1 2011-07-01T00:25:59 *** m4k3r
2 2011-07-01T00:27:45 *** RogerHaase
3 2011-07-01T00:56:29 *** Marchael
4 2011-07-01T04:17:52 *** Marchael
5 2011-07-01T04:19:06 <Marchael> moin
6 2011-07-01T05:21:43 <Marchael> ThomasWaldmann: I noticed what whoosh can't find document named like u'Trash/doc (2011-06-26 00:49:11 UTC)'
7 2011-07-01T05:23:43 <Marchael> btw, if item get trashed, whoosh can't find and remove revisions because they have different names
8 2011-07-01T05:31:17 <Marchael> s/different/too complicated/
9 2011-07-01T06:41:37 *** m4k3r
10 2011-07-01T06:46:07 <Marchael> ThomasWaldmann: also see ep 183
11 2011-07-01T06:57:25 <CIA-45> pkumar <contactprashantat@gmail.com> * 6b926426e3d2 r300 pytest2/MoinMoin/util/_tests/test_crypto.py: added test to check the stamp value in valid_token in util/crypto.py
12 2011-07-01T06:57:26 <CIA-45> Reimar Bauer <rb.proj AT googlemail DOT com> * 187a36c04c59 r301 pytest2/MoinMoin/storage/backends/ (fs19.py fs2.py): sqlalchemy.exceptions was removed in 0.7 of sqlalchemy. Since at least 0.6 the import for the core exception module is "from sqlalchemy import exc".
13 2011-07-01T06:57:27 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 7a77ae08161e r302 pytest2/MoinMoin/items/ (_tests/test_Item.py __init__.py): (log message trimmed)
14 2011-07-01T06:57:27 <CIA-45> items: cleanup contenttype handling in modify/_save, see below
15 2011-07-01T06:57:27 <CIA-45> contenttype in query string -> force this content type (put it into metadata)
16 2011-07-01T06:57:27 <CIA-45> If _save() does not find CONTENTTYPE in the new metadata, it uses:
17 2011-07-01T06:57:27 <CIA-45> 1. current content type (previous revision)
18 2011-07-01T06:57:27 <CIA-45> 2. the guessed content type (see below)
19 2011-07-01T06:57:28 <CIA-45> 3. application/octet-stream
20 2011-07-01T06:57:28 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * ec37a42f6fd2 r303 pytest2/MoinMoin/ (templates/modify_binary.html apps/frontend/views.py): item modify ui: remove cancel button (user can just click anywhere else than on SAVE)
21 2011-07-01T06:57:29 <CIA-45> Roger Haase <crosseyedpenguin@yahoo.com> * 58fc3bb9cee5 r304 pytest2/MoinMoin/themes/modernized/static/css/common.css: remove unused classes badinterwiki and moin-badinterwiki
22 2011-07-01T06:57:35 <CIA-45> RogerHaase * 144b6aa9516d r305 pytest2/MoinMoin/themes/modernized/static/ (img/white_clouds.png css/common.css img/white-clouds.jpg): replace white_clouds.png file with smaller white-clouds.jpg
23 2011-07-01T06:57:35 <CIA-45> RogerHaase * a74e93f21623 r306 pytest2/MoinMoin/converter/ (5 files in 2 dirs): correct rendering of interwiki and mailto links
24 2011-07-01T06:57:35 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * f464adcc3a51 r307 pytest2/MoinMoin/items/__init__.py: fix saving no data if no previous revision exists
25 2011-07-01T06:57:35 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * e93157b8b0bd r308 pytest2/MoinMoin/ (6 files in 3 dirs):
26 2011-07-01T06:57:35 <CIA-45> refactor modify views to use flatland, see details below
27 2011-07-01T06:57:35 <CIA-45> code is more OO now, the Item instance now handles modify view AND
28 2011-07-01T06:57:35 <CIA-45> post handling in do_modify() - moving this was needed so a flatland Form
29 2011-07-01T06:57:36 <CIA-45> subclass could be used that depends on the Item type.
30 2011-07-01T06:57:47 <CIA-45> moved item_modified event triggering to _save().
