1 2010-01-15T00:00:14 * eSyr is thinking how to change test without missing anything it tests
2 2010-01-15T00:01:29 <eSyr> oh, there already is import re >:3
3 2010-01-15T00:24:00 <eSyr> ThomasWaldmann, dreimark, diff for tests: http://paste.pocoo.org/show/165557/
4 2010-01-15T00:29:19 <ThomasWaldmann> http://moinmo.in/4ct10n/diff/MoinMoinBugs/LineNumberSpansForProcessInstructionsMissed please also add that one / test it
5 2010-01-15T00:30:10 * dreimark is writing mails currently can look at sources tom
6 2010-01-15T00:30:35 <ThomasWaldmann> eSyr: i am not sure I like re.search in the tests. does one see failure reason if it fails?
7 2010-01-15T00:30:47 <ThomasWaldmann> and is the "is not None" needed?
8 2010-01-15T00:31:40 <eSyr> ThomasWaldmann, i'm not good at asserts, so i est that matchobject is not None explicitly.
9 2010-01-15T00:32:08 <eSyr> ThomasWaldmann, now <span ...></span> is added to hml ouput, so test begin failing.
10 2010-01-15T00:32:08 <ThomasWaldmann> assert expects a truth value
11 2010-01-15T00:32:35 <ThomasWaldmann> re.search result will be either None (falsy) or a match object (truly)
12 2010-01-15T00:33:10 <eSyr> ok.
13 2010-01-15T00:33:13 <johill> dreimark: I thought autofilter worked on more than one
14 2010-01-15T00:33:22 <johill> dreimark: maybe I messed it up ... :(
15 2010-01-15T00:33:43 <johill> oh another thing ... I thought I had session cleanup code in the original code
16 2010-01-15T00:33:52 <johill> is that broken, or did that get lost?
17 2010-01-15T00:34:00 <johill> I had ~200k sessions yesterday for wireless.kernel.org
18 2010-01-15T00:35:24 <ThomasWaldmann> eSyr: 9 + for lineno in range(1, self.start_line): <- check off-by-one error
19 2010-01-15T00:36:04 <ThomasWaldmann> johill: it kills sessions now when they are expired
20 2010-01-15T00:36:21 <ThomasWaldmann> (you need a cron job with moin maint cleansessions)
21 2010-01-15T00:36:28 <johill> but I thought I had that in the code before
22 2010-01-15T00:36:44 <ThomasWaldmann> we have new session code, from werkzeug
23 2010-01-15T00:37:14 <johill> hm so it got lost
24 2010-01-15T00:37:16 <johill> bugger
25 2010-01-15T00:37:29 <johill> while at it, are there any other cronjobs I should set up?
26 2010-01-15T00:37:40 <ThomasWaldmann> for xapian, if you use it
27 2010-01-15T00:38:01 <johill> not right now
28 2010-01-15T00:38:47 <johill> requiring a cron job kinda sucks tho
29 2010-01-15T00:38:53 <johill> for web-hosting environments
30 2010-01-15T00:38:56 <johill> I don't care :)
31 2010-01-15T00:40:14 <ThomasWaldmann> do you remember where the old cleanup code was?
32 2010-01-15T00:42:41 <johill> it was like random, every 100 requests look at them
33 2010-01-15T00:43:29 <ThomasWaldmann> looking at all sessions can take long
34 2010-01-15T00:48:50 <ThomasWaldmann> found the old cleanup code
35 2010-01-15T00:56:43 <ThomasWaldmann> johill: do you think one could spread the purge more evenly?
36 2010-01-15T00:57:26 <johill> not sure
37 2010-01-15T00:57:40 <ThomasWaldmann> i mean instead of processing all sessions every 1000 requests, process 30 every 30 requests?
38 2010-01-15T00:58:01 <ThomasWaldmann> ehrm 1/30 of the sessions
39 2010-01-15T00:58:07 <johill> not easily
40 2010-01-15T01:00:10 <ThomasWaldmann> maybe a random choice of some specific size of all sessions
41 2010-01-15T01:00:33 <ThomasWaldmann> sounds like a task for a mathematician :D
42 2010-01-15T01:00:53 <johill> well you still have to list them all
43 2010-01-15T01:00:58 <johill> which can be fairly expensive already
44 2010-01-15T01:01:20 <ThomasWaldmann> maybe, maybe not :)
45 2010-01-15T01:01:56 <ThomasWaldmann> problem is if you have 1000 users, it would process 1000 sessions files now and then
46 2010-01-15T01:03:01 <ThomasWaldmann> we need auto-destructing files :D
47 2010-01-15T01:03:17 <johill> heh
48 2010-01-15T01:04:32 <eSyr> hm.
