1 2010-05-20T01:04:58 *** diofeher
2 2010-05-20T02:21:30 *** diofeher_
3 2010-05-20T02:24:08 *** diofeher
4 2010-05-20T02:24:41 *** diofeher_
5 2010-05-20T03:59:36 *** diofeher
6 2010-05-20T05:14:00 *** kursor
7 2010-05-20T05:42:51 *** kursor
8 2010-05-20T08:09:34 <dreimark> .oO ubuntu 10.04 [Thu May 20 08:02:06 2010] [error] Exception KeyError: KeyError(140622858094400,) in <module 'threading' from '/usr/lib/python2.6/threading.pyc'> ignored
9 2010-05-20T08:18:15 <dreimark> ouch i found also another filesystem limitation 31998 page dirs in ext3 only
10 2010-05-20T08:18:32 *** gwork
11 2010-05-20T08:19:24 <franklin> ThomasWaldmann: I made a few test with werkzeug 0.6 (WSGI wiki setup, accoutn creation, language setup, using some macro, enabling Xapian and a few other tests) nothing seemed broken so far.
12 2010-05-20T08:19:28 <franklin> Sorry for my late reply (I am fairly busy with real life at the moment).
13 2010-05-20T08:19:53 <dreimark> franklin: do you know how much dirs I can create in ext4?
14 2010-05-20T08:21:20 <dreimark> ah found it Right now the maximum possible number of sub directories contained in a single directory in Ext3 is 32000. Ext4 breaks that limit and allows unlimited number of sub directories.
15 2010-05-20T08:24:14 <franklin> dreimark: I never tested on Linux, but on Win2K3, a folder with 65000 entries plains sucks from performance PoV, especially to list the content.
16 2010-05-20T08:33:55 *** ei-grad
17 2010-05-20T09:00:15 <dreimark> franklin: i will see in the next hours, but in case i have performace issues i add a SSD Drive
18 2010-05-20T09:00:29 <dreimark> ei-grad: how is it going on
19 2010-05-20T09:00:49 <dreimark> brb
20 2010-05-20T09:51:49 <ei-grad> sorry, being away... dreimark: i'll start it later today
21 2010-05-20T09:53:50 <dreimark> ei-grad: please use the wiki to show what your are doing
22 2010-05-20T09:54:04 <dreimark> using the MonthCalendar macro
23 2010-05-20T10:58:17 *** blathijs
24 2010-05-20T11:10:02 <dreimark> ei-grad: do you join europython ?
25 2010-05-20T11:10:13 <dreimark> http://moinmo.in/EuroPython2010
26 2010-05-20T11:29:48 <valeuf> dreimark: looks interesting
27 2010-05-20T11:29:56 <valeuf> I will see if I can come
28 2010-05-20T12:28:35 *** ei-grad
29 2010-05-20T14:07:59 <ThomasWaldmann> franklin: good that it works now :)
30 2010-05-20T14:09:45 <ThomasWaldmann> dreimark: don't forget dir_index if you are creating a new file system (ext3, not sure about ext4)
31 2010-05-20T14:11:03 *** xjjk
32 2010-05-20T14:11:03 *** elmo
33 2010-05-20T14:11:04 *** dennda
34 2010-05-20T14:11:04 *** aa_
35 2010-05-20T14:11:56 *** ronny_
36 2010-05-20T14:13:03 *** ronny
37 2010-05-20T14:14:12 *** xjjk
38 2010-05-20T14:14:12 *** aa_
39 2010-05-20T14:14:12 *** dennda
40 2010-05-20T14:14:12 *** elmo
41 2010-05-20T14:22:39 *** DmitryAndreev
42 2010-05-20T14:25:28 <ThomasWaldmann> welcome back, dmitry!
43 2010-05-20T14:28:07 <eSyr> :)
44 2010-05-20T14:28:38 <DmitryAndreev> .)
45 2010-05-20T14:28:53 <eSyr> vim x86/mopdule.c
46 2010-05-20T14:29:03 <eSyr> oops, wrong window.
