1 2013-06-22T01:29:11 *** bilal
2 2013-06-22T02:56:55 *** brunomartin
3 2013-06-22T03:09:27 *** brunomartin
4 2013-06-22T03:19:29 *** brunomartin
5 2013-06-22T06:35:30 *** xjjk
6 2013-06-22T06:35:56 *** xjjk
7 2013-06-22T07:10:18 *** ashutosh1461
8 2013-06-22T07:26:03 <TheSheep> http://api.jquerymobile.com/category/methods/ <-- this doesn't look like jquery mobile site
9 2013-06-22T07:26:16 *** ashutosh1461
10 2013-06-22T07:27:44 <TheSheep> http://view.jquerymobile.com/1.3.1/dist/demos/ <-- this does
11 2013-06-22T07:51:16 <sharky93> moin
12 2013-06-22T08:23:44 *** breton
13 2013-06-22T08:56:09 *** hellboy2k8
14 2013-06-22T08:56:14 *** hellboy2k8
15 2013-06-22T08:58:41 *** penguinRaider
16 2013-06-22T09:14:23 *** ashutosh1461
17 2013-06-22T09:17:11 *** penguinRaider
18 2013-06-22T09:19:14 *** penguinRaider
19 2013-06-22T10:11:53 <ashutosh1461> ThomasWaldmann, what name would we add to the trail ? the fqname or just name ?
20 2013-06-22T10:12:20 <ashutosh1461> simple name wont make much sense in case the field is tags or itemid
21 2013-06-22T10:13:20 <ashutosh1461> but composite name can get large so we can use abbreviated name
22 2013-06-22T10:13:36 <ashutosh1461> abbreviated fqname*
23 2013-06-22T10:35:57 <ThomasWaldmann> the trail needs to have wikiname, namespace(s), name - all you need to build a url
24 2013-06-22T10:36:13 <ThomasWaldmann> ashutosh1461:
25 2013-06-22T10:37:38 <ThomasWaldmann> trail is a short term thing, so there isn't much point in using itemids
26 2013-06-22T10:40:33 <ThomasWaldmann> TheSheep: api site is not based on jqm
27 2013-06-22T10:44:38 <ThomasWaldmann> (the same theme is used for other api.*.com sites related to jquery)
28 2013-06-22T12:21:44 *** TheSheep
29 2013-06-22T12:21:56 *** TheSheep
30 2013-06-22T12:31:26 <ana_ananasova> ThomasWaldmann: what delimiter should we choose for subscriptions? apparently a colon can be a part of namespace name.
31 2013-06-22T12:32:36 <ana_ananasova> what i mean is `TAG:ns:tagname`
32 2013-06-22T12:41:44 <spy_> xiaq: Hi! It would be cool to hear your comments and suggestions from the point of view of your project on http://etherpad.osuosl.org/moin-commenting
33 2013-06-22T12:42:44 <ThomasWaldmann> ana_ananasova: you split off the leftmost part and then give the rest to the parsing code found in the interwiki module
34 2013-06-22T12:43:35 <ThomasWaldmann> (and keep an eye on what ashutosh1461 is doing there)
35 2013-06-22T12:44:47 <ana_ananasova> ok
36 2013-06-22T12:45:12 <ThomasWaldmann> (and currently, something like ns1:ns2 means nested namespaces, not : being part of a namespace name)
37 2013-06-22T12:46:24 <ThomasWaldmann> ana_ananasova: did you already play with whoosh to check whether the approach works?
38 2013-06-22T12:46:46 <ashutosh1461> ThomasWaldmann, i tested using the Flask 0.10.1, '@' goes to %40 , so what to do ?
39 2013-06-22T12:47:10 <ThomasWaldmann> ashutosh1461: how did you verify that?
40 2013-06-22T12:48:04 <ThomasWaldmann> and what's your werkzeug version?
41 2013-06-22T12:48:09 <ashutosh1461> wrote a dummy program
42 2013-06-22T12:48:19 <ThomasWaldmann> > pastebin
43 2013-06-22T12:48:56 <ashutosh1461> Werkzeug==0.9.1
44 2013-06-22T12:49:00 <ana_ananasova> simple tests - by using searcher.documents(). or you mean smth more heavy loaded to check the performance?