49 2010-01-15T01:05:04 <eSyr> i hink, only index with all sessions and last access time.
50 2010-01-15T01:05:12 <eSyr> and check it accidenly.
51 2010-01-15T01:05:30 <eSyr> (checking one dict/file wih dict is easier than checking bunch of files)
52 2010-01-15T01:07:24 <ThomasWaldmann> hm?
53 2010-01-15T01:07:28 <eSyr> hm, i think, generating of line number anchors should be refactored somehow...
54 2010-01-15T01:08:45 <eSyr> ThomasWaldmann, i propose storing session index (like user index)
55 2010-01-15T01:10:09 <johill> hm
56 2010-01-15T01:10:15 <johill> not working
57 2010-01-15T01:10:28 <johill> moin.py --config-dir=... maint ..
58 2010-01-15T01:10:36 <johill> MoinMoin.config.multiconfig:93 using farm config: /home/www/moin/farmconfig.pyc
59 2010-01-15T01:10:49 <johill> werkzeug.exceptions.NotFound: 404 Not Found
60 2010-01-15T01:12:12 <johill> from raise NotFound('<p>No wiki configuration matching the URL found!</p>')
61 2010-01-15T01:12:59 <johill> do I really need to give the url, can\t it do all at once?
62 2010-01-15T01:13:37 <eSyr> johill, when you use farm config, selecing of config file is based on wiki host name.
63 2010-01-15T01:13:44 <johill> well yeah
64 2010-01-15T01:13:57 <johill> but I wanted it to clean up _all_ of them
65 2010-01-15T01:14:04 <eSyr> oh, well.
66 2010-01-15T01:14:05 <johill> I sure don't want to make a list myself
67 2010-01-15T01:14:09 <eSyr> it's the problem.
68 2010-01-15T01:14:46 <ThomasWaldmann> johill: --wiki-url=http://...
69 2010-01-15T01:15:29 <johill> ^^
70 2010-01-15T01:15:39 <johill> I want --all
71 2010-01-15T01:15:47 <johill> that's why I have a farm, after all ;)
72 2010-01-15T01:17:15 <johill> I guess the cronjob idea isn't really viable right now ;)
73 2010-01-15T01:17:49 <johill> at this point, sleeping seems more productive, at least in medium term
74 2010-01-15T01:19:09 <ThomasWaldmann> hehe
75 2010-01-15T01:20:00 <johill> should I open a feature request?
76 2010-01-15T01:20:18 <johill> (the script framework doesn't seem clear enough to me that I could just whip it up)
77 2010-01-15T01:20:56 <ThomasWaldmann> yes, please
78 2010-01-15T01:23:30 <johill> done, bed time now
79 2010-01-15T01:37:08 <ThomasWaldmann> gn johill
80 2010-01-15T01:42:31 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5427:cd2696f05fc5 1.9/MoinMoin/parser/text_moin_wiki.py: parser/text_moin_wiki.py: Fix of http://www.moinmo.in/MoinMoinBugs/LineNumberSpansForProcessInstructionsMissed for text_moin_wiki parser
81 2010-01-15T01:42:31 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5428:82b60084aaf3 1.9/MoinMoin/parser/text.py: Fix for MoinMoinBugs/LineNumberAnchorsInPreformattedText (plain ext parser)
82 2010-01-15T01:42:33 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5429:4d3f51792160 1.9/MoinMoin/parser/highlight.py: Fix for MoinMoinBugs/LineNumberAnchorsInPreformattedText (highlight parser)
83 2010-01-15T01:42:36 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5430:c18b36185649 1.9/MoinMoin/parser/_tests/test_text_moin_wiki.py: Tests updated to conform with line number anchors in plain text parser blocks.
84 2010-01-15T01:42:37 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5431:5886d3b46063 1.9/MoinMoin/parser/highlight.py: Fix of MoinMoinBugs/LineNumberSpansForProcessInstructionsMissed for highlight parser.