47 2010-05-20T14:29:08 <TheSheep> file not found
48 2010-05-20T14:29:39 <DmitryAndreev> just printed my diploma,
49 2010-05-20T14:29:51 <eSyr> TheSheep: "x86/module.c" [New DIRECTORY]
50 2010-05-20T14:29:56 <eSyr> (i suppose)
51 2010-05-20T16:15:17 *** diofeher
52 2010-05-20T16:36:41 <dreimark> hi DmitryAndreev
53 2010-05-20T16:36:49 <DmitryAndreev> hello
54 2010-05-20T16:46:49 <TheSheep> hi diofeher
55 2010-05-20T16:46:55 <TheSheep> diofeher: any progress?
56 2010-05-20T16:48:03 <diofeher> not yet :/ i've checkout the projects
57 2010-05-20T16:48:27 <diofeher> and looked for how is working
58 2010-05-20T16:48:57 <diofeher> but it's not clear in my mind yet, you know?
59 2010-05-20T16:51:51 <TheSheep> we should go through it together
60 2010-05-20T16:51:59 <TheSheep> do you have time now?
61 2010-05-20T16:52:27 <TheSheep> I'm finishing my work in 10 minutes, so I'll be available in 45 minutes
62 2010-05-20T16:52:33 <TheSheep> (commuting)
63 2010-05-20T16:57:05 <diofeher> TheSheep: yes, i'm available now
64 2010-05-20T17:26:08 *** gwork
65 2010-05-20T17:48:24 <diofeher> TheSheep: hey, i have to go to college now :/
66 2010-05-20T17:48:39 <diofeher> be back in 6/7 hours
67 2010-05-20T17:48:42 <diofeher> bye
68 2010-05-20T17:48:43 *** diofeher
69 2010-05-20T18:05:08 *** DmitryAndreev
70 2010-05-20T18:23:01 <TheSheep> damn, took me longer than anticipated
71 2010-05-20T18:26:44 <ThomasWaldmann> maybe dio should put his schedule into the wiki, so we can optimize timing
72 2010-05-20T18:42:33 *** valeuf
73 2010-05-20T19:14:23 *** kursor
74 2010-05-20T19:22:35 *** valeuf
75 2010-05-20T19:23:42 *** ei-grad
76 2010-05-20T19:56:34 *** kursor
77 2010-05-20T20:39:29 *** ei-grad
78 2010-05-20T20:43:01 *** ei-grad
79 2010-05-20T20:50:28 <ThomasWaldmann> re
80 2010-05-20T21:01:02 <ei-grad> ThomasWaldmann: where ldap_testbase.py come from? is there any catalog of such samples for py.test? or it is just one, wroten especially for ldap, and spreading from project to project?
81 2010-05-20T21:02:51 * ThomasWaldmann looks, too long ago
82 2010-05-20T21:05:08 <ThomasWaldmann> ldap_testbase is from me
83 2010-05-20T21:08:19 <ei-grad> oops, I thought that I had seen it used in some another project... seems wrong)
84 2010-05-20T21:45:12 *** DmitryAndreev
85 2010-05-20T22:00:26 *** ei-grad
86 2010-05-20T22:20:28 <TheSheep> http://moinmo.in/RadomirDopieralski/RESTfulURLs <-- any feedback welcome
87 2010-05-20T22:25:14 <dreimark> re
88 2010-05-20T22:26:39 <dreimark> well there is no reason why you should not find foss code on other places
89 2010-05-20T22:43:03 <ThomasWaldmann> TheSheep: looks good
90 2010-05-20T22:43:18 <ThomasWaldmann> but one problem: we'll need markup conversion :|
91 2010-05-20T22:43:29 <TheSheep> why?
92 2010-05-20T22:43:55 <ThomasWaldmann> for the [...|...|&do=get]] or &action=whatever
93 2010-05-20T22:44:02 <TheSheep> markup stays the same, you just generate different urls for it
94 2010-05-20T22:45:41 <TheSheep> so [[||do=get]] will just generate a /+download/... url
95 2010-05-20T22:46:09 <ThomasWaldmann> but that'll make query args parsing complex
96 2010-05-20T22:46:11 <TheSheep> there is no coupling here
97 2010-05-20T22:46:19 <ThomasWaldmann> currently we just build the qs from that
98 2010-05-20T22:46:40 <ThomasWaldmann> then we build either qs or url from that, depeding on the keys?
