1 2011-06-15T00:12:38 <ThomasWaldmann> RogerHaase: hmm, just from reading the comments, it sounds still problematic.
2 2011-06-15T00:13:10 <RogerHaase> Use a less broken browser :-)
3 2011-06-15T00:13:45 <ThomasWaldmann> hehe
4 2011-06-15T00:13:47 <RogerHaase> did you look at http://unixpapa.com/js/testmouse.html
5 2011-06-15T00:14:58 <ThomasWaldmann> not yet
6 2011-06-15T00:18:16 <ThomasWaldmann> hmm, doesn't look that bad for me
7 2011-06-15T00:18:29 <RogerHaase> I am thinking that using [[SomePage|{{SomeImage}}]] will be rare, viewing transclusions will be rare, and using both with Firefox or Opera will rarer yet. Maybe Firefox and Opera will be fixed by the time we need it.
8 2011-06-15T00:19:40 <RogerHaase> The most annoying thing will be the popup blocker message from Firefox, but assuming the user says OK, it will happen only 1 time.
9 2011-06-15T00:21:01 <ThomasWaldmann> a link with an image is not that rare
10 2011-06-15T00:21:47 <ThomasWaldmann> e.g. buttons, logos linking to something
11 2011-06-15T00:25:13 <RogerHaase> Most people I know use shift-click. Only the MMB is a problem with FIrefox and Opera.
12 2011-06-15T00:26:16 <RogerHaase> Did you notice on unixpapa, that when you use the MMB to click on "click here to test", you get a new tab with javascript:void?
13 2011-06-15T00:26:47 <ThomasWaldmann> yup
14 2011-06-15T00:28:47 <RogerHaase> Do you see that kind of problem on other web pages when you use the MMB?
15 2011-06-15T00:29:17 <ThomasWaldmann> no, because i usually use it to open links in new tabs :)
16 2011-06-15T00:29:40 <ThomasWaldmann> and my thinpad has 3 touchpad buttons
17 2011-06-15T00:29:43 <ThomasWaldmann> +k
18 2011-06-15T00:29:52 <ThomasWaldmann> hmm, it's thin also :D
19 2011-06-15T00:31:14 <RogerHaase> But if you accidently click on a element with an onclick action instead of a real A tag, you should see this kind of a problem often.
20 2011-06-15T00:32:07 <RogerHaase> click/MMB click, or MMB non-click
21 2011-06-15T00:33:44 <ThomasWaldmann> i never noticed a problem
22 2011-06-15T00:33:51 <ThomasWaldmann> maybe i was lucky
23 2011-06-15T00:37:19 *** sinha
24 2011-06-15T00:37:52 <ThomasWaldmann> (me is not good at js and browser compatbility stuff, maybe some other people help better with this)
25 2011-06-15T00:39:13 <RogerHaase> OK, so what should I do next?
26 2011-06-15T00:40:30 <RogerHaase> I can look for examples where MMB fails with Firefox.
27 2011-06-15T00:41:46 <ThomasWaldmann> well, if it is really a bug in ff, a bug report would make sense
28 2011-06-15T00:42:52 <ThomasWaldmann> as they release rather often, it would not take centuries until everybody has a fixed one (only year[s])
29 2011-06-15T00:47:31 <RogerHaase> There are several bug reports in both Firefox and Opera related to MMB. I think the problem is there is not a standard related to the use of the MMB.
30 2011-06-15T00:50:00 <RogerHaase> gtg
31 2011-06-15T00:50:22 *** RogerHaase
32 2011-06-15T00:57:06 *** m4k3r
33 2011-06-15T05:58:48 *** MattMaker
34 2011-06-15T06:07:35 *** CIA-69
35 2011-06-15T06:14:15 *** CIA-72
36 2011-06-15T06:23:48 *** MattMaker
37 2011-06-15T07:16:52 *** m4k3r
38 2011-06-15T07:43:16 *** andialbrecht_
39 2011-06-15T07:45:49 *** MattMaker
40 2011-06-15T08:39:07 <dreimark> moin
41 2011-06-15T08:59:40 *** m4k3r
42 2011-06-15T09:05:23 <ThomasWaldmann> moin
43 2011-06-15T09:08:04 *** m4k3r
44 2011-06-15T09:08:52 *** MattMaker
45 2011-06-15T09:11:04 *** gwork
46 2011-06-15T09:11:42 <Marchael> moin
47 2011-06-15T09:17:33 <dreimark> moin
48 2011-06-15T09:18:50 *** gwork
49 2011-06-15T09:48:31 <ThomasWaldmann> Marchael: btw, how fast is your internet downstream?
50 2011-06-15T09:51:59 <Marchael> 2-3 mbits
51 2011-06-15T09:52:15 <Marchael> in university of course
52 2011-06-15T10:00:52 <ThomasWaldmann> i was just considering making a bigger xml file so you have more than 4 items to index
53 2011-06-15T10:11:52 <Marchael> ok
54 2011-06-15T10:12:08 <Marchael> how much is it?
55 2011-06-15T10:12:40 <Marchael> please, share this
56 2011-06-15T10:14:04 <ThomasWaldmann> i don't have it yet. but i have some 1.9 production wiki data.
57 2011-06-15T10:14:56 <ThomasWaldmann> is there a file size limit for your downloads?
58 2011-06-15T10:18:24 *** ebo^
59 2011-06-15T10:18:44 <Marchael> hm, if it >500 I can download it
60 2011-06-15T10:18:53 <Marchael> *500 Mb
61 2011-06-15T10:19:38 <ThomasWaldmann> > or < ?
