1 2010-01-17T00:00:04 <CIA-30> Radomir Dopieralski <moindev@sheep.art.pl> default * 5441:5bcd98504363 1.9/wikiconfig.py: revert wikiconfig.py
2 2010-01-17T00:03:46 <dreimark> do you have run py.test?
3 2010-01-17T00:03:54 <TheSheep> I did
4 2010-01-17T00:04:18 <TheSheep> creole 1.1 is backward compatible with 1.0
5 2010-01-17T00:04:37 <TheSheep> I'm now modifying the moin parser to take advantage of new features though
6 2010-01-17T00:04:51 <dreimark> what is new?
7 2010-01-17T00:05:27 <TheSheep> you can pass a Rules object to the parser -- you can set some options in it or override some of the rules
8 2010-01-17T00:06:00 <TheSheep> functionally, it now takes a list of url schemas for free link detection, and can enable WikiWords
9 2010-01-17T00:06:36 <TheSheep> I wonder if WikiWords should be enabled in moin by default
10 2010-01-17T00:08:08 <dreimark> there was some discussion to start in 2.0 with no camelcase
11 2010-01-17T00:08:20 <dreimark> and to change all help pages
12 2010-01-17T00:09:36 <TheSheep> well, this is for 1.9
13 2010-01-17T00:10:27 <CIA-30> Radomir Dopieralski <moindev@sheep.art.pl> default * 5442:9dcd78d8be5f 1.9/MoinMoin/parser/text_creole.py: creole parser: respect url_schemas, enable CamelCase links
14 2010-01-17T00:39:55 *** eSyr
15 2010-01-17T00:57:59 <dreimark> failing test in i18n in 1.9
16 2010-01-17T00:58:09 <dreimark> MoinMoin/i18n/_tests/test_i18n.py:66: AssertionError
17 2010-01-17T00:59:19 <dreimark> ok that is a local test not in 1.9 yet
18 2010-01-17T02:35:34 *** grzywacz
19 2010-01-17T07:10:40 *** sharadchhetri
20 2010-01-17T08:38:24 *** sharadchhetri
21 2010-01-17T10:07:29 *** kursor
22 2010-01-17T11:17:02 *** kursor
23 2010-01-17T11:58:24 <ThomasWaldmann> moin
24 2010-01-17T12:18:06 <dreimark> moin
25 2010-01-17T12:27:25 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5443:8588b6292d9f 1.9/MoinMoin/i18n/ (41 files): updated i18n
26 2010-01-17T12:29:09 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5444:48874d70abff 1.9/MoinMoin/parser/text_moin_wiki.py: accept rel attr for links
27 2010-01-17T14:08:21 *** grzywacz
28 2010-01-17T14:13:52 <CIA-30> Reimar Bauer <rb.proj AT googlemail DOT com> default * 471:992cfded6a60 extensions/htdocs/svg-edit/ (77 files in 10 dirs): updated to version of svg-edit Revision r1200
29 2010-01-17T14:13:54 <CIA-30> Reimar Bauer <rb.proj AT googlemail DOT com> default * 472:8c4c99da678d extensions/ (5 files in 4 dirs): SvgEditor: svg-edit adapted to moin-1.9.0
30 2010-01-17T15:28:58 *** kursor
31 2010-01-17T16:30:43 <ThomasWaldmann> TheSheep: is the creole stuff tested enough for a release?
32 2010-01-17T16:31:30 <CIA-30> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5445:2b641be7728c 1.9/MoinMoin/ (6 files in 6 dirs):
33 2010-01-17T16:31:33 <CIA-30> action/diff.py: Added navigation links to change revision in left/right pane separately.
34 2010-01-17T16:31:35 <CIA-30> NOTE: util/diff_html.py output has been changed: now in case of identical texts
35 2010-01-17T16:31:37 <TheSheep> ThomasWaldmann: I did a release of the library yesterday
36 2010-01-17T16:31:37 <CIA-30> passed it outputs message string 'No differences found!' inside table in td with
37 2010-01-17T16:31:39 <CIA-30> class "diff-same" (previously it outputs only this message string without any
38 2010-01-17T16:31:41 <CIA-30> table at all).
