then add necessary features to the external interface
2007-06-05T22:51:16 <xorAxAx> skipm: thats usually solved by using shared secret
2007-06-05T22:51:16 <xorAxAx> s
2007-06-05T22:51:49 <dreimark> skipm: do get it for me. if I do host two different wikis, i should have to sb servers running. Or can I exchange name of the database and do have only one sb server running?
2007-06-05T22:51:50 <ThomasWaldmann> Page is a bit more complex, it has revision, metadata, log, attachments, cache
2007-06-05T22:52:30 <lanius> right, i will begin with revision and metadata and leave the rest aside at first
2007-06-05T22:52:40 <grzywacz> sb is a separate server?
2007-06-05T22:53:11 <dreimark> grzywacz: application see SkipMontanaro/CoreServer
2007-06-05T22:53:16 <skipm> xorAxAx: i've used that before myself. I'm not sure I'd want to rely on it over the Internet where your secret can be snooped - i don't know if the xml-rpc server supports SSL out of the box
2007-06-05T22:53:41 <xorAxAx> skipm: yeah, ssl is hairy
2007-06-05T22:53:54 <xorAxAx> skipm: but thats an orthogonal issue somehow :)
2007-06-05T22:54:46 <skipm> grzywacz: yes - my core server is based on the SpamBayes POP3 proxy. I ripped out the POP3 stuff then implemented a simple plugin interface - xml-rpc is the first plugin
2007-06-05T22:54:55 <grzywacz> Nice.
2007-06-05T22:55:33 <skipm> note that i have never implemented a plugin-based system before, so those of you who ever decide to peek at the code must promise not to laugh. ;-)
2007-06-05T22:59:13 <dreimark> skipm: is the Training database somehow versioned
2007-06-05T22:59:58 <dreimark> moin rhoe
2007-06-05T23:00:43 <skipm> dreimark: no versioning - it hangs onto every message it is asked to score. you can then discard them or assign them as ham (good) or spam (bad) - this interface is untouched from the SB POP3 proxy and probably needs a tweak or two for better remote control
2007-06-05T23:01:20 <rhoe> moin dreimark
2007-06-05T23:01:30 <rhoe> moin altogether
2007-06-05T23:05:04 <dreimark> skipm: how much messages can it score at the same time?
2007-06-05T23:05:55 <ThomasWaldmann> hi rhoe
2007-06-05T23:06:10 <skipm> dreimark: i'm not sure what you're asking - how many messages per second or how many simultaneously?
2007-06-05T23:06:57 <dreimark> the last, because several users, several pages
2007-06-05T23:07:08 <neagulm_> I have problems with my internet connection :(
2007-06-05T23:07:42 <dreimark> neagulm_: oh,
2007-06-05T23:08:21 <dreimark> ThomasWaldmann: can you make now one of devlog file, so neagulm_ get it on the wiki
2007-06-05T23:08:31 <skipm> we can multi-thread it i suppose, but i suspect sequential access will be sufficient - let me make some measurements and send you some mail
2007-06-05T23:08:33 <neagulm_> dreimark, i hope i didn't miss some important stuff
2007-06-05T23:08:58 <skipm> neagulm_: you can read the chat logs just like everyone else. ;-)
2007-06-05T23:10:00 <dreimark> rhoe: thanks for inviting to the diploma speech
2007-06-05T23:11:56 <rhoe> yes. that's understood. however there are now a lof of refacotring tasks open to me..
2007-06-05T23:13:03 <skipm> dreimark, neagulm_: thanks again for patiently waiting for me, however i have a train to catch in about 15 minutes - any last questions for me?
2007-06-05T23:13:43 <skipm> dreimark, neagulm_: i will try and make some performance measurements and get things in shape well enough to merge my branch back to HEAD
2007-06-05T23:14:16 <neagulm_> skipm, Please send me a mail after finishing the code ;)
2007-06-05T23:14:28 <dreimark> :) the performace check is important with attachments too, 100MB
2007-06-05T23:14:52 <dreimark> or whatever we have to set a limit too
2007-06-05T23:14:52 <skipm> neagulm_: will do - although as you are aware, code will never be truly finished
2007-06-05T23:15:16 <skipm> dreimark: do you really see 100mb attachments? legitimate or spam?