62 2011-06-15T10:20:17 <Marchael> lesser than 500
63 2011-06-15T10:20:48 <Marchael> s/>/</ :D
64 2011-06-15T10:23:38 *** gwork
65 2011-06-15T11:00:46 *** greg_f
66 2011-06-15T11:19:42 *** sinha
67 2011-06-15T11:31:43 *** ronny
68 2011-06-15T11:37:07 <ThomasWaldmann> Marchael: you read the docs i mentioned?
69 2011-06-15T11:37:21 <Marchael> about bulk indexing?
70 2011-06-15T11:37:30 <ThomasWaldmann> yes
71 2011-06-15T11:37:36 * Marchael reading
72 2011-06-15T11:38:15 <Marchael> I found only incremental indexing
73 2011-06-15T11:38:37 <ThomasWaldmann> you'll need that also
74 2011-06-15T11:39:08 <ThomasWaldmann> see the scenarios we need to support (EP)
75 2011-06-15T11:41:05 <ThomasWaldmann> http://packages.python.org/Whoosh/batch.html
76 2011-06-15T11:42:43 <Marchael> bulk == batch ?
77 2011-06-15T11:43:11 <Marchael> I guess what we'll use this for all rev index
78 2011-06-15T11:43:21 <Marchael> *xes
79 2011-06-15T11:45:31 <ThomasWaldmann> yes, maybe even for both
80 2011-06-15T11:47:42 <Marchael> but how it'll working? Will we commit bunch of documents using time interval?
81 2011-06-15T11:48:25 <ThomasWaldmann> bulk, batch, full index build, it all means that you index many document at once
82 2011-06-15T11:49:30 <ThomasWaldmann> you just need to follow those docs i pointed you to
83 2011-06-15T11:50:17 <ThomasWaldmann> skip the stemming stuff there for now
84 2011-06-15T11:52:52 * Marchael looks on MultiSegmentWriter
85 2011-06-15T11:53:17 <Marchael> It'll be good for (re)building indexes in first time
86 2011-06-15T11:55:03 <ThomasWaldmann> yes
87 2011-06-15T11:57:53 <Marchael> but we need to know how much processor wiki admin wants to use for building indexes.
88 2011-06-15T11:58:00 <Marchael> *processors
89 2011-06-15T12:04:22 <ThomasWaldmann> app.cfg.index_procs
90 2011-06-15T12:09:16 <ThomasWaldmann> same for that limitmb
91 2011-06-15T12:11:21 <ThomasWaldmann> and after thinking a bit about it, i think the current index build algo can be improved
92 2011-06-15T12:12:29 <ThomasWaldmann> the latest revs index is much more important and much more used and much smaller than the all revs index
93 2011-06-15T12:15:05 <ThomasWaldmann> so, in case indexes need (re)building, it would be good to have a means to first build the latest revs index (should be rather quick, even for huge wikis), then already be able to use moin for 99% of use cases and then build all revs index (which can take many hours for huge wikis)
94 2011-06-15T12:20:35 <ThomasWaldmann> see EP also
95 2011-06-15T12:58:00 <Marchael> seems much faster
96 2011-06-15T12:58:16 <Marchael> at least for 5-10 seconds for my thin wiki :)
97 2011-06-15T12:59:40 <ThomasWaldmann> for 4 revs? :)
98 2011-06-15T13:00:36 <Marchael> yeah
99 2011-06-15T13:01:29 <Marchael> ThomasWaldmann: please, review http://codereview.appspot.com/4539114/patch/17002/17003
100 2011-06-15T13:17:13 *** MattMaker
101 2011-06-15T13:20:28 <ThomasWaldmann> review done
102 2011-06-15T13:20:33 * ThomasWaldmann needs to go
103 2011-06-15T13:22:07 * dreimark to and i have a question added
104 2011-06-15T13:25:38 <Marchael> dreimark: iirc pep-8 recomends to use empty line for dividing different packages
105 2011-06-15T13:25:51 <Marchael> I mean third part and local
106 2011-06-15T13:27:53 <Marchael> > procs and limitmb not intialized?
107 2011-06-15T13:27:53 <Marchael> dreimark: see default value in each option
108 2011-06-15T13:45:37 *** hardy89
109 2011-06-15T13:46:49 <hardy89> waldi: how do i use cProfile in my tests intead of using it in one single function?
110 2011-06-15T13:57:14 <waldi> the same way then profile. not sure how to use that with pytest
111 2011-06-15T14:00:05 <hardy89> waldi: i cud not find it in any of the tutorials.
112 2011-06-15T14:02:13 <waldi> i'm not sure what you try to do. please elaborate
113 2011-06-15T14:03:47 <hardy89> waldi: i'm trying to use python profilers for memory leak issue.
114 2011-06-15T14:04:06 <hardy89> waldi: all the tutorial explain how to use them for functions.
115 2011-06-15T14:04:18 <hardy89> how do i use them for pytest?
116 2011-06-15T14:06:52 * Marchael bbl
117 2011-06-15T14:07:04 *** Marchael
118 2011-06-15T14:10:00 <hardy89> waldi: even tried doing '$ py.test -m cProfile testname.py' which did not work
119 2011-06-15T14:10:12 <hardy89> that had to happen though
120 2011-06-15T14:10:22 <waldi> hardy89: i still don't get it. profiling lists function calls and time spent. memory leaks are usually buried objects somewhere
121 2011-06-15T14:11:24 <hardy89> waldi: It not only gives the total running time, but also times each function separately, and tells you how many times each function was called, making it easy to determine where you should make optimizations.
122 2011-06-15T14:16:46 <waldi> but this is not about memory leaks
123 2011-06-15T14:21:51 <hardy89> waldi: i tried using Dozer. Spent some time on it. But it did not help.
124 2011-06-15T14:22:06 <hardy89> waldi: so next thing i hopped on to was profilers.