39 2010-01-17T16:31:59 <TheSheep> ThomasWaldmann: the moin-specific stuff might need more of my attention though, I will look at it now
40 2010-01-17T16:32:56 * ThomasWaldmann wants to release 1.9.1 later today
41 2010-01-17T16:33:08 <TheSheep> ThomasWaldmann: give me half an hour
42 2010-01-17T16:33:55 <ThomasWaldmann> ok :)
43 2010-01-17T16:34:07 * ThomasWaldmann runs update-po again
44 2010-01-17T16:47:02 * ThomasWaldmann updated MoinI18n/de
45 2010-01-17T16:50:17 <TheSheep> hmm... a funny thing
46 2010-01-17T16:50:31 <TheSheep> I started wikiserver.py, but I'm getting ImportError: cannot import name Config
47 2010-01-17T16:50:56 <ThomasWaldmann> something wrong in wikiconfig or ..._local?
48 2010-01-17T16:51:14 <TheSheep> I pulled them from the repo
49 2010-01-17T16:51:22 <TheSheep> hg diff shows nothing
50 2010-01-17T16:52:09 <ThomasWaldmann> works for me
51 2010-01-17T16:52:27 * TheSheep scrathces his head
52 2010-01-17T16:53:13 * TheSheep pulls a fresh copy
53 2010-01-17T16:53:46 <TheSheep> this will take a little longer, sorry
54 2010-01-17T16:54:10 <ThomasWaldmann> you could try a local clone also
55 2010-01-17T16:55:04 <TheSheep> good idea
56 2010-01-17T16:55:29 <TheSheep> same
57 2010-01-17T16:57:54 <TheSheep> creating a wikiconfig_local.py helped
58 2010-01-17T16:59:46 <TheSheep> looks like the exact error message changed
59 2010-01-17T16:59:54 <TheSheep> and the code at the end of wikiconfig.py fails
60 2010-01-17T17:03:12 <ThomasWaldmann> TheSheep: it works for me, w/o wikiconfig_local
61 2010-01-17T17:03:50 <ThomasWaldmann> py 2.6.4 here
62 2010-01-17T17:09:54 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5446:709089f30164 1.9/MoinMoin/script/maint/cleansessions.py: moin maint cleansessions: catch KeyErrors for 'expires' and 'user.id' and make the best out of it
63 2010-01-17T17:12:03 <dreimark> TheSheep: do you may be have some local scripts not under rev control?
64 2010-01-17T17:15:15 * ThomasWaldmann wished all translators were so fast as de, pt-br and ru :)
65 2010-01-17T17:16:12 <dreimark> what do we do with http://moinmo.in/MoinMoinBugs/MoinSavesEmptySessions ?
66 2010-01-17T17:18:08 <ThomasWaldmann> well, the sessions tracks modifications and get saved if they are modified
67 2010-01-17T17:18:46 <ThomasWaldmann> so i think we either live with that effect or do not modify them
68 2010-01-17T17:23:34 * TheSheep is happy with the creole parser
69 2010-01-17T17:23:43 <ThomasWaldmann> :)
70 2010-01-17T17:24:01 <ThomasWaldmann> TheSheep: seen the shiny new diff action navigation?
71 2010-01-17T17:24:10 <TheSheep> I'm adding line counting to in in the next revision
72 2010-01-17T17:24:14 <TheSheep> ThomasWaldmann: no
73 2010-01-17T17:24:16 <johill> ThomasWaldmann: the trail code was explicitly written to not store a trail on the very first hit
74 2010-01-17T17:24:31 <johill> ThomasWaldmann: so that browsers that would never send a cookie wouldn't get a session file
75 2010-01-17T17:24:40 <TheSheep> revert, nice
76 2010-01-17T17:25:28 <johill> storing a session is only useful if we expect the browser to send the cookie again (openid for instance), or we _know_ it will
77 2010-01-17T17:25:38 <ThomasWaldmann> but how do you know it is the second hit if you don't store state?
78 2010-01-17T17:25:42 <johill> but if wget hits the site, there's no reason to store a session file
79 2010-01-17T17:25:49 <johill> because on the second hit the browser sent the cookie
80 2010-01-17T17:25:52 <TheSheep> ThomasWaldmann: 'revert to this version' means reverting to the version after or before the diff?