99 2010-05-20T22:46:47 <TheSheep> well, internal model leaks
100 2010-05-20T22:48:01 <ThomasWaldmann> it might be cleaner to convert markup
101 2010-05-20T22:48:38 <ThomasWaldmann> so that [[file.zip|foo|&do=get]] would be [[+get/file.zip|foo]]
102 2010-05-20T22:50:27 <TheSheep> you can have both work
103 2010-05-20T22:51:07 <ThomasWaldmann> what's with [[http://foo.org|foo get|&do=get]]?
104 2010-05-20T22:51:24 <TheSheep> I guess you treat them the old way
105 2010-05-20T22:51:38 <ThomasWaldmann> so sometimes it'll give qs, sometimes path?
106 2010-05-20T22:52:28 * ThomasWaldmann wouldn't want to write another markup converter
107 2010-05-20T22:53:53 <TheSheep> urls shouldn't be so tightly bound to markup
108 2010-05-20T22:54:18 <ThomasWaldmann> well, it's not just about wiki urls
109 2010-05-20T22:54:33 <ThomasWaldmann> our current markup supports generic links with qs
110 2010-05-20T22:55:21 <ThomasWaldmann> otoh, if it is a non-wiki url, you could also put the qs into the target part
111 2010-05-20T22:55:27 <TheSheep> the problem is not that it supports them, the problem is that it forces you to use them
112 2010-05-20T22:56:26 <TheSheep> they would be needed by 1% of users if not the horrible way attachment links work now
113 2010-05-20T22:57:35 <ThomasWaldmann> yeah
114 2010-05-20T22:58:55 <TheSheep> you need to change those urls anyways
115 2010-05-20T23:00:19 <TheSheep> from /PageName?action=AttachFile?do=get?target=AttachmentName to /PageName/AttachmentName?action=raw
116 2010-05-20T23:00:28 <ThomasWaldmann> yes, but without restful urls, it would be [[attachment:foo|...|&do=get]] -> [[/foo|...|&do=get]]
117 2010-05-20T23:01:18 <TheSheep> that can still work
118 2010-05-20T23:01:47 <TheSheep> so you will change 'action' into 'do'?
119 2010-05-20T23:02:00 <ThomasWaldmann> that's already done in 2.0
120 2010-05-20T23:02:07 <TheSheep> what about all the other urls?
121 2010-05-20T23:02:21 <ThomasWaldmann> (because it is shorter and because attachments use it often)
122 2010-05-20T23:02:40 <TheSheep> what if I did [[foo||&action=FindPage]] ?
123 2010-05-20T23:03:06 <ThomasWaldmann> there is no such action. but yes, in principle we would've to convert similar markup.
124 2010-05-20T23:03:15 <TheSheep> will you convert that to [[foo||&do=FindPage]]?
125 2010-05-20T23:03:28 <TheSheep> so you need to do it anyways
126 2010-05-20T23:03:33 <ThomasWaldmann> no (not yet at least)
127 2010-05-20T23:05:11 <TheSheep> personally I would specialcase &do and let people use both |&do=get and /+get/
128 2010-05-20T23:05:26 <TheSheep> generating /+get links in both cases
129 2010-05-20T23:05:52 <TheSheep> lets them avoid relearning
130 2010-05-20T23:06:08 <TheSheep> converting markup is easy, converting users is hard
131 2010-05-20T23:07:10 <ThomasWaldmann> so you want to look at the link target and handle &do=get depending on whether it is wikipage or external?