45 2013-06-22T12:51:15 <ashutosh1461> http://rn0.ru/show/ubP69HtcAbojYy6W1Qfp/
46 2013-06-22T12:51:17 <ThomasWaldmann> ana_ananasova: ashutosh1461: spy_: please write your diary entries "same day" (and thus: only for one day at one occasion)
47 2013-06-22T12:51:44 <ashutosh1461> ThomasWaldmann, ^^
48 2013-06-22T12:52:02 <ThomasWaldmann> ana_ananasova: POC
49 2013-06-22T12:54:25 <ana_ananasova> possible. tested before proposing
50 2013-06-22T12:58:32 <ThomasWaldmann> waldi: ashutosh1461: + and : get through "as is", @ -> %40, $ -> %24
51 2013-06-22T12:59:22 <ThomasWaldmann> ana_ananasova: i mean did you write some proof of concept code for that lookup?
52 2013-06-22T12:59:57 <ThomasWaldmann> (or any code inside moin already?)
53 2013-06-22T13:01:25 <ana_ananasova> i'll put it on pastebin
54 2013-06-22T13:03:06 *** penguinRaider
55 2013-06-22T13:03:48 *** penguinRaider
56 2013-06-22T13:05:02 <ThomasWaldmann> waldi: ashutosh1461: /%21%24%25%26%28%29%3D%2A+%7E-_.:%2C%3B|%5B%5D%7B%7D/ashutosh <- i put all sorts of chars into a flask routing rule and that is the url it makes from it
57 2013-06-22T13:12:52 <waldi> looks broken
58 2013-06-22T13:15:56 <ana_ananasova> http://rn0.ru/show/3303/
59 2013-06-22T13:19:34 *** penguinRaider
60 2013-06-22T13:21:39 *** penguinRaider
61 2013-06-22T13:27:38 *** penguinRaider
62 2013-06-22T13:29:05 <ThomasWaldmann> waldi: i just posted that to #pocoo, dasich wants to have a look later today
63 2013-06-22T13:29:16 <ThomasWaldmann> ana_ananasova: looking...
64 2013-06-22T13:31:00 <ThomasWaldmann> ana_ananasova: why is all_subscriptions a set?
65 2013-06-22T13:38:13 <ThomasWaldmann> ana_ananasova: you can just do for subscription in searcher.lexicon("subscribed_items"):
66 2013-06-22T13:38:20 <ThomasWaldmann> works as good
67 2013-06-22T13:38:42 <ThomasWaldmann> ok, so PoC works
68 2013-06-22T13:39:44 <ThomasWaldmann> waldi: did you recently look through the ideas on http://etherpad.osuosl.org/moin-mail ?
69 2013-06-22T13:48:57 <waldi> it looks weird. i thought whoosh should help solving the problem and not create new ones
70 2013-06-22T13:49:02 <ThomasWaldmann> ana_ananasova: xiaq: link your public repo from the table there: https://moinmo.in/GoogleSoc2013
71 2013-06-22T13:49:56 <ThomasWaldmann> waldi: is it related to the EP contents? if so, can we add stuff there / discuss there? line no?
72 2013-06-22T13:51:00 <waldi> ThomasWaldmann: the whole, how to find subscriptions pseudo code
73 2013-06-22T13:51:45 <waldi> if this is an index just first compile a search query and then let the index figure out all matching items. not loop over all
74 2013-06-22T13:57:33 <ana_ananasova> ThomasWaldmann: the public repo was linked long ago.
75 2013-06-22T14:39:25 <ThomasWaldmann> ehrm, that was for ashutosh1461, not ana_ananasova. ashutosh1461: link your public repo from the table there: https://moinmo.in/GoogleSoc2013
76 2013-06-22T16:26:21 <ThomasWaldmann> ashutosh1461: if you fixed stuff as reaction to the codereview, upload a new patchset to same codereview id
77 2013-06-22T16:28:46 <ashutosh1461> ThomasWaldmann, did you fully reviewed the cr? or you want me to update the test_interwiki.py file you reviewed ?
78 2013-06-22T16:54:58 <ana_ananasova> ThomasWaldmann: for subscribers determination create methods in Item class?
79 2013-06-22T16:58:17 <ThomasWaldmann> ashutosh1461: not yet, but it thought you could just fix what was already found
80 2013-06-22T16:58:21 <ThomasWaldmann> i*
81 2013-06-22T16:59:29 <ThomasWaldmann> ana_ananasova: not sure, it somehow is in between items and users...