81 2010-01-17T17:25:54 <johill> and it didn't on the first
82 2010-01-17T17:26:09 <TheSheep> ThomasWaldmann: I think it's not obvious
83 2010-01-17T17:26:41 <ThomasWaldmann> TheSheep: yes, that is a known problem, the (new) version rendered below the diff display is meant
84 2010-01-17T17:27:39 <ThomasWaldmann> esyr and me talked about changing nav even more
85 2010-01-17T17:27:54 <ThomasWaldmann> like moving those 3 buttons to the table headline
86 2010-01-17T17:28:24 <ThomasWaldmann> and moving the dates to the metadata section
87 2010-01-17T17:28:37 <ThomasWaldmann> but that requires new i18n strings at a very visible place
88 2010-01-17T17:30:07 <TheSheep> I see
89 2010-01-17T17:34:17 <ThomasWaldmann> johill: where is the place where "the trail code was explicitly written to not store a trail on the very first hit"?
90 2010-01-17T17:40:00 <johill> maybe that's not actually true, but that's how it worked
91 2010-01-17T17:40:14 <johill> it may well be that my old code ignored the trail when determining whether to save a session or not
92 2010-01-17T17:40:34 <johill> however, it does seem to me that even with anon sessions I don't get anything in the session file but 'expires' when using wget
93 2010-01-17T17:42:37 <dreimark> the patch works for me
94 2010-01-17T17:56:56 <johill> is should_save always set correctly?
95 2010-01-17T17:57:22 <johill> it might be better to just check if len(session) > 1?
96 2010-01-17T17:57:28 <johill> i.e. anything but 'expires' in it
97 2010-01-17T18:07:50 * ThomasWaldmann tries checking session.new
98 2010-01-17T18:08:37 <johill> I forgot all the details :(
99 2010-01-17T18:11:01 <johill> heh
100 2010-01-17T18:11:03 <johill> .should_save
101 2010-01-17T18:11:07 <johill> return self.modified or self.new
102 2010-01-17T18:11:18 <johill> so that's explicit in wekrzeug
103 2010-01-17T18:12:09 <dreimark> looks like i have too much http_auth session files
104 2010-01-17T18:12:20 <johill> hm?
105 2010-01-17T18:12:39 <johill> anyway I don't remember exactly how I did it, but I'm pretty sure 1.8 didn't store like that
106 2010-01-17T18:12:43 <johill> :)
107 2010-01-17T18:13:17 <johill> and apparently whoever rewrote the code didn't understand the old code, which I guess I can blame on myself for not commenting it well enough
108 2010-01-17T18:39:30 * ThomasWaldmann has found a sid problem
109 2010-01-17T18:44:37 <ThomasWaldmann> johill: werzeug.contrib.sessions:219
110 2010-01-17T18:44:39 <ThomasWaldmann> if not self.is_valid_key(sid) or not path.exists(fn):
111 2010-01-17T18:44:39 <ThomasWaldmann> return self.new()
112 2010-01-17T18:44:54 <ThomasWaldmann> thus I think it won't work like we want
113 2010-01-17T18:45:09 <ThomasWaldmann> if there is no session stored for some sid, it'll create a new sid
114 2010-01-17T19:09:38 <johill> uh
115 2010-01-17T19:09:46 <johill> even if the browser sent a cookie? that's kinda dumb
116 2010-01-17T19:12:23 <ThomasWaldmann> the session code just works with sids, not cookies
117 2010-01-17T19:12:31 <ThomasWaldmann> but I fixed .get() behaviour
118 2010-01-17T19:12:34 <johill> right but the sid is in the cookie
119 2010-01-17T19:12:58 <ThomasWaldmann> i am just polishing it a bit and then post a diff for review
120 2010-01-17T19:13:00 <johill> so my old code would just take the sid from the cookie even if there was no information stored for that yet
121 2010-01-17T19:13:16 <johill> which would then also indicate "hey, this browser knows how to send cookies"
122 2010-01-17T19:13:20 <ThomasWaldmann> yes, that's like it behaves now
123 2010-01-17T19:13:50 <johill> cool
124 2010-01-17T19:28:31 <ThomasWaldmann> http://pastebin.com/f6f78f2c0 < please review
125 2010-01-17T19:44:28 * johill looks
126 2010-01-17T19:45:31 <ThomasWaldmann> there is one problem now: if you have no cookie and POST the login form then, it won't work
127 2010-01-17T19:50:06 <johill> if you don't allow cookies you can't login anyway
128 2010-01-17T19:52:25 <johill> guess it looks ok to me
129 2010-01-17T19:54:22 <ThomasWaldmann> login doesn't work if cookies are allowed