125 2011-06-15T14:23:45 <hardy89> waldi: cos every search on memory leaks in python leads to either Dozer or Profilers
126 2011-06-15T14:29:30 <waldi> i would start with the gc module
127 2011-06-15T14:33:16 <hardy89> all right. let me check and work on it. thanks waldi
128 2011-06-15T14:52:52 *** hardy89
129 2011-06-15T14:53:25 *** hardy89
130 2011-06-15T14:54:17 *** Marchael
131 2011-06-15T14:54:52 *** hardy89
132 2011-06-15T14:56:51 *** hardy89
133 2011-06-15T14:58:56 *** pkumar
134 2011-06-15T15:01:38 *** brunomartin
135 2011-06-15T15:02:14 <Marchael> ThomasWaldmann: may be use 2 different functions for building partial indexes, and one as we have now.
136 2011-06-15T15:07:24 <waldi> i think there is a pytest meeting scheduled for _now_. anyone here?
137 2011-06-15T15:07:47 <pkumar> hi
138 2011-06-15T15:08:27 <hardy89> waldi: yes
139 2011-06-15T15:10:59 <pkumar> so, who goes first?
140 2011-06-15T15:11:15 <pkumar> waldi : ^
141 2011-06-15T15:15:06 <waldi> this is about coordination between you. postpone the rest for the general meeting
142 2011-06-15T15:15:31 <pkumar> ok
143 2011-06-15T15:15:51 <hardy89> pkumar: ok i'll do it
144 2011-06-15T15:16:18 <hardy89> LAST & BLOCKERS
145 2011-06-15T15:16:19 <hardy89> * For memory leaks problem, Dozer was used. Both with default config and even with some added changes. RAM usage still climb up to more than 2.5 GB.
146 2011-06-15T15:16:19 <hardy89> So Dozer finally did not work and ended up taking too many days.
147 2011-06-15T15:16:19 <hardy89> * Then tried memory profiling which again turned out to be useful only for functions and not for objects(which is the main cause of memory leaks).
148 2011-06-15T15:17:21 <hardy89> waldi: ^
149 2011-06-15T15:17:37 <pkumar> I think we need to report thi in general meeting
150 2011-06-15T15:17:41 <pkumar> this*
151 2011-06-15T15:17:50 <hardy89> waldi: we've anyhow discussed bout this.
152 2011-06-15T15:19:26 <hardy89> waldi: now i need to know how gc module will help in tracing object allocation and any kind of unwanted recursion if any?
153 2011-06-15T15:22:55 <dreimark> hardy89: this is the coordination meeting now
154 2011-06-15T15:23:19 <dreimark> between you waldi, pkumar and xorAxAx
155 2011-06-15T15:23:45 <hardy89> ok
156 2011-06-15T15:27:08 <waldi> there is some plan for next week. can this plan work? how do you want to devide the work?
157 2011-06-15T15:28:12 <pkumar> I think, 1/2 days more would be needed to work on the current part of project
158 2011-06-15T15:28:21 <pkumar> hardy: ?
159 2011-06-15T15:28:38 <pkumar> hardy89 : ^
160 2011-06-15T15:28:56 <hardy89> yes. i guess come up with something useful within next 2 days
161 2011-06-15T15:29:23 <pkumar> hardy89 : ok, EP should also be updated.
162 2011-06-15T15:29:48 <hardy89> waldi: what if we discuss this part on EP
163 2011-06-15T15:29:51 <hardy89> pkumar: yeah
164 2011-06-15T15:31:36 <pkumar> hardy89 : two more days should be enough I guess?
165 2011-06-15T15:31:45 <hardy89> pkumar: yes i guess
166 2011-06-15T15:31:58 <hardy89> pkumar: talk on EP itself
167 2011-06-15T15:32:56 <waldi> please talk here
168 2011-06-15T15:33:04 <hardy89> waldi: ok
169 2011-06-15T15:33:07 <hardy89> pkumar: ^
170 2011-06-15T15:33:50 <hardy89> pkumar: i guess there is too much of work from 18th to 25th june which has been listed.
171 2011-06-15T15:34:26 <hardy89> pkumar: we can shift not so important things like tutorial etc for the next week, thus focussing on main issues first.
172 2011-06-15T15:34:59 <pkumar> hardy89 : MoinMoin tests basically uses assert statements and it is quite compatible with pytest2
173 2011-06-15T15:35:46 <hardy89> pkumar: yes it is. then?
174 2011-06-15T15:37:36 <pkumar> it means, there is nothing much to do in the parts as per line 203 and 201
175 2011-06-15T15:38:37 <hardy89> pkumar: well yes, cos individual tests are anyhow very basic which are too far from being complex.
176 2011-06-15T15:39:00 *** Marchael
177 2011-06-15T15:39:05 *** Marchael
178 2011-06-15T15:39:19 <pkumar> since, most of them are assert statements
179 2011-06-15T15:39:21 <hardy89> pkumar: line 209 is a major issue, YES cos that is very much evident in modules specially storage and datastruct
180 2011-06-15T15:40:27 <hardy89> brb
181 2011-06-15T15:41:40 <Marchael> I have problem with internet connection via awful wi-fi, so if I'll disconnect, please don't think what I'm gone
182 2011-06-15T15:42:13 <pkumar> yeah, there are some failure and errors in storage and datastruct and a few in convertor modules
183 2011-06-15T15:44:44 *** ebo^
184 2011-06-15T15:49:48 <pkumar> xorAxAx : hi, I got this result after running the tests using fs2. http://paste.pocoo.org/show/408152/
185 2011-06-15T15:50:31 <pkumar> xorAxAx : but not OSError
186 2011-06-15T15:50:52 <pkumar> ThomasWaldmann : ^
187 2011-06-15T15:52:20 <pkumar> xorAxAx : I'm not sure if it is due to changes in conftest.py
188 2011-06-15T15:53:14 <pkumar> or it is pytest2
189 2011-06-15T15:54:55 <waldi> i have to catch something to eat
190 2011-06-15T15:58:50 <xorAxAx> re
191 2011-06-15T16:00:14 <xorAxAx> pkumar: so?
