1 2011-07-22T00:24:05 <brunomartin> I think I found a bug... when loading xml file, if you have metadata like meta_data = [["abc"],["asdf"," fsa"]] (list of lists), it raises: http://paste.pocoo.org/show/443890/
2 2011-07-22T00:25:54 <brunomartin> I made a workarroud that make it loads anyway, that solves my problem.... I just want to confirm if its a bug...
3 2011-07-22T00:29:07 <dreimark> brunomartin: file this to the issue tracker on bb
4 2011-07-22T00:29:22 <dreimark> ans also add the workaround
5 2011-07-22T00:29:44 <dreimark> s/ans/and/
6 2011-07-22T00:30:12 <brunomartin> yeah.... I'm gonna do it now... ;-)
7 2011-07-22T00:30:24 <dreimark> thx
8 2011-07-22T01:23:57 *** raignarok
9 2011-07-22T05:15:28 *** MattMaker_
10 2011-07-22T05:19:15 *** MattMaker
11 2011-07-22T05:19:38 *** MattMaker_
12 2011-07-22T05:57:59 <Marchael> moin
13 2011-07-22T05:58:33 <Marchael> ThomasWaldmann: when you'll get some free time please, assist me with this problem.
14 2011-07-22T08:05:32 *** pkumar
15 2011-07-22T09:32:06 <Marchael> also, whoosh 2.0 released
16 2011-07-22T09:32:21 <Marchael> I'll up to date soon
17 2011-07-22T09:43:13 *** sinha
18 2011-07-22T10:18:41 *** raignarok
19 2011-07-22T10:18:58 <ThomasWaldmann> moin
20 2011-07-22T10:27:25 <ThomasWaldmann> Marchael: review done
21 2011-07-22T10:33:43 *** pkumar
22 2011-07-22T10:34:40 *** sinha
23 2011-07-22T10:51:53 *** greg_f
24 2011-07-22T11:12:10 * Marchael answered
25 2011-07-22T11:13:23 <Marchael> I use that loop in 145 for catching error
26 2011-07-22T11:13:34 <Marchael> *145-152
27 2011-07-22T11:14:31 <Marchael> but if use normal result.highlight("field") I take the same error/
28 2011-07-22T11:27:58 <ThomasWaldmann> ok, answered there.
29 2011-07-22T11:28:18 <ThomasWaldmann> and iirc, it was .highlights() with s
30 2011-07-22T11:28:25 <Marchael> рьб нуы
31 2011-07-22T11:28:28 <Marchael> hm, yes
32 2011-07-22T11:28:36 <Marchael> but it isn't working
33 2011-07-22T11:29:00 <Marchael> But I noticed what for fields with standard types highlights() working
34 2011-07-22T11:29:08 <Marchael> I mean ID,TEXT
35 2011-07-22T11:30:20 <ThomasWaldmann> then maybe try to do some content result rendering and hightlighting today
36 2011-07-22T11:31:12 <ThomasWaldmann> if a quick glance into whoosh source (see your traceback for pointers) doesn't help to identify what's going wrong and whether this is a whoosh bug or your bug, I can have a look later today
37 2011-07-22T11:32:34 <ThomasWaldmann> btw, about whoosh 2.0: be careful with that upgrade, they merged py3 support code. so better test wether all is working as expected.