31 2011-07-01T06:57:47 <CIA-45> templates: use flatland generator to generate most form fields
32 2011-07-01T06:57:47 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 69c0d9f6981d r309 pytest2/MoinMoin/items/__init__.py: modify: put real base rev no into the hidden form field
33 2011-07-01T06:57:47 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * a11656ba1eab r310 pytest2/MoinMoin/ (templates/modify_applet.html util/forms.py items/__init__.py): modify views: use flatland also for the file upload form field
34 2011-07-01T06:57:47 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 52ca5dc1ab15 r311 pytest2/MoinMoin/items/__init__.py: add a JSON validator, use it to validate meta_text
35 2011-07-01T06:57:47 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 727e03c59dbf r312 pytest2/MoinMoin/ (apps/frontend/views.py items/__init__.py): simplify form validation check, fix indentation
36 2011-07-01T06:57:48 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 5028c04190a9 r313 pytest2/MoinMoin/converter/html_out.py: html_out converter: remove unused code from moinpage_object that crashed on transclusions with URLs without path. Fixes #30.
37 2011-07-01T06:58:46 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 1a1dab306214 r314 pytest2/MoinMoin/ (themes/__init__.py apps/frontend/views.py items/__init__.py): move the flatland form html generator to the global environment
38 2011-07-01T06:58:46 <CIA-45> Thomas Waldmann <tw AT waldmann-edv DOT de> * 20c18292bd02 r315 pytest2/MoinMoin/ (4 files in 3 dirs): (log message trimmed)
39 2011-07-01T06:58:47 <CIA-45> use flatland and some html5 features for search form, details see below
40 2011-07-01T06:58:47 <CIA-45> generate the top-right search form using flatland.
41 2011-07-01T06:58:47 <CIA-45> currently, search form is only rendered for "show" view.
42 2011-07-01T06:58:47 <CIA-45> use a Validator for the search form, currently it rejects if search query
43 2011-07-01T06:58:48 <CIA-45> (39 lines omitted)
44 2011-07-01T07:22:57 <Marchael> ThomasWaldmann: hm, if you could fix this https://bitbucket.org/thomaswaldmann/moin-2.0/issue/33/uuid-is-changing-for-default-documents then I could use uuid instead of name, and we can avoid troubles with item names
45 2011-07-01T08:15:47 <dreimark> Marchael: if this blocks you currently can you just live with that problem?
46 2011-07-01T08:16:55 <Marchael> dreimark: I already told what problem I have with name of trashed item
47 2011-07-01T08:19:51 <Marchael> dreimark: if don't solve this problem then update script starts duplicate trashed item
48 2011-07-01T08:22:39 <Marchael> I could use re for that, but it was great spike
49 2011-07-01T08:31:14 <dreimark> Marchael: i know, my question was if you can live with that "mistake" for now and continue
50 2011-07-01T08:31:33 <dreimark> you can add a ToDo into the code
51 2011-07-01T08:36:31 <Marchael> ah, ok
52 2011-07-01T08:42:40 * dreimark current problem with flatland http://paste.pocoo.org/show/425613/
53 2011-07-01T09:15:56 <ThomasWaldmann> moin
54 2011-07-01T09:16:59 <Marchael> hi ThomasWaldmann
55 2011-07-01T09:18:03 <ThomasWaldmann> dreimark: can you try again to get flatland? getting that url with browser worked.
56 2011-07-01T09:19:20 <dreimark> ThomasWaldmann: yes second run works
57 2011-07-01T09:19:21 <ThomasWaldmann> Marchael: there are quite some problems with uuid still. so i can try fixing issue 33, but no promises.
58 2011-07-01T09:20:27 <dreimark> brb
59 2011-07-01T09:26:04 <ThomasWaldmann> Marchael: btw, you did not specify exactly what problem you have / why you have that problem regarding names and trashing
60 2011-07-01T09:32:25 <Marchael> ThomasWaldmann: hm, how can I improve it?
61 2011-07-01T09:32:39 <Marchael> btw, please make review http://codereview.appspot.com/4662071/
62 2011-07-01T09:46:38 *** m4k3r
63 2011-07-01T09:47:44 *** m4k3r
64 2011-07-01T09:49:20 *** sinha
65 2011-07-01T09:49:52 <sinha> dreimark: blockers at diary page, http://moinmo.in/AkashSinha/Gsoc2011Diary/2011-06-30
66 2011-07-01T09:58:06 <ThomasWaldmann> Marchael: review done
67 2011-07-01T09:58:31 * Marchael looks
68 2011-07-01T10:06:20 <Marchael> ThomasWaldmann: see ep
69 2011-07-01T10:13:45 *** mkerrin
70 2011-07-01T10:52:21 <ThomasWaldmann> Marchael: < GrahamDum> Technically you can't rely on requests being completed before process shutdown. Various situations exist where process will be forcibly shutdown.