192 2011-06-15T16:01:54 *** m4k3r
193 2011-06-15T16:01:56 <pkumar> xorAxAx : I mean what are the conditions when this error occurs or is it solved?
194 2011-06-15T16:02:14 <pkumar> xorAxAx : since I'm not gettin those errors
195 2011-06-15T16:02:19 * ThomasWaldmann is here now
196 2011-06-15T16:02:34 <ThomasWaldmann> meeting time! :)
197 2011-06-15T16:02:47 <xorAxAx> pkumar: are you getting them or not?
198 2011-06-15T16:02:59 <xorAxAx> pkumar: ah, you mean oserrors? no idea
199 2011-06-15T16:03:08 <ThomasWaldmann> xorAxAx: pkumar: you want to begin with the formal part?
200 2011-06-15T16:03:15 <pkumar> xorAxAx : no. no errors like OSError
201 2011-06-15T16:03:23 <xorAxAx> pkumar: please begin :)
202 2011-06-15T16:03:29 <pkumar> ok
203 2011-06-15T16:03:47 <pkumar> Here it goes:
204 2011-06-15T16:04:00 <pkumar> The task was to work on issues like file leak issue when run with fs2
205 2011-06-15T16:04:12 <pkumar> used tools like objgraph to better understand the code execution
206 2011-06-15T16:04:24 <pkumar> ran tests twice, got some errors and failure(need to be fixed), but not 'OSError'
207 2011-06-15T16:04:40 <pkumar> It would be great if somebody else could run the tests using pytest2 test code and confirm/rectify the results.
208 2011-06-15T16:04:53 <pkumar> working on other issue(faster tests), should take 1/2 days more (hopefully one).
209 2011-06-15T16:05:02 <pkumar> for blockers:
210 2011-06-15T16:05:14 <pkumar> not such yet. will work on the issue regading 'setup for wsgi app from scratch' for a while myself and then ask on IRC if nothing works out.
211 2011-06-15T16:05:47 <pkumar> my report ends here.
212 2011-06-15T16:05:55 <pkumar> any questions?
213 2011-06-15T16:06:13 <xorAxAx> yes, didnt you say you wanted to work on logging?
214 2011-06-15T16:06:31 <pkumar> yeah, I worked last week
215 2011-06-15T16:06:45 <xorAxAx> ok
216 2011-06-15T16:06:52 <xorAxAx> then why didnt you say so?
217 2011-06-15T16:06:56 <pkumar> and we don't have any timing log outputs anymore on test failures
218 2011-06-15T16:07:17 <xorAxAx> your last pushed commit is 7/8 days ago, you didnt make any debugging progress
219 2011-06-15T16:07:22 <ThomasWaldmann> pkumar: did you watch file handle usage while running tests?
220 2011-06-15T16:07:35 <xorAxAx> sorry, but it looks to me as if you are not working fulltime on this project
221 2011-06-15T16:07:36 <ThomasWaldmann> like with lsof -p <pid of process>
222 2011-06-15T16:08:08 <pkumar> this week was for fixing issues
223 2011-06-15T16:08:38 *** ronny
224 2011-06-15T16:08:41 <pkumar> and this takes quite a lot of time and we really have less changessets to make in those caases
225 2011-06-15T16:08:48 <pkumar> ThomasWaldmann : no
226 2011-06-15T16:09:14 <xorAxAx> so you havent fixed a single issue yet? ok
227 2011-06-15T16:09:28 <pkumar> yes, logging one
228 2011-06-15T16:09:36 <ThomasWaldmann> then please do so :) it will only fail if it goes over the resource limit, but it can also leak maybe without going over the limit
229 2011-06-15T16:10:37 <pkumar> the issue regarding 'faster tests' will take 1/2 days more i guess
230 2011-06-15T16:10:39 <ThomasWaldmann> and about fulltime work: I (in general, not limited to pkumar) have also that impression
231 2011-06-15T16:10:41 <pkumar> xorAxAx : ^
232 2011-06-15T16:11:29 <ThomasWaldmann> (I see that while debugging, there is less "commit output", but I also miss communications, questions, etc.)
233 2011-06-15T16:11:30 *** grzywacz
234 2011-06-15T16:12:20 <ThomasWaldmann> ok, anything else for pkumar?
235 2011-06-15T16:12:30 <ThomasWaldmann> hardy89: waldi: you are there?
236 2011-06-15T16:12:35 <hardy89> ThomasWaldmann: yes
237 2011-06-15T16:12:56 <pkumar> ok, from now onwards I'll be more on IRC
238 2011-06-15T16:13:09 <xorAxAx> nice
239 2011-06-15T16:13:14 <pkumar> :)
240 2011-06-15T16:13:19 <ThomasWaldmann> pkumar: good :)
241 2011-06-15T16:13:32 <ThomasWaldmann> ok, hardy89 you're next
242 2011-06-15T16:13:48 <hardy89> ok
243 2011-06-15T16:14:02 <ThomasWaldmann> waldi: ^^vv
244 2011-06-15T16:14:03 <hardy89> LAST & BLOCKERS
245 2011-06-15T16:14:04 <hardy89> * For memory leaks problem, Dozer was used. Both with default config and even with some added changes. RAM usage still climb up to more than 2.5 GB.
246 2011-06-15T16:14:04 <hardy89> So Dozer finally did not work and ended up taking too many days.
