1 2013-04-04T01:28:30 *** dwcramer
2 2013-04-04T02:04:20 *** dwcramer
3 2013-04-04T02:19:00 *** dwcramer
4 2013-04-04T02:58:50 *** rihnapstor
5 2013-04-04T02:59:13 *** rihnapstor
6 2013-04-04T04:54:22 * rihnapstor says moin
7 2013-04-04T04:58:46 <rihnapstor> dreimark: ThomasWaldmann please review ,y code on issue#308 https://codereview.appspot.com/8248045/ .I am figuring out how can i link the common.css in atom.html. Any style changes made are not reflecting into browser.
8 2013-04-04T04:59:28 * rihnapstor going for college now
9 2013-04-04T04:59:32 <rihnapstor> :)
10 2013-04-04T04:59:35 *** rihnapstor
11 2013-04-04T05:30:57 *** birkenfeld_
12 2013-04-04T05:32:16 *** vpv_
13 2013-04-04T05:38:11 *** birkenfeld
14 2013-04-04T05:38:11 *** vpv
15 2013-04-04T05:38:14 *** birkenfeld_
16 2013-04-04T06:30:02 *** birkenfeld
17 2013-04-04T06:30:02 *** birkenfeld
18 2013-04-04T09:05:37 *** greg_f
19 2013-04-04T10:44:13 <ThomasWaldmann> moin
20 2013-04-04T10:49:43 *** dwcramer
21 2013-04-04T10:51:36 *** vpv_
22 2013-04-04T11:35:47 <sunu> ThomasWaldmann: https://codereview.appspot.com/8361043 A different and cleaner approach for web interface for index operations
23 2013-04-04T11:50:51 <ThomasWaldmann> review done
24 2013-04-04T11:51:00 * ThomasWaldmann brb
25 2013-04-04T12:01:13 *** ashutosh1461
26 2013-04-04T12:01:57 <ashutosh1461> moin everyone :)
27 2013-04-04T12:02:11 *** dwcramer
28 2013-04-04T12:08:20 <ThomasWaldmann> hi ashutosh1461
29 2013-04-04T12:20:58 <ashutosh1461> ThomasWaldmann, what is the problem in http://goo.gl/tz4d0 ? and the current revision is None in the case we create a new item.
30 2013-04-04T12:21:39 <ashutosh1461> i mean the block in the line(202)
31 2013-04-04T12:25:59 <ThomasWaldmann> point me to the place where you have seen that it is None
32 2013-04-04T12:31:57 <ashutosh1461> i commented that line196 should be executed in a try-catch block
33 2013-04-04T12:32:21 <ashutosh1461> if we get a keyerror there, we will know that the current-rev is none.
34 2013-04-04T12:33:27 <ThomasWaldmann> let's discuss about real code, not code that is yet to be written...
35 2013-04-04T12:33:33 * ThomasWaldmann bbl
36 2013-04-04T12:35:33 <ashutosh1461> In the real code, we can check here, https://bitbucket.org/thomaswaldmann/moin-2.0/src/1d8366403411a4b6f2b4eeebfe3278a52fe3d9d6/MoinMoin/items/__init__.py?at=default#cl-462
37 2013-04-04T12:36:31 *** dave_largo
38 2013-04-04T12:48:39 *** dwcramer
39 2013-04-04T12:51:15 <ThomasWaldmann> ashutosh1461: line number? (your link was truncated)
40 2013-04-04T12:51:36 <ashutosh1461> 462
41 2013-04-04T12:53:40 <ThomasWaldmann> well, that's THERE, but how is that related to what you are (not) doing in line 196..202?
42 2013-04-04T12:54:14 <ThomasWaldmann> so i suggest you just implement what you already got as a suggestion and update the codereview with improved code.
43 2013-04-04T12:56:40 <ashutosh1461> and should i attach the error to the form or the json element ?
44 2013-04-04T12:58:10 <ThomasWaldmann> well, just imagine yourself filling out a form with many fields
45 2013-04-04T12:58:37 <ThomasWaldmann> do you like if it just says "there was something wrong, fix it" or do you like getting pointed to the precise place that was wrong?