38 2011-07-22T11:32:47 <ThomasWaldmann> gtg/brb
39 2011-07-22T12:33:15 *** Marchael
40 2011-07-22T14:07:41 *** MattMaker
41 2011-07-22T14:15:23 *** Marchael
42 2011-07-22T14:28:50 *** Marchael
43 2011-07-22T14:38:16 *** Marchael
44 2011-07-22T15:15:04 <Marchael> ThomasWaldmann: I added short variant http://codereview.appspot.com/4802048 but error is the same http://paste.pocoo.org/show/444240/
45 2011-07-22T15:58:38 <ThomasWaldmann> well, you have teh source :)
46 2011-07-22T15:59:05 <Marchael> heh
47 2011-07-22T16:11:41 *** pkumar
48 2011-07-22T16:17:00 <dreimark> so only sinha is missing :(
49 2011-07-22T16:23:57 *** brunomartin
50 2011-07-22T16:24:14 <pkumar> xorAxAx : hey, I think that that the test in test_backends.test_new_item_create_revision() is incorrect. item.rollback does not removes the item but just effects rev._fs_file_meta and rev._fs_file_data
51 2011-07-22T16:24:55 <pkumar> it neither works for test_bakends_fs2 nor test_backends_fs
52 2011-07-22T16:25:55 <xorAxAx> well, depends on the spec :)
53 2011-07-22T16:26:13 <xorAxAx> i think the test is correct
54 2011-07-22T16:26:28 <xorAxAx> because items without revisions may not exist as far as i know
55 2011-07-22T16:29:02 <pkumar> but here in tests, item is created, then a revision for it and item is not removed on rollback
56 2011-07-22T16:29:26 <pkumar> xorAxAx : ^
57 2011-07-22T16:31:37 <pkumar> I traced the tests and everything seems to be working fine, except handling of item part in the test
58 2011-07-22T16:32:52 <xorAxAx> well, then the implementation is probably incorrect
59 2011-07-22T16:35:36 <pkumar> xorAxAx : I guess so. Also, if we use item.change_metadata() for an item and don't publish it, item shouldn't exist right?
60 2011-07-22T16:36:55 <pkumar> _change_item_metadata*
61 2011-07-22T16:37:20 <xorAxAx> do you have a test doing this that you can show me?
62 2011-07-22T16:38:54 *** brunomartin
63 2011-07-22T16:39:41 <pkumar> xorAxAx : yeah, test_item_metadata_without_publish() in storage/test_backends
64 2011-07-22T16:40:26 <pkumar> xorAxAx : should I paste it?
65 2011-07-22T16:40:26 <xorAxAx> yep, shouldnt exist
66 2011-07-22T16:40:29 <xorAxAx> no
67 2011-07-22T16:41:44 <pkumar> xorAxAx : that fails too due to same reason(handling of item in test)
68 2011-07-22T16:41:57 <xorAxAx> yeah
69 2011-07-22T16:43:31 <pkumar> xorAxAx : one more thing, in test_mixed_commit_metadata2() change_metadata would end up having a lock on the item?
70 2011-07-22T16:43:47 <xorAxAx> pkumar: no idea
71 2011-07-22T16:45:13 <pkumar> xorAxAx : storage.backends.FS2Backend._change_item_metadata says something like that, isn't it?
72 2011-07-22T16:45:53 <xorAxAx> would make sense somehow at least
73 2011-07-22T16:49:56 <pkumar> xorAxAx : actually in the test test_backend_fs2, http://paste.pocoo.org/show/444299/ line 28,
74 2011-07-22T16:50:59 <pkumar> so for this particular test lock needs to be released , that's what I was just confirming
75 2011-07-22T16:51:59 <pkumar> and it works when lock is released
76 2011-07-22T16:52:15 <pkumar> for that test
77 2011-07-22T16:52:37 <pkumar> i.e test_mixed_commit_metadata2()
78 2011-07-22T17:04:09 <xorAxAx> what is your question?
79 2011-07-22T17:09:26 <pkumar> xorAxAx : as item.change_metadata() in test_backends would put a lock on item, to pass the test in above mentioned line 28, I need to release the lock. so I was just confirming if I was correct for the lock part
80 2011-07-22T17:10:03 <xorAxAx> well, the test is badly written if the assumption is taht every test cleans up its locks
81 2011-07-22T17:12:11 <pkumar> seems that tests with item.change_metadata() are not cleaning up its locks
82 2011-07-22T17:12:17 <pkumar> xorAxAx : ^
83 2011-07-22T17:12:47 <xorAxAx> they should call rollback in finally, right?
84 2011-07-22T17:14:01 <pkumar> xorAxAx : where exactly are you referring to?