71 2011-07-01T10:53:19 <ThomasWaldmann> graham is author of mod-wsgi
72 2011-07-01T10:57:00 <Marchael> ok, but we already decided to use normal writer
73 2011-07-01T10:57:59 <Marchael> ThomasWaldmann: also, could you give a hint how avoid while True ?
74 2011-07-01T10:59:18 <dreimark> sinha: answered
75 2011-07-01T10:59:45 * sinha lloking
76 2011-07-01T11:00:03 * sinha looking
77 2011-07-01T11:05:23 <sinha> dreimark: i am preferring GET, because we can directly feed the url and also we can use the permalink to elsewhere
78 2011-07-01T11:06:07 <sinha> and regarding this +index/startswith/letter, can i use the same function in views.py for this and for url +index ?
79 2011-07-01T11:15:42 <ThomasWaldmann> Marchael: first know what to do, then code. you still have no idea what to do, thus you do not need to think about coding.
80 2011-07-01T11:16:11 * ThomasWaldmann is thinking about whether atexit could help us
81 2011-07-01T11:17:43 <Marchael> ThomasWaldmann: first what I think was qeue. When commit happens, we give item an id and put it to qeue
82 2011-07-01T11:18:23 <Marchael> then we wainting until first item id become current item id
83 2011-07-01T11:18:44 <Marchael> s/qeue/queue/
84 2011-07-01T11:23:27 <ThomasWaldmann> you could not yet explain that idea, so i still think it doesn't make any sense.
85 2011-07-01T11:25:04 <Marchael> at least you may told why it's pontless
86 2011-07-01T11:25:17 <Marchael> *told me
87 2011-07-01T11:30:05 <ThomasWaldmann> it's just random code, it doesn't solve any problem
88 2011-07-01T11:31:37 <ThomasWaldmann> and if you can't explain what the problem is that you are trying to solve and in what way your code solves that problem, it maybe is just pointless
89 2011-07-01T11:32:13 <ThomasWaldmann> to bring this topic forwards, I'll just tell you now:
90 2011-07-01T11:32:58 <ThomasWaldmann> one problem you have is concurrency on thread and process level. something else might have the writelock on the index and keep it for some unknown time.
91 2011-07-01T11:32:59 <dreimark> sinha: you can separate it. no +index/startswith/letter means show all#
92 2011-07-01T11:33:07 <dreimark> with letter show only that
93 2011-07-01T11:33:45 <ThomasWaldmann> the other problem you have, is that the process that runs your code might get terminated at any time.
94 2011-07-01T11:34:09 <ThomasWaldmann> and the goal is to put data into the index and usually not lose index data
95 2011-07-01T11:34:35 <dreimark> sinha: and think about if you can do it as +index/letter if there are never other parameters needed
96 2011-07-01T11:35:37 <ThomasWaldmann> usually means that i don't think 100.0% safety is needed, but we should try to get as near to that as possible
97 2011-07-01T11:35:53 <Marchael> ok
98 2011-07-01T11:37:22 *** greg_f
99 2011-07-01T11:37:24 <ThomasWaldmann> atexit.register maybe helps, maybe just try: finally: helps, if something else keeps the lock forever, you don't want to hang forever also
100 2011-07-01T11:37:53 <ThomasWaldmann> you don't want to consume 100% cpu while waiting for a lock
101 2011-07-01T11:39:35 <ThomasWaldmann> AsyncWriter somehow sounds nice, but behaviour on process termination is unclear (if a process terminates, all threads end with it)
102 2011-07-01T11:55:04 <ThomasWaldmann> Marchael: for now, just use AsyncWriter
103 2011-07-01T11:55:33 <Marchael> hm, ok
104 2011-07-01T12:33:01 <Marchael> ThomasWaldmann: I noticed what after update&move, indexes become look like http://paste.pocoo.org/show/425739/ So we have here indexes before build and after
105 2011-07-01T12:33:25 <Marchael> I guess what we should remove all stuff from index_dir before moving indexes there
106 2011-07-01T12:46:50 <ThomasWaldmann> i don't understand what you mean
107 2011-07-01T12:48:07 <ThomasWaldmann> you said "update", which is iirc in-place. did you mean "build from scratch in index_dir_tmp"?