85 2010-01-15T01:42:39 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5432:172d77f67e0a 1.9/MoinMoin/parser/text.py: Fix of MoinMoinBugs/LineNumberSpansForProcessInstructionsMissed for plain text parser.
86 2010-01-15T01:44:58 <eSyr> ThomasWaldmann, and what do you think about http://moinmo.in/FeatureRequests/LeaveRedirectInPlaceWhenRenamingPages ?
87 2010-01-15T03:48:52 *** eSyr
88 2010-01-15T03:51:26 *** grzywacz
89 2010-01-15T06:22:36 *** xjjk
90 2010-01-15T06:25:28 *** grzywacz
91 2010-01-15T07:00:38 *** grzywacz
92 2010-01-15T07:31:07 *** grzywa
93 2010-01-15T07:46:24 *** grzywacz
94 2010-01-15T08:22:29 *** grzywa
95 2010-01-15T09:08:43 *** moinBot
96 2010-01-15T09:10:23 *** waldi
97 2010-01-15T09:10:23 *** elmo
98 2010-01-15T09:10:23 *** ThomasWaldmann
99 2010-01-15T09:10:23 *** xorAxAx
100 2010-01-15T09:10:23 *** ChanServ
101 2010-01-15T09:10:23 *** moinBot`
102 2010-01-15T09:10:23 *** dreimark
103 2010-01-15T09:10:23 *** ser
104 2010-01-15T09:10:23 *** ronny
105 2010-01-15T09:12:43 *** ChanServ
106 2010-01-15T09:12:43 *** elmo
107 2010-01-15T09:12:43 *** waldi
108 2010-01-15T09:12:43 *** ThomasWaldmann
109 2010-01-15T09:12:43 *** dreimark
110 2010-01-15T09:12:43 *** ser
111 2010-01-15T09:12:43 *** xorAxAx
112 2010-01-15T09:12:43 *** ronny
113 2010-01-15T09:12:43 *** irc.freenode.net
114 2010-01-15T11:24:01 *** eSyr
115 2010-01-15T12:18:29 <ThomasWaldmann> moin
116 2010-01-15T12:33:45 *** ChanServ
117 2010-01-15T12:33:51 *** ChanServ
118 2010-01-15T12:33:51 *** irc.freenode.net
119 2010-01-15T12:36:37 <dreimark> moin
120 2010-01-15T12:42:07 *** ChanServ
121 2010-01-15T12:42:43 *** ChanServ
122 2010-01-15T12:42:43 *** irc.freenode.net
123 2010-01-15T13:30:09 *** eSyr
124 2010-01-15T14:35:50 *** kursor
125 2010-01-15T15:40:14 <eSyr> hm, isnt uid also always ascii?
126 2010-01-15T15:43:45 <eSyr> Then, http://paste.pocoo.org/show/165711/
127 2010-01-15T15:43:56 <dreimark> I wonder why it needs the abs path
128 2010-01-15T15:45:24 <dreimark> hmm, strange it has only a problem with the location of the sessions.py file ?
129 2010-01-15T15:46:17 <eSyr> session file, because sid got from cookie is unicode, not bytestring.
130 2010-01-15T15:46:27 <ThomasWaldmann> that line looks crappy anyway :)
131 2010-01-15T15:47:36 <dreimark> can you check if the problem is really the seesion sid file and not the path where it is located
132 2010-01-15T15:48:22 * dreimark has had in the past some problems because of wrong encoding of dir names
133 2010-01-15T15:48:24 <eSyr> dreimark, path is bytestring, sid filename is unicode, so at concainating python tries to convert path to unicode and fails
134 2010-01-15T15:48:51 <eSyr> because it uses ascii parser, and path can be in any encoding.
135 2010-01-15T15:49:26 <eSyr> more, because in this place filename is generating, the produced string should be also bytestring.
136 2010-01-15T15:50:25 <eSyr> so the right solution is to keep sid as bytestring (likely ascii, or using getfilesystemencoding).
137 2010-01-15T15:51:48 <dreimark> why is path bytestring?
138 2010-01-15T15:52:30 <eSyr> err, why it is not?