85 2011-07-22T17:14:35 <xorAxAx> to the tests that do not clean up
86 2011-07-22T17:18:05 <pkumar> xorAxAx : I think that the assumption is that rollback would remove the newly created item but it doesn't. Also, those with item.change_metadata() are having cleaning up problems
87 2011-07-22T17:18:28 <xorAxAx> you are repeating yourself
88 2011-07-22T17:18:36 <xorAxAx> the latter tests - how should they clean up`?
89 2011-07-22T17:19:40 <xorAxAx> which method do they need to call?
90 2011-07-22T17:21:15 <pkumar> xorAxAx : for tests with lock problems 'item._item._fs_metadata_lock.release()' helps.
91 2011-07-22T17:21:36 <xorAxAx> pkumar: thats not a public API
92 2011-07-22T17:24:31 <pkumar> xorAxAx : so, does it mean we can't call it in tests?
93 2011-07-22T17:24:41 <xorAxAx> indeed
94 2011-07-22T17:24:57 <xorAxAx> wouldnt make sense because it is a layering violation
95 2011-07-22T17:25:51 <pkumar> xorAxAx : can you please explain : layering violation?
96 2011-07-22T17:26:03 <xorAxAx> software consists of layers
97 2011-07-22T17:26:30 <xorAxAx> and if layer A calls layer C without asking B to do that, it is a layering violation
98 2011-07-22T17:26:43 <xorAxAx> because the layering concept of the layers is violated
99 2011-07-22T17:27:17 <xorAxAx> the fs/fs2 locks are broken automatically after a timeout, right?
100 2011-07-22T17:28:08 <pkumar> xorAxAx : yeah, I guess so, 30 it seems
101 2011-07-22T17:29:40 <pkumar> xorAxAx : problem is lock still exists though it should not
102 2011-07-22T17:30:02 <pkumar> in case of tests
103 2011-07-22T17:30:35 <xorAxAx> well, there should be a method that undoes the change_metadata step
104 2011-07-22T17:30:41 <xorAxAx> is rollback that method?
105 2011-07-22T17:31:02 <pkumar> lemme check
106 2011-07-22T17:31:13 *** sinha
107 2011-07-22T17:37:25 <pkumar> xorAxAx : rollback is possible only after a revision is made. tests like test_mixed_commit_metadata2 don't have a revision
108 2011-07-22T17:37:55 <pkumar> else it shows NoneType' object has no attribute '_fs_file_meta
109 2011-07-22T17:40:58 <sinha> ThomasWaldmann: dreimark: pls see this http://moinmo.in/AkashSinha/Gsoc2011Diary/2011-07-22
110 2011-07-22T17:41:23 *** MattMaker_
111 2011-07-22T17:45:27 *** MattMaker
112 2011-07-22T17:45:28 *** MattMaker_
113 2011-07-22T17:46:05 <pkumar> xorAxAx : how about item.publish_metadata().
114 2011-07-22T17:49:48 <pkumar> xorAxAx : seems like item.publish_metadata() works
115 2011-07-22T17:58:21 *** sinha
116 2011-07-22T18:04:59 <pkumar> bbl
117 2011-07-22T18:05:04 *** pkumar
118 2011-07-22T18:36:11 <ThomasWaldmann> btw, pkumar shouldn't waste too much time on itemlevel metadata.
119 2011-07-22T18:36:33 <ThomasWaldmann> that's a feature that might become obsolete, once we have all indexing in place
120 2011-07-22T18:37:10 <ThomasWaldmann> ehrm, sinha could share a bit more time with us
121 2011-07-22T18:48:53 <dreimark> re
122 2011-07-22T18:50:42 <dreimark> I guess since he is at home he has bad internet connection
123 2011-07-22T19:14:34 *** pkumar
124 2011-07-22T19:15:47 <pkumar> ThomasWaldmann : ok, # read the chat logs
125 2011-07-22T19:50:59 *** sinha
126 2011-07-22T19:53:54 *** greg_f
127 2011-07-22T19:56:43 *** pkumar
128 2011-07-22T20:18:53 <dreimark> http://blogs.sitepointstatic.com/examples/tech/speech-input/index.html
129 2011-07-22T20:19:06 <dreimark> thats an example for speech input
130 2011-07-22T20:20:57 <ThomasWaldmann> sinha: instead of that green expaning dl button, do you think that the whole background of that window could have that dropzone functionality?