2007-06-05T23:15:20 <dreimark> and documentation never * never
2007-06-05T23:15:52 <dreimark> I often see 100MB presentations
2007-06-05T23:16:38 <dreimark> ThomasWaldmann: lanius and xorAxAx have had a discussion about 4GB files in one of the last logs
2007-06-05T23:16:41 <neagulm_> and privacy, in case of SB out of DMZ
2007-06-05T23:16:41 <neagulm_> this is also important
2007-06-05T23:16:41 <neagulm_> privacy
2007-06-05T23:17:06 <xorAxAx> yes, moin users have 4 GB PPT files
2007-06-05T23:17:07 <skipm> dreimark: if they don't contain any plain text to munch on you can just arbitrarily trim them to a smaller size - as long as we get a size token and maybe a content-type
2007-06-05T23:17:34 <skipm> neagulm_: can you expand on privacy in an email?
2007-06-05T23:17:57 <dreimark> sometimes a summary in the wiki is fine
2007-06-05T23:18:03 <dreimark> after some emails
2007-06-05T23:18:25 <dreimark> or instead of ...
2007-06-05T23:18:45 <skipm> dreimark: sure wiki is fine too.
2007-06-05T23:18:50 <neagulm_> skipm, dreimark: simmilar to the current approach
2007-06-05T23:18:50 <neagulm_> skipm, yes
2007-06-05T23:19:01 <skipm> ciao for now...
2007-06-05T23:20:31 <neagulm_> dreimark, I will have to quit IRC to... I will be back early in the morning
2007-06-05T23:20:45 <dreimark> good night neagulm_
2007-06-05T23:21:04 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2081:917bb59fdbb8 1.7-jabber-knowak/MoinMoin/events/JabberNotification.py:
2007-06-05T23:21:04 <CIA-27> moin: Catch Exception as well, because xmlrpclib lets other exceptions propagate.
2007-06-05T23:21:04 <CIA-27> moin: Also a style fix.
2007-06-05T23:21:06 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2082:9ee68c8b409b 1.7-jabber-knowak/MoinMoin/ (events/__init__.py userform.py): Style fixes as suggested by Thomas.
2007-06-05T23:21:07 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2083:258d6e648eca 1.7-jabber-knowak/MoinMoin/ (events/JabberNotification.py events/__init__.py userform.py): Add a JabberIDUnsetEvent; make some private symbols internal in MoinMoin.events
2007-06-05T23:21:11 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2084:b37dd51e566e 1.7-jabber-knowak/MoinMoin/jabber/xmlrpcbot.py: Use decorators to assure that shared secret is always checked. Add 2 commands.
2007-06-05T23:21:14 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2085:9cebfffe27af 1.7-jabber-knowak/MoinMoin/events/JabberNotification.py: Don't forget to pass the value returned by event handlers back to the caller.
2007-06-05T23:21:17 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2086:da36e7cc80a3 1.7-jabber-knowak/MoinMoin/events/JabberNotification.py: Handle JID set/unset events on the wiki side.
2007-06-05T23:21:20 <CIA-27> moin: Karol 'grzywacz' Nowak <grzywacz@sul.uni.lodz.pl> * 2087:ab5e7026413a 1.7-jabber-knowak/MoinMoin/jabber/ (xmlrpcbot.py xmppbot.py commands.py): Move shared commands to a separate file, to make imports cleaner.