46 2013-04-04T13:01:11 <ashutosh1461> but there is only one name field in the whole form, so the user would get an idea of whats wrong? and what i think is that the user should be able to see the error response in the first look.
47 2013-04-04T13:03:07 <ThomasWaldmann> we usually attach error to the specific form field, except if there is no clear field (e.g. if on password change, the 2 passwords entered are different - you don't know which of them was wrong)
48 2013-04-04T13:04:37 <ashutosh1461> alright, fine ! :)
49 2013-04-04T13:04:40 <ThomasWaldmann> so, how about just modifying the validator already present for that field and maybe keeping the inner code (just the name validation) in a function (or validator) called from there?
50 2013-04-04T13:05:10 <ThomasWaldmann> sooner or later, that names field won't be in that json block any more
51 2013-04-04T13:12:50 <ashutosh1461> yes, that would be good, the JSON validation would also be done there. and regarding your point that set(names) could be None, should we also have a check that case when the user has removed the meta name from the json?
52 2013-04-04T13:19:34 <ThomasWaldmann> it should be there, at least a empty list
53 2013-04-04T13:20:00 <ThomasWaldmann> there is already some validation on the storage layer, but also work in progress
54 2013-04-04T13:33:11 *** ashutosh1461
55 2013-04-04T13:36:04 *** RogerHaase
56 2013-04-04T13:49:37 *** dwcramer
57 2013-04-04T14:12:48 *** dwcramer
58 2013-04-04T14:59:00 *** dwcramer
59 2013-04-04T15:07:21 *** bruno_
60 2013-04-04T15:13:02 *** rihnapstor
61 2013-04-04T15:39:09 *** sharky93
62 2013-04-04T16:03:44 *** dwcramer
63 2013-04-04T16:20:51 *** sharky93
64 2013-04-04T16:22:36 * rihnapstor brb
65 2013-04-04T16:22:39 *** rihnapstor
66 2013-04-04T16:29:03 *** dwcramer
67 2013-04-04T16:30:06 <sunu> ThomasWaldmann: https://codereview.appspot.com/8361043 updated. I didn't quite get what multiple processes comment though. Could you explain ?
68 2013-04-04T17:03:41 <ThomasWaldmann> if you use a real webserver (e.g. apache with mod-wsgi), it might create multiple processes of moin, multiple threads per process.
69 2013-04-04T17:04:22 <ThomasWaldmann> so, if you somehow late-initialize the indexing layer for one process, the other won't be influenced by that and still not work
70 2013-04-04T17:08:24 <sunu> ThomasWaldmann: if we fix the index and then restart the app, everything should work ?
71 2013-04-04T17:32:38 *** greg_f
72 2013-04-04T18:07:18 <ThomasWaldmann> sunu: yes.
73 2013-04-04T18:08:10 <ThomasWaldmann> or, maybe we could somehow delay initialization of indexing (until when it is really needed)
74 2013-04-04T18:56:44 <ThomasWaldmann> sunu: btw, maybe registering a new blueprint (or using the admin blueprint?) would make sense
75 2013-04-04T18:57:27 <ThomasWaldmann> because all the wiki-using blueprints need the index sooner or later, but if the cli stuff is separate, these pathes would not be used
76 2013-04-04T19:30:13 *** dave_largo
77 2013-04-04T19:34:54 *** dave_largo
78 2013-04-04T20:07:54 *** dwcramer
79 2013-04-04T20:23:34 *** dwcramer
80 2013-04-04T20:53:45 *** dwcramer
81 2013-04-04T21:17:21 *** dave_largo
82 2013-04-04T21:27:03 *** RogerHaase
83 2013-04-04T22:02:56 *** dwcramer
84 2013-04-04T22:08:02 *** dwcramer
85 2013-04-04T22:17:46 *** ashutosh1461
86 2013-04-04T22:20:27 <ashutosh1461> ThomasWaldmann, my idea is to define the NameNotunique Error class (used in items/__init__.py) in the forms.py file and later import that ?
87 2013-04-04T22:37:21 <ashutosh1461> There is also an ImportError if I try to import form items/__init__.py to forms.py