82 2013-06-22T17:00:29 <ana_ananasova> MoinMoin/util/subscriptions.py ?
83 2013-06-22T17:10:48 <ThomasWaldmann> yeah, you can start there. if we find a better place, we'll move.
84 2013-06-22T17:16:15 <ThomasWaldmann> ashutosh1461: we are discussing the %XX issue on #pocoo
85 2013-06-22T17:29:52 <ThomasWaldmann> ashutosh1461: btw, we can ignore that detail for now and just use it as is. if they fix it later or we choos other prettier chars, we can do a quick systematic change...
86 2013-06-22T17:30:43 <ashutosh1461> is there a specific reason they quote '@' and not '+' ?
87 2013-06-22T17:31:54 <ThomasWaldmann> no idea. they randomly quoted or not quoted + in the past, depending on werkzeug version :D
88 2013-06-22T17:33:05 <ashutosh1461> and what about the whoosh issue ?
89 2013-06-22T17:34:11 <ashutosh1461> https://bitbucket.org/mchaput/whoosh/issue/321/
90 2013-06-22T17:34:57 <ana_ananasova> btw, get_subscribers would return a set of profile itemids or emails?
91 2013-06-22T17:36:41 <ThomasWaldmann> ashutosh1461: until whoosh makes that more consistent (if ever, is an issue with backwards compat), you could emulate sane behaviour in your code :)
92 2013-06-22T17:37:42 <ashutosh1461> yes, we can do that.
93 2013-06-22T17:44:16 <ana_ananasova> ThomasWaldmann: ^
94 2013-06-22T17:48:37 <ashutosh1461> ThomasWaldmann, does not look clean http://imagebin.org/index.php?mode=image&id=262195
95 2013-06-22T17:54:16 <ThomasWaldmann> ana_ananasova: neither
96 2013-06-22T17:55:24 <ThomasWaldmann> if you return userprofile itemids, you would have to do another lookup (expensive) to get data you (mostly) already had in the search results
97 2013-06-22T17:56:25 <ThomasWaldmann> if you return just the email addresses, think about translations - you want to send german users emails in german (not english, not the language of the current editor)
98 2013-06-22T17:56:45 <ThomasWaldmann> s/german users/german subscribers/
99 2013-06-22T17:57:50 <ThomasWaldmann> ashutosh1461: instead of @itemid/1234... you would rather show itemid: 1234...
100 2013-06-22T17:59:35 <ana_ananasova> return User objects then
101 2013-06-22T18:02:16 <ThomasWaldmann> ashutosh1461: see whoosh.formats.tokens - you see it is handling lists/tuples differently than non-lists/tuples
102 2013-06-22T18:07:12 <ashutosh1461> yes
103 2013-06-22T18:08:42 <ashutosh1461> but we can write our own analyzer anyways
104 2013-06-22T18:12:59 <ThomasWaldmann> there is another issue: if you have NAME = ["a/b", "c/d"] and you analyze/tokenize that all, you would end up with ["a", "b", "c", "d", ...] token
105 2013-06-22T18:13:36 <ThomasWaldmann> if you then search for b/c, will it find a match on that item?
106 2013-06-22T18:14:45 <ashutosh1461> if you tokenize the query and then search yes
107 2013-06-22T18:27:13 <ThomasWaldmann> ashutosh1461: you know how to work around this whoosh issue now?
108 2013-06-22T18:28:40 <ashutosh1461> ThomasWaldmann, yes i think i would be able to handle it, should not be tough !
109 2013-06-22T18:29:05 <ThomasWaldmann> how would you do it?
110 2013-06-22T18:29:27 <ronny> sup
111 2013-06-22T18:32:12 <ashutosh1461> ThomasWaldmann, write a analyzer function of my own
112 2013-06-22T18:32:44 <ThomasWaldmann> and what would it do?
113 2013-06-22T18:33:27 <ashutosh1461> that will tokenize all the elements present in the name list
114 2013-06-22T18:33:42 <ashutosh1461> the job that the existing analyzer is expected to perform
115 2013-06-22T18:34:13 <ThomasWaldmann> if it is a list, it won't call your analyzer
116 2013-06-22T18:36:21 <ashutosh1461> which is the root of the problem, i see.