2007-06-05T23:21:33 <neagulm_> good night
2007-06-05T23:22:13 <dreimark> rhoe: I realized a few weeks ago that I have to refactor all of my 3rd party plugins which aren't yet included to MM too
2007-06-05T23:22:43 <dreimark> So you are not alone with that kind of job
2007-06-05T23:25:02 <rhoe> yeah, that sucks.. but maybe it is not that much work as i expect it
2007-06-05T23:26:50 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2045:b3ddb4504ec6 1.7/MoinMoin/parser/text_moin_wiki.py: text_moin_wiki: bug fix of {{{ aaa {{{\n }}} bbb }}}
2007-06-05T23:26:52 <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2046:1bdbf54a33ff 1.7/MoinMoin/parser/text_moin_wiki.py: text_moin_wiki:bug fix of NoURLsAfterBlock
2007-06-05T23:29:25 <grzywacz> O_o;
2007-06-05T23:30:03 <dreimark> grzywacz: ?
2007-06-05T23:30:33 <grzywacz> I'm waiting for Thomas to point out all my mistakes. :D
2007-06-05T23:33:21 <xorAxAx> dreimark: do we have unittests for that {{{ {{{ issues? it looks like you could have catched a lot bugs by trivil tests ...
2007-06-05T23:33:41 <xorAxAx> 22:57:14 -!- Irssi: New peak in #moin-dev@freenode : 23
2007-06-05T23:33:55 <xorAxAx> mysterious figure
2007-06-05T23:34:18 <grzywacz> What's mysterious about it?
2007-06-05T23:35:53 <xorAxAx> http://en.wikipedia.org/wiki/23_(number)
2007-06-05T23:36:04 <dreimark> xorAxAx: johill has found that bug I have fixed it and there will be added unittests too.
2007-06-05T23:36:07 <xorAxAx> http://en.wikipedia.org/wiki/23_(numerology)
2007-06-05T23:36:18 <xorAxAx> dreimark: it makes sense to write the test first
2007-06-05T23:36:27 <xorAxAx> afterwards, its only half as good
2007-06-05T23:36:35 <xorAxAx> and its not like this is not easily testable
2007-06-05T23:40:03 <dreimark> It's not easy to write a test if you have no idea why it is wrong in general.
2007-06-05T23:42:03 <dreimark> to make that clear I know some more issues related to that which may have easy workarounds
2007-06-05T23:42:37 <dreimark> but before fixing them I try to add some unittests about
2007-06-05T23:44:28 <xorAxAx> dreimark: well, in the commit message, you supplied an erronous scenario
2007-06-05T23:44:50 <xorAxAx> that would be a candiate for a test
2007-06-05T23:45:13 <xorAxAx> and you know this scenario before starting to code, so you can write the test beforehand
2007-06-05T23:46:26 <dreimark> true, and my only argument is the complexity of the parser
2007-06-05T23:46:52 <xorAxAx> the test would be a 3 liner :)
2007-06-05T23:47:03 <xorAxAx> so thats not the problem
2007-06-05T23:47:22 <xorAxAx> there are things that are much harder to test in moin but something like that is trivial
2007-06-05T23:47:42 <dreimark> the parser test itselfs is removed currently or isn't it
2007-06-05T23:48:28 <xorAxAx> no
2007-06-05T23:48:37 <xorAxAx> some tests are skipped because of line number issues
2007-06-05T23:49:14 <dreimark> ohh, then apologize, I have missed that
2007-06-05T23:58:01 <xorAxAx> dreimark: i cleaned it up in 1.7
2007-06-05T23:58:26 <xorAxAx> have a look at the tests that lanius wrote if you want to see nice examples :)
2007-06-05T23:58:43 <dreimark> :)
2007-06-05T23:59:18 <dreimark> there are several outstanding patches for 1.7 how to proceed with them?
2007-06-05T23:59:34 <xorAxAx> well, we need to keep some kind of list
2007-06-05T23:59:42 <xorAxAx> otherwise it will get mixed up
2007-06-05T23:59:46 <xorAxAx> (list of hashes
MoinMoin: MoinMoinChat/Logs/moin-dev/2007-06-05 (last edited 2007-10-29 19:10:51 by localhost)