247 2011-06-15T16:14:04 <hardy89> * Then tried memory profiling which again turned out to be useful only for functions and not for objects(which is the main cause of memory leaks).
248 2011-06-15T16:14:33 <hardy89> NEXT
249 2011-06-15T16:14:33 <hardy89> * Now trying garbage collector module as told by waldi. Need some more time for fixing memory leaks while I work with gc module.
250 2011-06-15T16:14:33 <hardy89> * Along with this, tests which skip, show error, or fail will be fixed. Specially tests in modules Storage and Datastruct.
251 2011-06-15T16:15:00 <ThomasWaldmann> what did you use for "Then tried memory profiling"?
252 2011-06-15T16:15:01 <waldi> re
253 2011-06-15T16:15:23 <hardy89> ThomasWaldmann: i used cProfile for mem leak problem
254 2011-06-15T16:15:53 <ThomasWaldmann> isn't that just for finding cpu/time expensive stuff?
255 2011-06-15T16:16:21 <hardy89> ThomasWaldmann: i called by calling it from within the code, and even invoked it while running script
256 2011-06-15T16:16:47 <hardy89> ThomasWaldmann: yeah. that + It not only gives the total running time, but also times each function separately, and tells you how many times each function was called, making it easy to determine where you should make optimizations.
257 2011-06-15T16:17:07 <hardy89> ThomasWaldmann: but the point is, its for functions and our main problem is objects here.
258 2011-06-15T16:17:15 <ThomasWaldmann> yes, that just wasn't the point :)
259 2011-06-15T16:17:38 <hardy89> luckily it was corrected by waldi on time before it wud have been another waste of time like Dozer
260 2011-06-15T16:17:43 <ThomasWaldmann> btw, at europython there will be a debugging talk, could be interesting for such topics
261 2011-06-15T16:18:20 <hardy89> ThomasWaldmann: i saw a link regarding it somewhere in stackoverflow while reading docs
262 2011-06-15T16:18:29 <hardy89> has to be helpful
263 2011-06-15T16:18:55 <ThomasWaldmann> in general, if you are not coming forward on that goal, I suggest to not waste multiple weeks on it, but rather work on more productive stuff
264 2011-06-15T16:19:27 <ThomasWaldmann> (that stuff isn't easy)
265 2011-06-15T16:19:37 <hardy89> ThomasWaldmann: i understand that, but i feel i want a bit more time with gc module. i'll do it :)
266 2011-06-15T16:19:45 <hardy89> ThomasWaldmann: not at all, but still.
267 2011-06-15T16:20:27 <ThomasWaldmann> btw (sorry, i didn't read backlog yet), is the "porting" itself finished?
268 2011-06-15T16:20:33 <hardy89> ThomasWaldmann: in case, (if) i feel its going out of proportion for me, i'll definitely report
269 2011-06-15T16:20:51 <hardy89> ThomasWaldmann: porting, to be honest, did not much/any work.
270 2011-06-15T16:21:07 <hardy89> cos even without making any changes , all the tests worked on pytest
271 2011-06-15T16:21:30 <hardy89> ThomasWaldmann: that is the only reason we ended up just changing py/py.test to pytest
272 2011-06-15T16:21:46 <hardy89> just to avoid confusion for someone who reads it for the first time
273 2011-06-15T16:21:52 <ThomasWaldmann> that means every test is running, with same error / success count as before, no pytest2 related issues?
274 2011-06-15T16:22:06 <hardy89> ThomasWaldmann: no
275 2011-06-15T16:22:42 <pkumar> I case of porting behavior of tests regarding generator functions have been changed.
276 2011-06-15T16:23:21 <ThomasWaldmann> that means?
277 2011-06-15T16:23:28 <hardy89> ThomasWaldmann: there are tests which have error/fail/skip etc, but they need some sanitization and refactoring which is yet to come
278 2011-06-15T16:24:18 <ThomasWaldmann> well, sounds like porting is not yet finished
279 2011-06-15T16:24:48 <pkumar> now according to the code, we iterate through the tests in case of generator functions
280 2011-06-15T16:24:48 <ThomasWaldmann> xorAxAx: waldi: don't we want to first finish porting before doing other stuff?
281 2011-06-15T16:25:05 <xorAxAx> ThomasWaldmann: i think that can run parallely
282 2011-06-15T16:25:21 <xorAxAx> but they should complete that task soon
283 2011-06-15T16:25:23 <hardy89> ThomasWaldmann: no the count hasn't changed and we were asked not to make changes in code if they fail/have error/ skip etc while porting
284 2011-06-15T16:26:03 <ThomasWaldmann> i didn't ask for test fixes unrelated to pytest2.
285 2011-06-15T16:26:11 <pkumar> Also, we have less messy MoinTestFunction class(tanks to Ronny) which was a part of porting
286 2011-06-15T16:26:21 <pkumar> thanks*
287 2011-06-15T16:26:56 <ThomasWaldmann> ok
288 2011-06-15T16:27:22 <hardy89> ThomasWaldmann: the point is, the major part is done. the minor tweaks can be done parallely
289 2011-06-15T16:27:28 <pkumar> further more, it would be good if we go according to current plan and port if we get any more porting ideas
290 2011-06-15T16:27:31 <ThomasWaldmann> waldi: anything else for hardy89?
291 2011-06-15T16:27:39 <waldi> no
292 2011-06-15T16:28:02 <pkumar> all the current tests are compatible with pytest2
293 2011-06-15T16:28:04 <hardy89> and yes, even i'll be seen more on #moin-dev from now on.
294 2011-06-15T16:28:05 <ThomasWaldmann> hardy89: can you also try to work fulltime AND be present and active here on IRC?
295 2011-06-15T16:28:18 <ThomasWaldmann> :)
296 2011-06-15T16:28:33 <ThomasWaldmann> ok, dreimark and sinha are you here?