108 2011-07-01T12:48:46 <ThomasWaldmann> Marchael:
109 2011-07-01T12:48:58 <ThomasWaldmann> gtg/bbl
110 2011-07-01T12:49:03 <Marchael> hm
111 2011-07-01T12:49:53 <Marchael> I mean situation when admin first make --for ... --action update && --for ... --action move
112 2011-07-01T12:50:34 <Marchael> but old indexes with greater number will stays and whoosh perhaps will think what they are latest
113 2011-07-01T13:15:46 <dreimark> Marchael: human failure
114 2011-07-01T13:17:05 <Marchael> am, but supposed what better remove all stuff in index_dir before moving indexes
115 2011-07-01T13:17:30 <Marchael> *I supposed
116 2011-07-01T13:19:01 <Marchael> also, if wiki admin moving new index to index_dir he absolutely sure what he'll use new indexes
117 2011-07-01T13:59:20 *** m4k3r
118 2011-07-01T13:59:32 * Marchael gtg/bbl
119 2011-07-01T13:59:39 *** m4k3r
120 2011-07-01T14:48:49 *** pkumar
121 2011-07-01T14:54:23 *** m4k3r
122 2011-07-01T15:02:14 *** sinha
123 2011-07-01T15:02:29 *** sinha
124 2011-07-01T15:03:03 <sinha> dreimark: i am trying to do this, @frontend.route('/+index/<startswith:letter>'), i am getting a exception in werkzeug routing, where i have to add any entry for this map ?
125 2011-07-01T15:24:41 <ThomasWaldmann> sinha: it sounds like you maybe want /+index/<prefix>
126 2011-07-01T15:25:14 <sinha> ThomasWaldmann: yes, but we already have that +index/itemname
127 2011-07-01T15:25:24 <sinha> so it was going into that function
128 2011-07-01T15:25:38 <ThomasWaldmann> ah, right
129 2011-07-01T15:25:42 <sinha> hence i did +index/startswith/prefix
130 2011-07-01T15:26:05 <ThomasWaldmann> no, that's wrong
131 2011-07-01T15:26:17 <sinha> so how to do this ?
132 2011-07-01T15:26:32 <ThomasWaldmann> maybe you want a qs argument for that
133 2011-07-01T15:26:50 <sinha> okay, qs argument also works
134 2011-07-01T15:26:55 <sinha> but i am getting this error
135 2011-07-01T15:26:58 <ThomasWaldmann> so url looks like /+index?prefix=A
136 2011-07-01T15:27:21 <sinha> yes i tried to generate this url by doing this url_for('frontend.index', startsWith=letter)
137 2011-07-01T15:27:25 <ThomasWaldmann> then it'll only list items starting with that prefix
138 2011-07-01T15:27:32 <sinha> but i was getting Build Error
139 2011-07-01T15:27:44 <sinha> But when i tried url_for('frontend.index', item_name='', startsWith=letter), it worked
140 2011-07-01T15:27:51 <sinha> i mean the url generation
141 2011-07-01T15:27:59 <ThomasWaldmann> it'll work
142 2011-07-01T15:28:24 <ThomasWaldmann> it the argument is not part of the routing spec pattern, it'll just put it into qs
143 2011-07-01T15:28:29 <sinha> ThomasWaldmann: this is the traceback url_for('frontend.index', item_name='', startsWith=letter)
144 2011-07-01T15:28:40 <sinha> ah, this one http://paste.pocoo.org/show/425630/
145 2011-07-01T15:28:48 <sinha> ThomasWaldmann: ^
146 2011-07-01T15:28:53 <ThomasWaldmann> but you have to make some thoughts about whether you want to really mix this into /+index
147 2011-07-01T15:29:18 <ThomasWaldmann> that rule is wrong anyway
148 2011-07-01T15:29:30 <sinha> i guess it is showing index only, not doing anything else
149 2011-07-01T15:29:50 <sinha> so how do i generate this url +index?startswith=letter
150 2011-07-01T15:30:12 <ThomasWaldmann> so, first answer the question "shall this and if yes how shall this integrate with current /+index behaviour?"
