1 2010-02-23T00:14:32 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5574:8b9acb287705 1.9/MoinMoin/util/filesys.py: (log message trimmed)
2 2010-02-23T00:14:33 <CIA-47> filesys.rename: add debug logging and retries (see below)
3 2010-02-23T00:14:33 <CIA-47> Without the retries, this stuff does not work as expected:
4 2010-02-23T00:14:33 <CIA-47> If the win32 transacted "atomic" rename finds that there is already some other
5 2010-02-23T00:14:33 <CIA-47> transacted rename running for the target file, it returns ret=6800. In the
6 2010-02-23T00:14:33 <CIA-47> original code this meant falling back to another (pseudo-atomic) rename, which
7 2010-02-23T00:14:33 <CIA-47> then likely had the same problem, causing another fallback to the non-atomic
8 2010-02-23T00:34:34 *** grzywacz
9 2010-02-23T09:20:48 <dreimark> moin
10 2010-02-23T09:30:14 <ThomasWaldmann> moin
11 2010-02-23T10:17:02 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5575:17cf01154e12 1.9/MoinMoin/web/session.py: sessions: remove code not needed any more due to new werkzeug session code
12 2010-02-23T10:17:04 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5576:6b94d284ad11 1.9/MoinMoin/ (user.py web/session.py): avoid writing key/value to session dict if value doesn't change
13 2010-02-23T10:17:05 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5577:092ce221f03a 1.9/MoinMoin/web/session.py: session finalize: refactor, don't save cleared session on logout
14 2010-02-23T10:18:56 <ThomasWaldmann> ^^ this reduces session store update frequency significantly as long as trail doesn't change
15 2010-02-23T10:19:21 <ThomasWaldmann> e.g. if one loads a page that shows a bunch of attached images
16 2010-02-23T10:32:56 <dreimark> ThomasWaldmann: also if one reloads the same page ?
17 2010-02-23T10:33:49 <dreimark> what happens if you show a page by refresh and have a cookie lifetime
18 2010-02-23T10:34:07 <dreimark> is the user logged out because the session file gets no time stamp?
19 2010-02-23T10:34:11 <dreimark> +new
20 2010-02-23T10:34:50 <ThomasWaldmann> if session contents does not change, it does not save
21 2010-02-23T10:35:21 <dreimark> that is a problem
22 2010-02-23T10:35:31 <ThomasWaldmann> timestamp is update with 60s granularity
23 2010-02-23T10:35:36 <ThomasWaldmann> +d
24 2010-02-23T10:36:21 <dreimark> it should keep the login state for a user using refresh as pragma on a page
25 2010-02-23T10:36:30 <dreimark> otherwise this one becomes logged out
26 2010-02-23T10:36:44 <ThomasWaldmann> no
27 2010-02-23T10:36:58 <dreimark> then i misinterpreted it
28 2010-02-23T10:37:13 <ThomasWaldmann> try it
29 2010-02-23T10:37:20 <ThomasWaldmann> should work
30 2010-02-23T10:37:52 * dreimark will do in some hours
31 2010-02-23T10:38:29 <dreimark> after midday
32 2010-02-23T11:01:16 <ThomasWaldmann> current code now runs on moinmo.in
33 2010-02-23T11:27:17 <dreimark> can we have somehow justifiy text ?
34 2010-02-23T11:28:37 <dreimark> oh we can too easy
35 2010-02-23T15:50:32 *** kursor
36 2010-02-23T18:29:30 <CIA-47> Reimar Bauer <rb.proj AT googlemail DOT com> default * 505:944bccb3f184 extensions/data/plugin/macro/CollectLists.py: macro.CollectLists: better error message for no search result
37 2010-02-23T18:32:39 *** kursor
38 2010-02-23T18:33:15 *** kursor
39 2010-02-23T18:36:30 *** kursor
40 2010-02-23T20:00:30 *** ser
41 2010-02-23T20:05:34 *** ser
42 2010-02-23T20:07:07 *** kursor
43 2010-02-23T20:28:52 <ThomasWaldmann> re
44 2010-02-23T21:03:19 * ThomasWaldmann updates docs/CHANGES
45 2010-02-23T21:16:36 *** grzywacz
46 2010-02-23T21:23:26 <dreimark> i think it is related to cookie_lifetime
47 2010-02-23T21:37:17 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5578:d7a546ed1596 1.9/docs/CHANGES: updated docs/CHANGES
48 2010-02-23T21:42:09 <ThomasWaldmann> dreimark: yes, I didn't consider anonymous sessions being disabled
49 2010-02-23T21:46:13 *** kursor
50 2010-02-23T22:24:47 <ThomasWaldmann> that session.finalize is twisted X)
51 2010-02-23T22:25:06 <dreimark> .oO
52 2010-02-23T22:48:01 *** waldi
53 2010-02-23T22:48:06 *** waldi
54 2010-02-23T23:01:39 <ThomasWaldmann> ok, i think i have it right now :)
55 2010-02-23T23:02:06 <ThomasWaldmann> i also added a check for expired sessions in get_session
56 2010-02-23T23:02:31 <ThomasWaldmann> usually the browser should kill the cookie, but who knows ... :)
57 2010-02-23T23:02:52 <ThomasWaldmann> moin waldi
58 2010-02-23T23:13:32 <dreimark> i was thinking about if we can get a problem if one changes the session lifetime somehow and doesn't kill existing sessions
59 2010-02-23T23:13:40 <dreimark> or if this is necessary to do so
60 2010-02-23T23:16:11 <ThomasWaldmann> stored sessions will run until they expire
61 2010-02-23T23:18:07 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5579:e535351eab08 1.9/MoinMoin/web/session.py: sessions: fix not being able to log out if anon sessions disabled, check expiry
62 2010-02-23T23:25:28 <ThomasWaldmann> gna
63 2010-02-23T23:32:55 <dreimark> py.test ?
64 2010-02-23T23:44:48 <dreimark> that testwiki i have figures shome failures and it is still stucking in search tests
65 2010-02-23T23:45:26 <CIA-47> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5580:af56baebf4d1 1.9/MoinMoin/web/session.py: sessions: fix expiry check
66 2010-02-23T23:45:30 <ThomasWaldmann> search tests suck
67 2010-02-23T23:50:38 <dreimark> 99 4.3 18:53.36 py.test
MoinMoin: MoinMoinChat/Logs/moin-dev/2010-02-23 (last edited 2010-02-22 23:15:02 by IrcLogImporter)