139 2010-01-15T15:52:37 <johill> it should always be
140 2010-01-15T15:52:51 <johill> you could even have a path with completely bogus encoding
141 2010-01-15T15:52:54 <johill> :)
142 2010-01-15T15:53:13 <eSyr> or no encoding at all.
143 2010-01-15T15:53:24 <johill> or just random byte garbage
144 2010-01-15T15:53:28 <eSyr> (but for what...)
145 2010-01-15T15:56:19 <dreimark> I think I don't understand something here. We set in our wiki the data path.
146 2010-01-15T15:57:27 <dreimark> and if I have nonascii chars i have to set it by u'мойн/data'
147 2010-01-15T15:57:35 <dreimark> or not?
148 2010-01-15T15:58:05 <eSyr> no, 'мойн/data
149 2010-01-15T15:58:07 <eSyr> '
150 2010-01-15T15:58:25 <eSyr> (and in this case config file encoding should be the same as filesystem encoding)
151 2010-01-15T15:58:38 <ThomasWaldmann> yeah X)
152 2010-01-15T15:58:56 <dreimark> ah ok
153 2010-01-15T15:59:19 <ThomasWaldmann> or data_dir = u"...".encode(fsencoding)
154 2010-01-15T16:00:10 <eSyr> sys.getfilesystemencoding()
155 2010-01-15T16:00:39 <dreimark> then concatinating would work for sid.encode(fsencoding)
156 2010-01-15T16:00:56 <eSyr> yes.
157 2010-01-15T16:01:48 <ThomasWaldmann> sid is ascii, you don't need encode(), str() works fine then
158 2010-01-15T16:02:03 <dreimark> and it is only needed for concantinating? I mean we don't need to change how sid is stored.
159 2010-01-15T16:02:59 <eSyr> well, sid should be stored in any encoding which support ascii character set :)
160 2010-01-15T16:03:16 <eSyr> (and convert it to ascii when loading from storage)
161 2010-01-15T16:03:44 <CIA-44> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5433:c800980890aa 1.9/MoinMoin/web/session.py: Fix for MoinMoinBugs/WerkzeugErrorWithNonAsciiPath ? making sid always ascii
162 2010-01-15T16:09:48 <eSyr> ThomasWaldmann, can i also push this patch: http://paste.pocoo.org/show/165711/ ?
163 2010-01-15T16:10:48 <ThomasWaldmann> you need to review what self.id can be
164 2010-01-15T16:11:09 <eSyr> ok.
165 2010-01-15T16:11:37 <ThomasWaldmann> also, your werkzeug ticket's patches will likely get rejected because of causing other problems
166 2010-01-15T16:11:55 <ThomasWaldmann> https://bugs.launchpad.net/bzr/+bug/41183 < see there for example
167 2010-01-15T16:13:37 <ThomasWaldmann> maybe the only reasonable thing one could do at werkzeug session level is something like either
168 2010-01-15T16:13:57 <ThomasWaldmann> assert isinstance(sid, str)
169 2010-01-15T16:13:59 <ThomasWaldmann> or
170 2010-01-15T16:14:07 <ThomasWaldmann> sid = str(sid)
171 2010-01-15T16:15:10 <ThomasWaldmann> (and add documentation about param typing and why it should be done like that)
172 2010-01-15T16:15:15 <eSyr> sys.getfilesystemencoding() or 'ascii' should also work.
173 2010-01-15T16:15:24 <eSyr> hm.
174 2010-01-15T16:16:00 <ThomasWaldmann> (there is not much reason to have non-ascii sids anyway)
175 2010-01-15T16:17:03 <ThomasWaldmann> the good thing with assert is that it would fail in any case where caller gives unicode
176 2010-01-15T16:17:26 <ThomasWaldmann> the str method would only fail for some cases, when in doesn't convert to ascii str
177 2010-01-15T16:17:35 <ThomasWaldmann> one can see that as pro or con :)
178 2010-01-15T16:18:05 <ThomasWaldmann> assert might be cleaner, but cause more trouble
179 2010-01-15T16:18:43 <eSyr> well, i'm voting for working in all cases and in most appropriate fashion, but werkzeug devs may have other opinion :)
180 2010-01-15T16:19:04 <ThomasWaldmann> btw, if you review User code anyway, please check that "...None..." thing, that looks dirty
181 2010-01-15T16:19:15 <eSyr> yes, of course.
