1 2013-04-11T00:42:20 *** pooryorick
2 2013-04-11T00:43:28 *** pooryorick
3 2013-04-11T01:26:52 *** dwcramer
4 2013-04-11T02:03:30 <xiaq> ThomasWaldmann: an OT question: what's the grammatical gender of Shell (as in bash, zsh...) in German?
5 2013-04-11T02:03:33 <xiaq> dreimark: ^^
6 2013-04-11T02:04:29 <xiaq> ah, feminine. http://de.wikipedia.org/wiki/Betriebssystem-Shell says "Die Shell ..."
7 2013-04-11T02:17:39 *** dwcramer
8 2013-04-11T02:30:22 *** dwcramer
9 2013-04-11T02:59:20 *** dwcramer
10 2013-04-11T03:59:48 *** dwcramer
11 2013-04-11T04:18:10 *** dwcramer
12 2013-04-11T04:20:03 *** sharky93
13 2013-04-11T04:55:47 *** rihnapstor
14 2013-04-11T05:14:48 *** rihnapstor
15 2013-04-11T05:26:23 *** rihnapstor
16 2013-04-11T06:35:34 *** ronny
17 2013-04-11T06:35:35 *** ronny
18 2013-04-11T07:06:04 *** ashutosh1461
19 2013-04-11T07:28:09 *** rihnapstor
20 2013-04-11T08:02:49 *** rihnapstor
21 2013-04-11T08:45:11 *** greg_f
22 2013-04-11T08:57:40 *** rihnapstor
23 2013-04-11T09:09:57 *** rihnapstor
24 2013-04-11T09:30:21 *** rihnapstor
25 2013-04-11T11:10:13 *** puneet
26 2013-04-11T11:14:08 <ThomasWaldmann> >>> from random import sample
27 2013-04-11T11:14:08 <ThomasWaldmann> >>> ''.join(map(chr, sample(range(33,126),9)))
28 2013-04-11T11:14:08 <ThomasWaldmann> '{F+hod,n>'
29 2013-04-11T11:14:15 <ThomasWaldmann> nice :)
30 2013-04-11T11:26:13 <sharky93> ThomasWaldmann, your take on https://bitbucket.org/thomaswaldmann/moin-2.0/issue/340/image-scaling-markup-does-not-work
31 2013-04-11T11:32:01 <ThomasWaldmann> well, aren't you just asking about something obious?
32 2013-04-11T11:32:12 <ThomasWaldmann> obvious*
33 2013-04-11T11:32:16 <dreimark> moin
34 2013-04-11T11:33:46 <ThomasWaldmann> sharky93: the correct approach to something like that is to first precisely define what we want to support and how the markup should look like. then write correct tests. then make the tests work.
35 2013-04-11T11:57:35 * sharky93 likes the approach
36 2013-04-11T11:58:44 <ThomasWaldmann> for defining what we want, also have a look at moin 1.9 markup
37 2013-04-11T11:59:31 <ThomasWaldmann> there are quite some users who will want to upgrade their wikis and if we don't have a good reason against using same markup, we should use same markup.
38 2013-04-11T12:00:03 <ThomasWaldmann> also note that there are 2 different things about scaling:
39 2013-04-11T12:00:16 <ThomasWaldmann> a) putting width and heigth attrs into img tag
40 2013-04-11T12:00:38 <ThomasWaldmann> b) putting &w=...&h=... into the href url of imag tag
41 2013-04-11T12:05:31 <ashutosh1461> ThomasWaldmann, can you please review https://codereview.appspot.com/8180044/
42 2013-04-11T12:06:12 <ashutosh1461> and is this https://bitbucket.org/thomaswaldmann/moin-2.0/issue/156/ still an issue ?
43 2013-04-11T12:08:37 <sharky93> ThomasWaldmann, i understand that the option (b) for scaling would actually make the server transfer the whole "big" image.
44 2013-04-11T12:08:44 <sharky93> and is costlier
45 2013-04-11T12:09:35 <sharky93> ^^ sry option(a)
46 2013-04-11T12:10:24 <sharky93> is there a downside to (b) ?
47 2013-04-11T12:11:08 *** sharky93
48 2013-04-11T12:11:24 *** sharky93
49 2013-04-11T12:22:45 <ThomasWaldmann> a works everywhere, b is more efficient when used with moin as img src
50 2013-04-11T12:25:10 <ashutosh1461> ThomasWaldmann, in reply to your comment i think my logic is correct, we have to make sure that no other revision exists before destroying the data related to that item. Suppose if we have <rev1, rev2, rev3> existing i cannot delete the data until i make sure that say rev1 or rev2 are destroyed.