297 2011-06-15T16:28:36 <sinha> yes
298 2011-06-15T16:28:39 * dreimark yes
299 2011-06-15T16:28:39 <pkumar> so not much porting work is required/available
300 2011-06-15T16:28:56 <sinha> POSTING IT
301 2011-06-15T16:28:58 <sinha> LAST: i) Completed bookmark functionality for global history results
302 2011-06-15T16:28:59 <sinha> changeset: https://bitbucket.org/sinha/moin2/changeset/3076415fdaba
303 2011-06-15T16:28:59 <sinha> ii) Minor modification in grouping of items also, now changes done for a item on the same day is only shown.
304 2011-06-15T16:28:59 <sinha> changeset: https://bitbucket.org/sinha/moin2/changeset/71988ea0f53e
305 2011-06-15T16:28:59 <sinha> CURRENT: Working on paging feature
306 2011-06-15T16:28:59 <sinha> BLOCKER: i) Somehow we can set offset for next page, but how to get the offset for previous one
307 2011-06-15T16:28:59 <ThomasWaldmann> ok, sinha your stage :)
308 2011-06-15T16:28:59 <sinha> ii) During grouping for a item i use to suppress the changes which occured in previous days, but as we move on to next page we are rejecting the revisions which falls under offset, so the item revisions which occured on previous days are shown here.
309 2011-06-15T16:28:59 <sinha> For example: Page 1) http://moinmo.in/AkashSinha/Gsoc2011Diary/2011-06-15?action=AttachFile&do=view&target=pagin1.png
310 2011-06-15T16:29:00 <sinha> Page 2) http://moinmo.in/AkashSinha/Gsoc2011Diary/2011-06-15?action=AttachFile&do=view&target=paging2.png
311 2011-06-15T16:29:00 <sinha> Item June31 has changes in both 4th june as well as 3rd june, in page 1 it is coming under 4th june, but its first change is coming under 3rd june in page2.
312 2011-06-15T16:29:00 <hardy89> ThomasWaldmann: i started working full time from 11th only as indicated by diary entry. I'm finally over with my college so this is the only thing for me now.
313 2011-06-15T16:29:24 * dreimark strange chars
314 2011-06-15T16:30:30 <waldi> they are replacement characters
315 2011-06-15T16:30:53 * sinha depends on the irc client
316 2011-06-15T16:31:09 <ThomasWaldmann> sinha: i can't look at all your links right now, but the general means to give state to next page is to put it in url qs
317 2011-06-15T16:31:29 <sinha> but for previous page
318 2011-06-15T16:31:51 <sinha> we are not following any standard rule here, what we calculate in current page we set it as offset for next page
319 2011-06-15T16:31:53 <ThomasWaldmann> you can do same thing. the other way is to be able to compute it somehow.
320 2011-06-15T16:32:09 <ThomasWaldmann> not easy if it is not exact and depends on quite some stuff.
321 2011-06-15T16:32:37 <sinha> and about the blocker ii) ?
322 2011-06-15T16:33:17 <ThomasWaldmann> i don't think i understand the issue
323 2011-06-15T16:33:26 <ThomasWaldmann> the algorithm is the same, for any page
324 2011-06-15T16:33:49 <sinha> yes but if u get time just look into those links
325 2011-06-15T16:34:09 <sinha> as during item grouping, i am suppressing the changes occured in previous days
326 2011-06-15T16:34:31 <dreimark> sinha: looks like you calculate them for another "bookmark" like time
327 2011-06-15T16:34:37 <dreimark> so they appear twice
328 2011-06-15T16:35:27 <sinha> dreimark: https://bitbucket.org/sinha/moin2/src/71988ea0f53e/MoinMoin/apps/frontend/views.py#cl-556
329 2011-06-15T16:35:35 <ronny> 3~re
330 2011-06-15T16:35:59 <ThomasWaldmann> ehrm
331 2011-06-15T16:36:04 <sinha> in this code if the latest rev of item doesnt comes, then its earlier revision gets selected
332 2011-06-15T16:36:19 <ronny> im a bit busy atm, i'll be there around 1800-1900
333 2011-06-15T16:36:31 <sinha> yes in the absence of bookmark, it happens
334 2011-06-15T16:36:59 <ThomasWaldmann> maybe we need slightly different way of doing things depending on whether there is a bookmark being used or not
335 2011-06-15T16:37:42 <sinha> okay
336 2011-06-15T16:38:03 <dreimark> yes
337 2011-06-15T16:38:06 <ThomasWaldmann> if a bookmark is used, it would be enough to show most recent change - the diff would always diff bookmark up to that change
338 2011-06-15T16:38:28 <ThomasWaldmann> if no bookmark is used, I guess you should not suppress stuff, but show it on the previous day(s)
339 2011-06-15T16:39:26 <sinha> you mean no item grouping, in case of no bookmark ?
340 2011-06-15T16:39:30 <ThomasWaldmann> (but even that mode will confuse some people, so maybe we should cluster changes only within one day for everybody)
341 2011-06-15T16:39:30 *** brunomartin
342 2011-06-15T16:39:47 <sinha> ah ok
343 2011-06-15T16:39:51 <dreimark> that won't be good
344 2011-06-15T16:40:22 <dreimark> it should not result in having an entry on each day the item was changed
345 2011-06-15T16:40:25 <ThomasWaldmann> might lead to more work when processing rc, though
346 2011-06-15T16:40:44 <ThomasWaldmann> viewing same diff parts multiple times
347 2011-06-15T16:42:04 <ThomasWaldmann> btw, one solution to the paging related problem is to compute ALL the stuff all the time, but just show the part that should be on current page
348 2011-06-15T16:42:14 <ThomasWaldmann> not very efficient, but would work :)
349 2011-06-15T16:42:52 <ThomasWaldmann> ALL meaning "up to bookmark" or up to N days, with N defaulting to some sane value
350 2011-06-15T16:43:27 <ThomasWaldmann> except that, is there any other question for sinha?