151 2011-07-01T15:31:15 <ThomasWaldmann> because we currently have 2 times "index" with different stuff, you'ld add a third thing to it
152 2011-07-01T15:31:23 <ThomasWaldmann> so first plan, then code
153 2011-07-01T15:31:59 <sinha> i am using a regex to search for the item having this letter in start, this is the modified flat_index http://paste.pocoo.org/show/425829/
154 2011-07-01T15:32:43 <sinha> but this is just a parameter passing, so it could be done there
155 2011-07-01T15:34:54 <ThomasWaldmann> looks like indenting fail
156 2011-07-01T15:35:02 <sinha> ThomasWaldmann: yeah cleaning is yet to be done
157 2011-07-01T15:35:44 <ThomasWaldmann> and also pep8 fails, because of startsWith
158 2011-07-01T15:36:02 <ThomasWaldmann> and you don't import re in the middle of a function
159 2011-07-01T15:36:25 <sinha> i have seen such import in the function above
160 2011-07-01T15:36:38 <ThomasWaldmann> item_re is no good
161 2011-07-01T15:36:55 <sinha> as of now i am trying to make things run, i will do things perfect when all are in place
162 2011-07-01T15:37:23 <ThomasWaldmann> return does not have () afterwards
163 2011-07-01T15:38:14 <ThomasWaldmann> and if you just want to check for startswith, that function is MUCH too complicated
164 2011-07-01T15:39:01 <Marchael> ThomasWaldmann: please review http://codereview.appspot.com/4662071/diff2/8001:12001/MoinMoin/script/maint/index.py
165 2011-07-01T15:39:20 <dreimark> sinha: this pep8 stuff was recommended already in the codereview
166 2011-07-01T15:39:57 <sinha> dreimark: yes i had seen the comment out there, just stuck in doing other changes
167 2011-07-01T15:40:02 <sinha> will do it now
168 2011-07-01T15:44:35 <sinha> dreimark: need hep in this url generation, {{ url_for('frontend.index', startswith=letter) }} isn't working
169 2011-07-01T15:56:19 <dreimark> i know i readed backlog and ThomasWaldmann gave already hints
170 2011-07-01T15:57:29 <dreimark> bbl
171 2011-07-01T16:08:43 <Marchael> ThomasWaldmann: MoinMoin.storage.backend.indexing.ItemIndex.update_item() updates metadata only for item_revision or whole item?
172 2011-07-01T16:22:11 <pkumar> bbl
173 2011-07-01T16:22:14 *** pkumar
174 2011-07-01T16:23:44 <ThomasWaldmann> Marchael: there is a docstring
175 2011-07-01T16:24:30 <sinha> ThomasWaldmann: so should i use new url other than +index to show index for a prefix
176 2011-07-01T16:25:28 <ThomasWaldmann> sinha: make a good plan. if you have a nice idea about how to integrate stuff, then do that. if you find out it should be rather separate, do that.
177 2011-07-01T16:26:06 <ThomasWaldmann> keep usability in mind, in the end we need something that somehow behaves like some end user would except it and would be able to cope with
178 2011-07-01T16:26:41 <sinha> ThomasWaldmann: yes, i guess +index?prefix=a, makes sense to user end
179 2011-07-01T16:26:54 <ThomasWaldmann> Marchael: what you maybe didn't realize yet is that items may have item-level metadata
180 2011-07-01T16:27:11 <ThomasWaldmann> also, items have revisions and revisions have metadata and data
181 2011-07-01T16:27:29 <ThomasWaldmann> item-level metadata is not used much, though
182 2011-07-01T16:27:47 <Marchael> hm, what is item-level metadata?
183 2011-07-01T16:28:43 <ThomasWaldmann> something we maybe won't need in the end :) but in general, it is some information that applies to whole item and is not revisioned
184 2011-07-01T16:29:08 <ThomasWaldmann> e.g. the current name or the always-same uuid
185 2011-07-01T16:29:37 <ThomasWaldmann> but you maybe want to rather care about revision metadata
186 2011-07-01T16:30:51 <Marchael> do we really need this item-level stuff? All and latest revision indexes can store enough for searching
187 2011-07-01T16:31:46 <Marchael> ah, ok
188 2011-07-01T16:34:04 <Marchael> ThomasWaldmann: btw, which functions of ItemIndex using outside this object?