130 2010-01-17T19:54:37 <ThomasWaldmann> because that session is also "new"
131 2010-01-17T19:55:50 <johill> but it should no longer be new after the get
132 2010-01-17T19:55:57 <johill> I mean, the login form GET is new
133 2010-01-17T19:56:02 <johill> and then the POST no longer
134 2010-01-17T19:57:40 <ThomasWaldmann> 1. request, no cookie, POSTs login data -> does not store session
135 2010-01-17T19:57:59 <ThomasWaldmann> 2. request, GET -> has no user.id because there is no session stored
136 2010-01-17T19:58:04 <johill> well 1. request should obviously set the cookie
137 2010-01-17T19:58:11 <johill> but _not_ save the session file
138 2010-01-17T19:58:17 <ThomasWaldmann> then you have a cookie but no session file
139 2010-01-17T19:58:21 <johill> yeah
140 2010-01-17T19:58:28 <ThomasWaldmann> and the userid is in session file
141 2010-01-17T19:58:38 <johill> it's like an anonymous session then
142 2010-01-17T19:58:52 <johill> when it has no stored data
143 2010-01-17T20:00:14 <ThomasWaldmann> no stored data, no user id, no logging in
144 2010-01-17T20:00:27 <johill> I don't really understand
145 2010-01-17T20:00:36 <johill> if the first request sets the cookie, but doesn't store data
146 2010-01-17T20:00:44 <ThomasWaldmann> wait for next patch :)
147 2010-01-17T20:00:46 <johill> then the POST request will save the user id and store the session
148 2010-01-17T20:01:10 <ThomasWaldmann> if you have not enabled anon session, a anon request won't set a cookie
149 2010-01-17T20:01:24 <johill> ah, but it still could
150 2010-01-17T20:01:29 <johill> it just wouldn't ever save it
151 2010-01-17T20:01:44 <johill> kinda a pointless exchange though
152 2010-01-17T20:01:49 <johill> but yeah
153 2010-01-17T20:01:53 <johill> my original solution to this was different
154 2010-01-17T20:02:01 <ThomasWaldmann> i try if i can fix it otherwise
155 2010-01-17T20:02:06 <johill> it would store the session, even if it was new, if it contained important data
156 2010-01-17T20:02:22 <johill> so the POST would create and store it
157 2010-01-17T20:05:00 <ThomasWaldmann> http://pastebin.com/f4a34184a
158 2010-01-17T20:07:01 <johill> guess that check works
159 2010-01-17T20:08:09 <ThomasWaldmann> johill: do you think the werkzeug stuff is a bug or intended behaviour?
160 2010-01-17T20:14:15 <johill> good question
161 2010-01-17T20:14:30 <johill> few people seem to ever think about such details ;)
162 2010-01-17T20:30:05 *** kursor
163 2010-01-17T20:35:08 *** grzywacz
164 2010-01-17T20:46:36 <dreimark> ThomasWaldmann: that patch does not work with the recent editmoin
165 2010-01-17T20:47:09 <ThomasWaldmann> which?
166 2010-01-17T20:48:02 <dreimark> http://paste.pocoo.org/show/166421/
167 2010-01-17T20:48:34 *** eSyr
168 2010-01-17T20:49:06 <dreimark> http://moinmo.in/ReimarBauer?action=AttachFile&do=view&target=editmoin
169 2010-01-17T20:50:00 <ThomasWaldmann> dreimark: http://pastebin.com/f4a34184a use that
170 2010-01-17T20:52:40 <dreimark> :)
171 2010-01-17T20:55:24 * dreimark missed the second one
172 2010-01-17T20:55:49 <dreimark> may be we can add some tests for that based on the editmoin aproach
173 2010-01-17T21:03:23 <ThomasWaldmann> i guess i have to subclass werkzeug stuff to fix that
174 2010-01-17T21:03:51 <ThomasWaldmann> because even if mitsuhiko fixes that now, it maybe won't get into some packages anytime soon
175 2010-01-17T21:46:57 <johill> ick
176 2010-01-17T21:47:04 <johill> can we have a new release soon with all the openid fixes/
177 2010-01-17T21:48:13 <dreimark> it was planned for today
178 2010-01-17T21:48:21 <johill> oh
179 2010-01-17T21:50:08 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5447:fed925dfdc0d 1.9/MoinMoin/web/session.py: (log message trimmed)
180 2010-01-17T21:50:11 <CIA-30> improve moin's session handling / fix werkzeug by subclassing (see below)
181 2010-01-17T21:50:13 <CIA-30> moin created lots of session files, esp. if you had anon sessions enabled,
182 2010-01-17T21:50:15 <CIA-30> but the user agent that did not support cookies (ab, wget, ...).