117 2013-06-22T18:37:44 <ashutosh1461> i will think about it, once the current namespace stuff is over
118 2013-06-22T18:38:06 <ThomasWaldmann> I'ld also suggest to solve this in a generic way.
119 2013-06-22T18:39:17 <ashutosh1461> currently i am re-writing items/__init__.py to support composite names
120 2013-06-22T19:16:46 *** ashutosh1461
121 2013-06-22T21:09:59 <sharky93> TheSheep: ThomasWaldmann j-m is customisable but it is not much easy to develop a complex website with it.
122 2013-06-22T21:10:40 <sharky93> It's proving out to be that way.
123 2013-06-22T21:14:52 <ThomasWaldmann> what specifically did you find, sharky93 ?
124 2013-06-22T21:29:10 <sharky93> ThomasWaldmann: very less support of the common grid system, you'll have to build it all by yourself i think.
125 2013-06-22T21:29:52 <sharky93> The basic styles are not suitable for dekstop viewing i think.. individual elements look ok, but when put together looks a lot like early man :(
126 2013-06-22T21:31:33 <spy_> ThomasWaldmann: would it be possible to show items by its itemid? In my particular case, do you think we need the possibility to open/edit a single comment item on its separate +show/+modify view? Or just provide its rendering/modification on the +show view of the content item (the comment is made for)?
127 2013-06-22T21:32:21 <sharky93> ThomasWaldmann: i'm trying out basic things, it already looks not good. @ http://web.iiit.ac.in/~rishabh.raj/jm/jmobile.html
128 2013-06-22T21:32:26 <sharky93> TheSheep: ^
129 2013-06-22T21:57:18 <sharky93> the mobile first UI is hurting a bit on desktops
130 2013-06-22T22:05:18 <ThomasWaldmann> spy_: yes, ashutosh is working on access by itemid
131 2013-06-22T22:06:12 <ThomasWaldmann> (to be not blocked in case this is not working when you need it, you can also put the itemid into the name and access by that, until it works by itemid)
132 2013-06-22T22:09:42 <ThomasWaldmann> sharky93: that doesn't scale too well with shrinking width
133 2013-06-22T22:12:38 * ana_ananasova says gn and see you tomorrow at the weekly meeting
134 2013-06-22T22:13:32 <ReimarBauer> gn
135 2013-06-22T22:14:39 <ThomasWaldmann> gn ana_ananasova
136 2013-06-22T22:17:49 <sharky93> ThomasWaldmann: yes, we'll have to specify media queries for that, at width below a certain value.
137 2013-06-22T22:21:13 <spy_> ThomasWaldmann: ok, thanks
138 2013-06-22T22:24:28 <ThomasWaldmann> sharky93: can you try to add a item context menu somehow (with the different views?)
139 2013-06-22T22:33:02 <sharky93> ThomasWaldmann: you mean just imitating some different views? you want to see the AJAX loading?
140 2013-06-22T22:40:20 <ThomasWaldmann> foobar has a global linkbar on the left and item context (views) on the right
141 2013-06-22T22:40:55 <ThomasWaldmann> something that can be used to get to different views, not necessarily a link bar
142 2013-06-22T22:44:08 <sharky93> ThomasWaldmann: i used the layout on the left.
143 2013-06-22T22:44:15 <sharky93> - @ http://web.iiit.ac.in/~rishabh.raj/jm/jmobile.html
144 2013-06-22T22:44:19 <sharky93> looks rather ugly :(
145 2013-06-22T22:50:01 <ThomasWaldmann> well, the margins around the main text are too small. also, maybe much of that should not be permanently visible
146 2013-06-22T22:55:17 <ThomasWaldmann> TheSheep: ^^
147 2013-06-22T22:56:46 <sharky93> ThomasWaldmann: hmm, i fixed that, but it still doesnt feel right, i mean look at this @ http://icon.sheep.art.pl/moin/ and you instantly feel happy :)
148 2013-06-22T22:58:49 <sharky93> plus from a frameworks perspective, i've had to put a number of customizations which should not be ideally there ..
149 2013-06-22T22:58:49 <ThomasWaldmann> yes, looks better. but that is also due to the colours.
150 2013-06-22T23:56:59 <sharky93> gn