189 2011-07-01T16:34:52 <ThomasWaldmann> yeah, i am still undecided. at a quite early time in design this was also some kind of "easy to access latest item info place" kind of a cache, so that one does not have to determine
190 2011-07-01T16:35:03 <ThomasWaldmann> the latest item revision first and then look there for infos
191 2011-07-01T16:35:06 <Marchael> I see what some of functions can be safely remove from this class
192 2011-07-01T16:35:15 <ThomasWaldmann> but with indexing, this is kind of superceded
193 2011-07-01T16:35:52 <ThomasWaldmann> i'ld suggest you do not remove anything from indexing for now, you ADD your whoosh stuff
194 2011-07-01T16:36:23 <ThomasWaldmann> and we remove the sqla stuff as soon as we are sure whoosh has replaced it
195 2011-07-01T16:37:57 <Marchael> could you review http://codereview.appspot.com/4662071/diff2/8001:12001/MoinMoin/script/maint/index.py
196 2011-07-01T16:39:19 <ThomasWaldmann> Marchael: about the tuple/lists comment of last review:
197 2011-07-01T16:39:24 <Marchael> I know about 23-25, but can't to find appropriate name for that
198 2011-07-01T16:39:36 <Marchael> *names
199 2011-07-01T16:40:21 <Marchael> but this tuples are convenient
200 2011-07-01T16:40:25 <ThomasWaldmann> tuples are typically used for some kind of structured data. e.g. the 1st tuple element is x, the 2nd tuple element is y - so elements at different positions mean somehow something different
201 2011-07-01T16:41:14 <ThomasWaldmann> lists typically are just lists of same kind of things, like pets = ["dog", "cat", ...]
202 2011-07-01T16:41:33 <ThomasWaldmann> so each element is same kind (a pet in this case)
203 2011-07-01T16:41:58 <Marchael> but first tuple element is indexname and second is schema
204 2011-07-01T16:42:21 <ThomasWaldmann> sometimes tuples are also used as immutable lists, though. matter of taste whether one considers this as an abuse of tuples or not.
205 2011-07-01T16:43:17 <ThomasWaldmann> except for line25, there you have a tuple that should be a list rather and is not of same structure than 23 24
206 2011-07-01T16:44:41 <ThomasWaldmann> Marchael: looks definitely better now
207 2011-07-01T16:45:04 <Marchael> ok, in 25 I should ues list of course
208 2011-07-01T16:45:34 <Marchael> so 23,24 ok?
209 2011-07-01T16:45:51 <ThomasWaldmann> yes
210 2011-07-01T16:46:14 <Marchael> ok, done. May I commit?
211 2011-07-01T16:46:25 <ThomasWaldmann> does it work? :)
212 2011-07-01T16:46:46 <Marchael> you could test it at any time :)
213 2011-07-01T16:47:08 <Marchael> but for me yes, althought still have issues
214 2011-07-01T16:47:21 <Marchael> like problem with trashed items
215 2011-07-01T16:47:38 <ThomasWaldmann> ehrm, what was the point of shutil.move again for moving a single file?
216 2011-07-01T16:48:15 <Marchael> I try don't touch existing indexes
217 2011-07-01T16:48:26 <Marchael> but seems like I have problem with it
218 2011-07-01T16:48:35 <ThomasWaldmann> i mean opposed to os.rename
219 2011-07-01T16:48:51 <Marchael> + os.mkdir
220 2011-07-01T16:49:02 <Marchael> because script will fail without this dir
221 2011-07-01T16:49:22 <Marchael> index_dir_tmp
222 2011-07-01T16:49:34 <ThomasWaldmann> ah, ok
223 2011-07-01T16:50:02 <Marchael> also, after commit you could clone my repo and try to test rebuilding with big dump
224 2011-07-01T16:50:29 <ThomasWaldmann> i likely will be rather busy the rest of the day and also at the weekend
225 2011-07-01T16:50:34 <Marchael> my internet connection limited
226 2011-07-01T16:51:11 <Marchael> ok, you could share this dump and try to download it when I'll have good opportunities
227 2011-07-01T16:51:23 <Marchael> *and I try
228 2011-07-01T16:51:25 <ThomasWaldmann> but yes, we need more data in moin for testing. 4 items is not very realistic.
229 2011-07-01T16:51:34 <Marchael> :)
230 2011-07-01T16:51:54 <sinha> ThomasWaldmann: please tell me what is wrong in this line {{ url_for('frontend.index', startswith=letter) }} ?
231 2011-07-01T16:51:58 <ThomasWaldmann> I have a client that has many gigabytes of stuff in his moin.
232 2011-07-01T16:52:24 <Marchael> ubuntu? :D
233 2011-07-01T16:52:29 <ThomasWaldmann> sinha: at first look, this looks ok
234 2011-07-01T16:52:47 <ThomasWaldmann> no, you wouldn't recognize anyway.