351 2011-06-15T16:43:30 <dreimark> also the result is more recent if something changes between two clicks
352 2011-06-15T16:43:49 <ThomasWaldmann> (we can discuss details later, maybe on EP)
353 2011-06-15T16:44:18 <sinha> yeah sure
354 2011-06-15T16:44:33 <ThomasWaldmann> Marchael: you're there?
355 2011-06-15T16:44:38 <Marchael> This week I working on fixing tests for whoosh schemas and writing flask-script for rebuilding indexes.
356 2011-06-15T16:44:38 <Marchael> Schemas has been tested and now working, but ThomasWaldmann not sure about fate WhooshIndex class, so may be we'll refuse
357 2011-06-15T16:44:38 <Marchael> from it in future, but now it using for opening index dir and containing referers to open all and latest rev indexes.
358 2011-06-15T16:44:38 <Marchael> Next week I'll done with building indexes and try to do some part of integrating indexing to moin.
359 2011-06-15T16:44:38 <Marchael> blockers:
360 2011-06-15T16:44:38 <Marchael> Last 2 days I was busy because I unluckily bought defective bike and spent more time to turn it back
361 2011-06-15T16:44:38 <Marchael> Now I have no blockers
362 2011-06-15T16:44:51 <ThomasWaldmann> that was the long answer :)
363 2011-06-15T16:45:14 <ThomasWaldmann> but nice that stuff is being prepared for meeting.
364 2011-06-15T16:46:06 * Marchael just go direct way :)
365 2011-06-15T16:46:10 <ThomasWaldmann> btw, it would be nice to get a more opinions about whether we want to have such a WhooshIndex class or not.
366 2011-06-15T16:47:00 <Marchael> ThomasWaldmann: I'm not quite understand your point, why WhooshIndex is bad?
367 2011-06-15T16:47:18 <Marchael> or may be bad
368 2011-06-15T16:47:24 <ThomasWaldmann> well, what does it give you yet?
369 2011-06-15T16:48:10 <Marchael> we can use one WhooshIndex object for wiki wotthout code duplication in different places
370 2011-06-15T16:48:21 <Marchael> s/wotthout/without/
371 2011-06-15T16:48:42 <ThomasWaldmann> https://bitbucket.org/marchael/moin-2.0/src/d5c124222533/MoinMoin/search/indexing.py this is what we talk about
372 2011-06-15T16:49:51 <ThomasWaldmann> well, the most code there just deals with the case when there is no index.
373 2011-06-15T16:50:26 <ThomasWaldmann> we could just require that there is an index when moin is started (and that admin has to use some cmdline too to make it).
374 2011-06-15T16:51:32 <ThomasWaldmann> if that would be the case, the only thing needed from there are 1 or 2 calls to open_dir().
375 2011-06-15T16:51:45 <ThomasWaldmann> which, btw often are NOT needed both, but just one of them.
376 2011-06-15T16:52:01 <Marchael> but I'm not quite understand, will we open indexes each time when commit happens?
377 2011-06-15T16:52:11 <dreimark> yes
378 2011-06-15T16:52:16 <ThomasWaldmann> for that case, you'll need both, of course
379 2011-06-15T16:52:17 <dreimark> you need to add new content
380 2011-06-15T16:52:28 <ThomasWaldmann> but for searching, it is often just the index with latest stuff
381 2011-06-15T16:52:42 <ThomasWaldmann> except when the wiki user says "i want to search all history"
382 2011-06-15T16:53:34 <ThomasWaldmann> but we can delay that potential refactoring until later and keep it now as is
383 2011-06-15T16:54:36 <ThomasWaldmann> btw, try to work 5 full days a week, otherwise you might run out of time
384 2011-06-15T16:54:58 <ThomasWaldmann> progress is too slow currently
385 2011-06-15T16:55:37 *** Marchael
386 2011-06-15T16:56:04 <ThomasWaldmann> for now, we'll get indexing and searching working for the metadata, but there is still lots to do with data, and search results presentation
387 2011-06-15T16:56:24 *** Marchael
388 2011-06-15T16:56:40 <Marchael> sorry, I have problem with internet connection
389 2011-06-15T16:56:52 <ThomasWaldmann> last thing you got was?
390 2011-06-15T16:57:03 <Marchael> (20:51:33) ThomasWaldmann: but we can delay that potential refactoring until later and keep it now as is
391 2011-06-15T16:57:13 <ThomasWaldmann> (16:54) < ThomasWal> btw, try to work 5 full days a week, otherwise you might run out of time
392 2011-06-15T16:57:16 <ThomasWaldmann> (16:55) < ThomasWal> progress is too slow currently
393 2011-06-15T16:57:18 <ThomasWaldmann> (16:56) < ThomasWal> for now, we'll get indexing and searching working for the metadata, but there is still lots to do with data, and search results presentation
394 2011-06-15T16:57:21 <pkumar> bbl
395 2011-06-15T16:57:21 *** pkumar
396 2011-06-15T16:57:56 <Marchael> yes, I know.
397 2011-06-15T16:58:18 <Marchael> I spent to many times to stupid mistakes :(
398 2011-06-15T16:58:38 <ThomasWaldmann> that'll get better with more experience :)
399 2011-06-15T16:58:57 <ThomasWaldmann> learn from mistakes, own and other peoples :D
400 2011-06-15T16:59:00 <Marchael> I also try work at night to communicate with you
401 2011-06-15T16:59:15 *** ronny
402 2011-06-15T16:59:31 <Marchael> ok :)
403 2011-06-15T16:59:45 <ThomasWaldmann> well, I am also there at daytimes, but I have work to do then
404 2011-06-15T17:00:04 <ThomasWaldmann> ok, questions for Marchael?