183 2010-01-17T21:50:17 <CIA-30> now we avoid creating session files until we KNOW that user agent supports
184 2010-01-17T21:50:19 <CIA-30> cookies (because it sent us one).
185 2010-01-17T21:50:21 <CIA-30> sessions with a logged-in (valid) user are still a bit problematic, because we
186 2010-01-17T21:51:37 <dreimark> i have headache all the day so it may be me who has problems with http_auth now
187 2010-01-17T21:53:27 <ThomasWaldmann> maybe that changeset also fixes http_auth, try it
188 2010-01-17T22:01:55 <dreimark> GivenAuth works but HTTPAuthMoin not
189 2010-01-17T22:01:57 <dreimark> [Sun Jan 17 22:01:38 2010] [error] [client 127.0.0.1] Digest: client used wrong authentication scheme `Basic': /mywiki_wsgi/Aktuelle\xc3\x84nderungen, referer: http://localhost/mywiki_wsgi/RecentChanges
190 2010-01-17T22:04:52 <ThomasWaldmann> basic vs digest?
191 2010-01-17T22:09:17 <dreimark> the failure is misleading but it is my fault
192 2010-01-17T22:09:38 <dreimark> there was a setting left over
193 2010-01-17T22:15:56 <dreimark> ThomasWaldmann: http://paste.pocoo.org/show/166447/
194 2010-01-17T22:18:32 <dreimark> ThomasWaldmann: cleansessions fails HTTPAuthMoin
195 2010-01-17T22:29:49 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5448:5518d41fc686 1.9/MoinMoin/web/session.py: move SID listing functionality to FixedFilesystemSessionStore
196 2010-01-17T22:31:46 * ThomasWaldmann looks at it
197 2010-01-17T22:36:17 <ThomasWaldmann> well, it is rather clear that this will happen, but currently no good idea about how to fix that best
198 2010-01-17T22:37:09 <ThomasWaldmann> if the HTTPAuthMoin doesn't get auth headers, it tries requesting them by aborting and responding with some appropriate http header
199 2010-01-17T22:39:29 <ThomasWaldmann> dreimark: can you try adding if not isinstance(request, ScriptContext): ?
200 2010-01-17T22:41:05 <dreimark> moin command at all does not work with HTTPAuthMoin
201 2010-01-17T22:41:50 <ThomasWaldmann> line 86
202 2010-01-17T22:42:16 <ThomasWaldmann> sure, because command line is not http X)
203 2010-01-17T22:42:24 <dreimark> yes
204 2010-01-17T22:44:35 <CIA-30> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5449:8946d234bc25 1.9/docs/CHANGES: docs/CHANGES: Added information about 2b641be7728c.
205 2010-01-17T22:44:37 <CIA-30> Eugene Syromyatnikov <evgsyr@gmail.com> default * 5450:7b0e8d07b146 1.9/MoinMoin/ (script/maint/cleansessions.py web/session.py): Merge
206 2010-01-17T22:54:17 * ThomasWaldmann runs update-po
207 2010-01-17T22:58:16 <dreimark> i wonder because with GivenAuth it works
208 2010-01-17T23:02:10 <dreimark> at least cleaning cache
209 2010-01-17T23:20:08 <dreimark> eSyr: nice diff :)
210 2010-01-17T23:33:35 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5451:12c5e510548f 1.9/MoinMoin/i18n/ (43 files): updated i18n
211 2010-01-17T23:44:24 <CIA-30> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5452:cc8f05e87bbc 1.9/wiki/underlay.tar: updated underlay
212 2010-01-17T23:53:19 <eSyr> dreimark, thanks.