235 2011-07-01T16:53:17 <sinha> ThomasWaldmann: But i am getting this error http://paste.pocoo.org/show/425630/
236 2011-07-01T16:53:27 <ThomasWaldmann> sinha: letter name for a unicode object?
237 2011-07-01T16:53:51 <sinha> ThomasWaldmann: yes, but it works if i just add item_name='' inside this
238 2011-07-01T16:54:36 <ThomasWaldmann> how does frontend.index look like right now?
239 2011-07-01T16:55:09 <sinha> ThomasWaldmann: showing the indexes, you need a screenshot ?
240 2011-07-01T16:55:19 <ThomasWaldmann> i mean teh source :)
241 2011-07-01T16:57:22 <ThomasWaldmann> Marchael: http://hg.moinmo.in/moin/2.0/file/7a81ff51008a/MoinMoin/util/filesys.py#l31 portable file rename functions with specific semantics
242 2011-07-01T16:57:48 <ThomasWaldmann> and if you want to see windows insanity, look into werkzeug.posixemulation
243 2011-07-01T16:58:23 <ThomasWaldmann> the windows touch() starting from line 70 is also nice example.
244 2011-07-01T16:58:49 <Marchael> Is os.rename not enought?
245 2011-07-01T16:58:55 <sinha> ThomasWaldmann: it looks like this http://paste.pocoo.org/show/425888/, currently i have moved the code to fetch first letters of item in flat_index itself ( will move it soon)
246 2011-07-01T16:58:55 <ThomasWaldmann> and line 91+ is LOL anyway
247 2011-07-01T16:59:13 <ThomasWaldmann> Marchael: for every sane OS, it usually is
248 2011-07-01T16:59:52 * ThomasWaldmann needs to leave soon and will be offline the rest of the days mostly
249 2011-07-01T17:00:20 <ThomasWaldmann> sinha: that is global_index
250 2011-07-01T17:00:34 <ThomasWaldmann> not index
251 2011-07-01T17:00:38 <sinha> yes +index points to global_index
252 2011-07-01T17:00:50 <ThomasWaldmann> url_for('frontend.index'
253 2011-07-01T17:01:11 <ThomasWaldmann> you wanted frontend.global_index maybe?
254 2011-07-01T17:01:41 <sinha> But this is the route for gloabl_index function @frontend.route('/+index') , right ?
255 2011-07-01T17:02:13 <ThomasWaldmann> btw, just as an announcement for everybody: i have flask 0.7.1 stuff in my workdir, it works now, after some fighting
256 2011-07-01T17:02:31 *** m4k3r
257 2011-07-01T17:02:39 <ThomasWaldmann> as soon as I commit and push that and you merge that, you will be required to upgrade to flask 0.7.1
258 2011-07-01T17:02:57 <ThomasWaldmann> like with pip install --upgrade flask
259 2011-07-01T17:03:19 <ThomasWaldmann> (but wait until you have that changeset, if you do it before, everything will break)
260 2011-07-01T17:03:37 <ThomasWaldmann> (if you don't upgrade flask, everything will be broken also after that cs)
261 2011-07-01T17:04:00 <ThomasWaldmann> Marchael: sinha: also tell you non-present fellow students
262 2011-07-01T17:04:18 <Marchael> ok
263 2011-07-01T17:11:07 <Marchael> ThomasWaldmann: rename_no_overwrite can be using only for regular files?
264 2011-07-01T17:11:19 <ThomasWaldmann> yes
265 2011-07-01T17:12:07 <Marchael> I thought what you point me to use it instead of os.rename()
266 2011-07-01T17:12:51 <sinha> ThomasWaldmann: yes frontend.global_index works, but i am really confused due to that frontend.route thing, anyways i better keep moving
267 2011-07-01T17:14:25 <ThomasWaldmann> sinha: I also have sometimes that BuildError and don't find it too informative
268 2011-07-01T17:14:34 <ThomasWaldmann> but usually the reason can be found
269 2011-07-01T17:16:30 <Marchael> os.rename() can't rename if dst_dir contains some files
270 2011-07-01T17:16:54 <ThomasWaldmann> on posix, it can
271 2011-07-01T17:17:19 <Marchael> OSError: [Errno 39] Directory not empty
272 2011-07-01T17:17:34 <ThomasWaldmann> and btw, shutil.move likely just calls os.rename internally
273 2011-07-01T17:17:49 <ThomasWaldmann> are you renaming files or directories?