405 2011-06-15T17:01:22 <ThomasWaldmann> ok, looks like not. thanks to everybody for being here and being prepared this time. :)
406 2011-06-15T17:02:03 <Marchael> thank you ThomasWaldmann
407 2011-06-15T17:02:30 <hardy89> ThomasWaldmann: thank you sarge :P
408 2011-06-15T17:02:46 <ThomasWaldmann> Marchael: btw, you import entoken, but you do not need it (in indexing.py).
409 2011-06-15T17:03:02 <ThomasWaldmann> hardy89: no military here
410 2011-06-15T17:03:09 <hardy89> ThomasWaldmann: :/
411 2011-06-15T17:03:11 <hardy89> ok
412 2011-06-15T17:03:12 <ThomasWaldmann> except General Failure
413 2011-06-15T17:03:24 <ThomasWaldmann> and Major Mistake
414 2011-06-15T17:03:45 <Marchael> :D
415 2011-06-15T17:03:47 <hardy89> and Colonel(kernel) etc
416 2011-06-15T17:03:54 <hardy89> ThomasWaldmann: :)
417 2011-06-15T17:05:17 <ThomasWaldmann> Marchael: you are currently working on index update?
418 2011-06-15T17:05:53 <ThomasWaldmann> whoosh docs have some sample code for that kind of operation. it is filesystem there, but you could use quite similar code for moin revisions.
419 2011-06-15T17:06:17 <Marchael> I think what we can do 3 functions. One for building both, one for all revs, and one for latest.
420 2011-06-15T17:08:02 <hardy89> bbl
421 2011-06-15T17:08:16 *** hardy89
422 2011-06-15T17:09:20 <ThomasWaldmann> Marchael: yes
423 2011-06-15T17:10:04 <ThomasWaldmann> but that's not what I meant, I was one step further, making sure that an existing index is completely uptodate
424 2011-06-15T17:13:29 <Marchael> so first we should check what all docs from backend are existing in index
425 2011-06-15T17:14:24 *** m4k3r
426 2011-06-15T17:18:22 <Marchael> I'll have short bike riding
427 2011-06-15T17:18:25 <Marchael> get back soon
428 2011-06-15T17:18:28 *** Marchael
429 2011-06-15T17:22:39 *** xorAxAx
430 2011-06-15T17:26:19 *** TheSheep
431 2011-06-15T17:26:24 *** TheSheep
432 2011-06-15T17:26:33 *** xorAxAx
433 2011-06-15T17:30:09 *** andialbrecht
434 2011-06-15T18:04:18 *** ronny
435 2011-06-15T18:29:27 *** hardy89
436 2011-06-15T18:36:35 *** Marchael
437 2011-06-15T18:50:19 *** brunomartin
438 2011-06-15T18:53:17 *** hardy89
439 2011-06-15T18:59:21 *** hardy89
440 2011-06-15T19:02:47 <hardy89> ThomasWaldmann: can we change the time of our weekly meeting. Even pkumar wants to change it cos of his hostel timings.
441 2011-06-15T19:02:54 *** RogerHaase
442 2011-06-15T19:04:03 <ThomasWaldmann> hardy89: no problem from my side, make a doodle, post it on the wiki gsoc page, talk to people to take part in finding a new time
443 2011-06-15T19:04:51 <hardy89> ThomasWaldmann: sounds good :)
444 2011-06-15T19:06:29 *** Marchael
445 2011-06-15T19:42:05 <hardy89> ThomasWaldmann: done :)
446 2011-06-15T19:45:24 *** greg_f
447 2011-06-15T19:47:10 *** hardy89
448 2011-06-15T19:55:01 <sinha> ThomasWaldmann, dreimark: need to discuss regarding paging at EP.
449 2011-06-15T20:45:36 *** m4k3r
450 2011-06-15T21:06:49 *** m4k3r
451 2011-06-15T22:35:48 *** CIA-72
452 2011-06-15T22:49:41 *** CIA-70
453 2011-06-15T22:58:22 *** grzywacz
454 2011-06-15T23:06:01 *** ronny
455 2011-06-15T23:06:50 <ThomasWaldmann> re
456 2011-06-15T23:07:32 <ThomasWaldmann> sinha: if you start discussing instead of telling we need to, you save 1 roundtrip time
457 2011-06-15T23:08:05 <sinha> ThomasWaldmann: yeah, did with dreimark
458 2011-06-15T23:08:23 <ThomasWaldmann> ah, ok :)
459 2011-06-15T23:08:36 <sinha> just have problem in finding previous link, trying to figure out that .
460 2011-06-15T23:08:44 <sinha> otherwise he told the other things
461 2011-06-15T23:11:25 <dreimark> http://www.doodle.com/76d9saphni46xwq9
462 2011-06-15T23:11:40 <dreimark> just found a new doodle on our gsoc page
463 2011-06-15T23:12:31 <dreimark> for which day is that?
464 2011-06-15T23:13:38 <dreimark> gn
465 2011-06-15T23:29:24 *** ronny
466 2011-06-15T23:38:32 <RogerHaase> ThomasWaldmann: for the [[http://moinmo.in|{{../Logo}}]] issue, I am trying to wrap the A tag rather than the IMG tag, see here: http://paste.pocoo.org/show/408455/
467 2011-06-15T23:40:32 <RogerHaase> But doing that in html_out looks impossible, and doing it within moinwiki_in seems difficult as well.
468 2011-06-15T23:41:07 <RogerHaase> So I was going to try cutting and pasting the dom with javascript. Have any better ideas?