182 2010-01-15T16:22:09 <eSyr> hehe, looks like -i've found a bug in some migration code.
183 2010-01-15T16:23:29 <eSyr> http://paste.pocoo.org/show/165721/
184 2010-01-15T16:24:34 <ThomasWaldmann> what's the bug?
185 2010-01-15T16:24:51 <eSyr> the second option should be uid.
186 2010-01-15T16:25:04 <ThomasWaldmann> be careful, that might be a different "User" class
187 2010-01-15T16:25:07 <eSyr> (looks like in past user_dir was passed, but now it is not)
188 2010-01-15T16:25:12 <eSyr> oh.
189 2010-01-15T16:25:16 <eSyr> uhoh.
190 2010-01-15T16:25:18 <ThomasWaldmann> all those mig scripts are rather lowlevel hacks
191 2010-01-15T16:25:37 <eSyr> hm, looks like you're right.
192 2010-01-15T16:25:43 * ThomasWaldmann looks forward to storage api
193 2010-01-15T17:21:46 <eSyr> ThomasWaldmann, http://paste.pocoo.org/show/165738/
194 2010-01-15T17:34:08 <dreimark> eSyr: http://moinmo.in/FeatureRequests/LeaveRedirectInPlaceWhenRenamingPages
195 2010-01-15T17:35:15 <dreimark> btw is our edit lock warning broken?
196 2010-01-15T17:37:07 <eSyr> ehm.
197 2010-01-15T17:37:18 <dreimark> becaue you should have seen that i was editing
198 2010-01-15T17:38:14 <eSyr> hm, i haven't got any edit lock warnings (just checked my browser history)
199 2010-01-15T17:39:25 <dreimark> lets try again. I am looking this time can you edit the FR now?
200 2010-01-15T17:39:45 <dreimark> and tell
201 2010-01-15T17:40:00 <eSyr> hm, got no warning...
202 2010-01-15T17:40:08 <dreimark> i edit now
203 2010-01-15T17:40:17 <dreimark> This page was opened for editing or last previewed at 2010-01-15 18:39:43 by EugeneSyromyatnikov.
204 2010-01-15T17:40:32 <dreimark> that is on top of the edit window
205 2010-01-15T17:40:49 <dreimark> just below Edit "Feature Requests/Leave Redirect In Place When Renaming Pages"
206 2010-01-15T17:41:20 <eSyr> yes, as it should, isn't it?
207 2010-01-15T17:42:00 <dreimark> well i was wondering why i got an edit conflict http://moinmo.in/4ct10n/info/FeatureRequests/LeaveRedirectInPlaceWhenRenamingPages?action=info
208 2010-01-15T17:42:20 <eSyr> yes, i've seen.
209 2010-01-15T17:43:50 <dreimark> because i have had no warning i started to edit and when i tried to save i got edit conflict.
210 2010-01-15T17:44:20 <dreimark> so i thought you haven't got the message i was editing
211 2010-01-15T17:44:41 <eSyr> yes, i've understood you.
212 2010-01-15T17:45:07 <eSyr> ThomasWaldmann, can you figure out by logs, who starts editing first?
213 2010-01-15T17:47:12 <dreimark> bbl
214 2010-01-15T18:13:01 *** grzywacz
215 2010-01-15T18:13:27 *** xjjk
216 2010-01-15T19:27:23 *** kursor
217 2010-01-15T19:37:42 *** eSyr
218 2010-01-15T23:41:29 *** moinBot`
219 2010-01-15T23:41:53 *** xjjk
220 2010-01-15T23:41:53 *** moinBot
221 2010-01-15T23:41:53 *** dennda
222 2010-01-15T23:41:53 *** vpv
223 2010-01-15T23:41:53 *** CIA-44
224 2010-01-15T23:42:44 *** xjjk
225 2010-01-15T23:42:59 *** dennda
226 2010-01-15T23:50:50 *** franklin
227 2010-01-15T23:50:50 *** TheSheep
228 2010-01-15T23:50:55 *** vpv
229 2010-01-15T23:51:05 *** franklin
230 2010-01-15T23:54:17 *** CIA-48
231 2010-01-15T23:54:38 *** TheSheep
232 2010-01-15T23:56:59 *** CIA-48
233