132 2010-05-20T23:07:26 <TheSheep> yes
133 2010-05-20T23:08:17 <ThomasWaldmann> that can create problematic cases
134 2010-05-20T23:08:38 <TheSheep> actually, depedning on whether it's attachment: or external
135 2010-05-20T23:09:19 <TheSheep> if you want to avoid converting, you need to keep the attachment: support
136 2010-05-20T23:09:36 <TheSheep> so, you have a good indicator of 'legacy' syntax
137 2010-05-20T23:09:44 <TheSheep> and you only need to specialcase that
138 2010-05-20T23:09:54 <ThomasWaldmann> hmm, ok
139 2010-05-20T23:10:15 <ThomasWaldmann> but [[+get/file]] would be the new way
140 2010-05-20T23:10:27 <TheSheep> yes, it would be intuitive then
141 2010-05-20T23:10:39 <TheSheep> and people wouldn't make titles starting with +
142 2010-05-20T23:11:30 <ThomasWaldmann> so we just disallow pages starting with + instead of doubling it in the url?
143 2010-05-20T23:12:21 <ThomasWaldmann> otherwise you won't know whether it is item name or action
144 2010-05-20T23:13:17 <TheSheep> I really don't know, the escape chaacter option looks promising
145 2010-05-20T23:13:33 <ThomasWaldmann> hmm, another issue:
146 2010-05-20T23:14:23 <ThomasWaldmann> hm, no, no issue, but we can't internally use relativ urls
147 2010-05-20T23:14:43 <TheSheep> why?
148 2010-05-20T23:15:33 <ThomasWaldmann> because [[foo]] needs to link to /foo, no matter where you currently are
149 2010-05-20T23:15:47 <TheSheep> it would lead to the same action for a differnt page
150 2010-05-20T23:16:15 <TheSheep> well, [[foo]] only gets rendeed in content with the view action anyways
151 2010-05-20T23:16:39 <TheSheep> maybe in previews too, but I'm not sure links should even work in previews
152 2010-05-20T23:17:00 <TheSheep> diffs...
153 2010-05-20T23:17:20 <TheSheep> hmm... I guess it needs a per-case analysis
154 2010-05-20T23:17:33 <TheSheep> for things like /+get it makes sense
155 2010-05-20T23:20:52 <TheSheep> ThomasWaldmann: btw, things like 'rename' or 'delete' should display a form on GET and only do the action on POST
156 2010-05-20T23:21:16 <ThomasWaldmann> yeah, sure
157 2010-05-20T23:21:43 <TheSheep> ah, and once we are at it, how about doing a 303 redirect after POSTs?
158 2010-05-20T23:22:16 <ThomasWaldmann> btw, can you try whether a pagename like "foo?" in a restful url gets parsed correctly by werkzeug?
159 2010-05-20T23:22:19 <TheSheep> I know it makes it harder to display the message, but message may be stored in the session
160 2010-05-20T23:22:53 <TheSheep> ThomasWaldmann: foo? is foo
161 2010-05-20T23:23:06 <TheSheep> ThomasWaldmann: unescaped ? is always the beginning of a query string
162 2010-05-20T23:23:14 <TheSheep> ThomasWaldmann: that's handled by the web server
163 2010-05-20T23:23:25 <TheSheep> ThomasWaldmann: before python gets to it
164 2010-05-20T23:23:58 <ThomasWaldmann> escaped
165 2010-05-20T23:24:12 <ThomasWaldmann> (should work, but does it?)
166 2010-05-20T23:24:23 <TheSheep> http://test.hatta-wiki.org/foo%3F
167 2010-05-20T23:25:31 <TheSheep> feel free to play there, but I do some silly / escaping, so slashes may behave strangely
168 2010-05-20T23:25:47 <ThomasWaldmann> ok
169 2010-05-20T23:27:39 <ThomasWaldmann> works
170 2010-05-20T23:30:14 <TheSheep> I will write down the notes from this tomorrow, good night
171 2010-05-20T23:35:15 <TheSheep> you can't use [[+get/ItemName]]
172 2010-05-20T23:35:27 <TheSheep> because for nested items you would need absolute paths
173 2010-05-20T23:35:55 <TheSheep> [[+get/Item/SubItem]] when you can do just [[/SubItem]]
174 2010-05-20T23:36:15 <TheSheep> better just do [[/SubItem||+get]]
175 2010-05-20T23:45:23 <ThomasWaldmann> gn TheSheep
176 2010-05-20T23:46:25 <ThomasWaldmann> yeah, so we have + now additionally to &. mixing?