274 2011-07-01T17:17:57 <Marchael> directories of course
275 2011-07-01T17:18:11 <ThomasWaldmann> ok, that doesn't work with os.rename
276 2011-07-01T17:18:39 <Marchael> so then I should first remove all stuff from index_dir
277 2011-07-01T17:19:05 <Marchael> because old indexes still located there
278 2011-07-01T17:19:10 <ThomasWaldmann> we already had that topic
279 2011-07-01T17:19:41 <ThomasWaldmann> if you kill the complete index_dir, you kill all indexes, not only the one(?) you are dealing with currently
280 2011-07-01T17:19:51 <ThomasWaldmann> anyway, gtg...
281 2011-07-01T17:20:46 <Marchael> bye
282 2011-07-01T17:41:46 <Marchael> for renaming regular files(if dst exists) with rename_no_overwrite I get OSError: [Errno 17] File exists
283 2011-07-01T17:42:03 <Marchael> (delete_old=True)
284 2011-07-01T18:34:48 <CIA-45> Michael Mayorov <marchael@kb.csu.ru> * 3ec32fa2fe01 r296 moin-2.0/ (3 files in 2 dirs): Merge update index and build index script to one, improvements in cli syntax
285 2011-07-01T18:34:49 <CIA-45> Michael Mayorov <marchael@kb.csu.ru> * 7ad7f9472ae0 r297 moin-2.0/ (3 files in 3 dirs):
286 2011-07-01T18:34:49 <CIA-45> Fixing some names for better usability,
287 2011-07-01T18:34:49 <CIA-45> changes in move functions(not it remove old indexes before add new),
288 2011-07-01T18:34:49 <CIA-45> fix bug with copy empty indexes.
289 2011-07-01T18:34:49 <CIA-45> TODO: whoosh searcher can't find items moving to trash, so it continuely try to add trashed revisions to index.
290 2011-07-01T18:39:14 <Marchael> bbl
291 2011-07-01T18:39:18 *** Marchael
292 2011-07-01T18:41:54 *** RogerHaase
293 2011-07-01T18:54:24 *** mkerrin
294 2011-07-01T18:55:15 *** RogerHaase
295 2011-07-01T18:56:50 *** RogerHaase
296 2011-07-01T19:43:00 *** greg_f
297 2011-07-01T20:45:20 <dreimark> re
298 2011-07-01T21:38:32 <sinha> dreimark: what is pep8 error in this line def flat_index(self, startswith=None):` ?
299 2011-07-01T21:41:08 <CIA-45> pkumar <contactprashantat@gmail.com> * caf03e340f75 r333 pytest2/MoinMoin/themes/_tests/test_navi_bar.py: added test for location_breadcrumbs in themes/test_navi_bar
300 2011-07-01T21:41:08 <CIA-45> pkumar <contactprashantat@gmail.com> * e9c80fcc5ae5 r334 pytest2/MoinMoin/themes/_tests/test_navi_bar.py: test_location_breadcrumbs improved to check segment_name and item_name in themes/test_navi_bar
301 2011-07-01T21:41:09 <CIA-45> pkumar <contactprashantat@gmail.com> * 27dc049120b1 r335 pytest2/MoinMoin/themes/_tests/test_navi_bar.py: test for shorten_item_name() added in themes/test_navi_bar
302 2011-07-01T21:41:09 <CIA-45> pkumar <contactprashantat@gmail.com> * ed593361022e r336 pytest2/MoinMoin/themes/_tests/test_navi_bar.py: test added for contenttype_to_class() in themes/test_navi_bar
303 2011-07-01T21:45:37 *** sinha
304 2011-07-01T21:45:52 *** sinha
305 2011-07-01T21:53:55 <ThomasWaldmann> re
306 2011-07-01T22:02:34 *** sinha
307 2011-07-01T22:02:54 *** sinha
308 2011-07-01T22:07:59 <dreimark> sinha: before it was the upper case letter in that word
309 2011-07-01T22:08:08 <sinha> okay
310 2011-07-01T22:09:06 <dreimark> look sometimes with pylint on the code
311 2011-07-01T22:09:26 <dreimark> it is a bit too verbose but you can get an idea what is not sane
312 2011-07-01T22:20:21 <sinha> dreimark: http://www.moinmo.in/AkashSinha/Gsoc2011Diary/2011-07-01
313 2011-07-01T22:32:33 *** sinha
314 2011-07-01T22:50:04 *** m4k3r
315 2011-07-01T23:58:41 *** RogerHaase
316