131 2011-07-22T20:21:10 <ThomasWaldmann> ehrm s/dl/upload/
132 2011-07-22T20:23:47 <sinha> ThomasWaldmann: actually i have seen that drag/drop working on my browser just once (so far), i am using Firefox 4.0 but still that thing is not working properly, so i dont know if i could experiment with that, i may write the scripts but dont know how good they will be .
133 2011-07-22T20:23:49 <dreimark> i would prefer the whole window too
134 2011-07-22T20:24:05 <sinha> I will try upgrading my browser, then look into that
135 2011-07-22T20:24:24 <dreimark> http://weston.ruter.net/projects/google-tts/ for speech
136 2011-07-22T20:24:54 <dreimark> output
137 2011-07-22T20:30:16 <dreimark> sinha: for me it works with 3.6.18
138 2011-07-22T20:30:23 <sinha> dreimark: reg including that file uploader to forms.html, there are few javascripts/stylesheet also related with that uploader
139 2011-07-22T20:31:43 <sinha> dreimark: i am not saying it doesnt works, i have seen it working, but very very few times
140 2011-07-22T20:32:31 <dreimark> we have to find out why
141 2011-07-22T20:33:05 <sinha> may be some sort of support issues
142 2011-07-22T20:33:06 <dreimark> what is the reason to point on javascripts/stylesheet for moving it to a macro of forms?
143 2011-07-22T20:33:29 <sinha> one of the script is very much custom to this index page
144 2011-07-22T20:33:40 <sinha> one has to rewrite it for every other page
145 2011-07-22T20:37:29 <dreimark> why, the uploading mechanism without showing the recent uploads should be possible to use on an other place
146 2011-07-22T20:39:19 <dreimark> sinha: ^
147 2011-07-22T20:39:50 <sinha> after every upload a callback is called, which is implemented in same same script
148 2011-07-22T20:40:03 <sinha> in that jfu.js it has views for every action
149 2011-07-22T20:45:28 <dreimark> yes, but why did this matter?
150 2011-07-22T20:49:41 <sinha> i mean, the current callback has elements (div) of this index page, but that element wont be present on other pages, so when it will try to execute that it may end up by encountering null ones. what i am saying is i will include all the scripts in that forms.html but will keep this one here, and for other plave we may need to do some small-small changes
151 2011-07-22T20:50:06 <sinha> s/plave/place
152 2011-07-22T20:51:42 <dreimark> ok
153 2011-07-22T20:56:30 <dreimark> top1 looks good
154 2011-07-22T20:59:19 <sinha> oh you seen it
155 2011-07-22T20:59:40 <sinha> dreimark: but the items are getting mixed together, i mean below "Recently uploaded.." all are together, i am thinking of adding a line below the Recently uploaded items
156 2011-07-22T21:00:51 <dreimark> just try, so that it looks nice
157 2011-07-22T21:02:10 <sinha> ok
158 2011-07-22T21:26:50 *** sinha
159 2011-07-22T21:27:49 <dreimark> http://craigsworks.com/projects/qtip2/download
160 2011-07-22T21:28:01 <dreimark> this may be able to handle all this previews
161 2011-07-22T21:30:27 <dreimark> http://craigsworks.com/projects/qtip/demos/content/thumbnail
162 2011-07-22T22:00:08 *** MattMaker
163 2011-07-22T22:34:11 <ThomasWaldmann> re
164 2011-07-22T22:50:54 <ThomasWaldmann> Marchael: how's the outcome of your highlighting research?
165 2011-07-22T23:21:57 <CIA-114> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 295:72009982fc8d 2.0/MoinMoin/converter/ (__init__.py text_out.py): add a dom -> text/plain output converter, use it to create indexable content
166 2011-07-22T23:22:09 <ThomasWaldmann> Marchael: ^^ easiest converter ever :)
167 2011-07-22T23:25:46 *** raignarok
168