51 2013-04-11T12:25:58 <ThomasWaldmann> no
52 2013-04-11T12:27:04 <ThomasWaldmann> the only issue you need to solve is that if multiple revision share the SAME (identical) data object, then the data must not get deleted if there are any other references to it still
53 2013-04-11T12:29:03 <ThomasWaldmann> (and data is always related to a revision (not an item))
54 2013-04-11T12:29:52 <ashutosh1461> can two different revisions having not the same <itemid> share the same dataid ?
55 2013-04-11T12:30:34 <ThomasWaldmann> we have nothing that forbids that (otoh, we also have nothing yet that creates such things)
56 2013-04-11T12:31:00 <ThomasWaldmann> but it is no problem if you just assume "yes"
57 2013-04-11T12:31:21 <ashutosh1461> exactly and on renaming or modifying the itemid gets preserved.
58 2013-04-11T12:31:58 <ThomasWaldmann> the itemid is rather irrelavant to the problem
59 2013-04-11T12:32:12 <ThomasWaldmann> irrelevant*
60 2013-04-11T12:33:14 <ThomasWaldmann> maybe draw a picture if that helps you, label the "pointers" to the objects and find out which pointers/references need counting
61 2013-04-11T12:34:02 <ashutosh1461> i took itemid into consideration as i stated it gets preserved on renaming/modifying and it would be easy to figure out through itemid whether any previous version exists or not.
62 2013-04-11T12:34:47 <ThomasWaldmann> but that's not the semantically correct approach for the problem in question
63 2013-04-11T12:35:16 <ThomasWaldmann> the problem we have is that we must make sure noone kills data something else is still referring to
64 2013-04-11T12:35:42 <ashutosh1461> fine, i will think about it.
65 2013-04-11T12:35:45 <ThomasWaldmann> and that is pretty much the whole problem description
66 2013-04-11T12:35:59 <ashutosh1461> and yes one more thing, is this https://bitbucket.org/thomaswaldmann/moin-2.0/issue/156/ still an issue ?
67 2013-04-11T12:36:37 <ThomasWaldmann> you can easily find that out yourself
68 2013-04-11T12:36:57 <ThomasWaldmann> just do a logged in edit, do a anon edit and look at history afterwards
69 2013-04-11T12:37:40 <ThomasWaldmann> if the last edit is not anon, the bug is still there (if it is anon, please put a comment onto the issue tracker)
70 2013-04-11T12:40:28 <sharky93> ThomasWaldmann, I had a look at the moin 1.9 markup, dont really see a reason to modify it.
71 2013-04-11T12:41:41 <ThomasWaldmann> ok, then write some tests for that with image scaling url / image scaling attrs
72 2013-04-11T12:43:09 <sharky93> i dont have much experience with writing tests, can you suggest a way to begin ?
73 2013-04-11T12:44:35 <ashutosh1461> ThomasWaldmann, i am not able to reproduce the 156 problem but found another problem, take a look https://codereview.appspot.com/8612044/
74 2013-04-11T12:50:00 <ThomasWaldmann> ashutosh1461: yes, correct. I am neither able to... - I am closing it, thanks for pointing to that.
75 2013-04-11T12:51:32 <ThomasWaldmann> ehrm, wait, it was about "rename" - trying again...
76 2013-04-11T12:52:47 <ThomasWaldmann> hmm, that doesn't happen any more either
77 2013-04-11T12:53:41 * ThomasWaldmann closed it
78 2013-04-11T12:54:35 *** puneet
79 2013-04-11T12:55:34 <ashutosh1461> i am getting that problem, the rename one, just do rename 1->2 with any comment , and then rename 2->3 (without entering any comment) then check history/3
80 2013-04-11T12:56:15 *** puneet
81 2013-04-11T13:01:15 <sharky93> ThomasWaldmann, by putting "&w" into the href url of the image tag, did you mean modifying something like this -> "http://localhost:8080/+get/+8112c6929fb74503a1e93dcb815ed827/python.jpg" or getting "&w" into the query field of IRI when the DOM is being constructed.. ?
82 2013-04-11T13:04:16 <sharky93> for which an appropriate URL is constructed by the server
83 2013-04-11T13:08:04 *** puneet
84 2013-04-11T13:10:18 *** puneet
85 2013-04-11T13:10:29 <ThomasWaldmann> well, as query string args as well as tag attrs are something we generically support for [[x|y|there]] or {{x|y|there}}, both should make it into the dom representation somehow
86 2013-04-11T13:19:23 <sharky93> can you give one more example where a query string is necessary for modifications ?
87 2013-04-11T13:21:31 <dreimark> sharky93: have you looked on history e.g. a diff
88 2013-04-11T13:23:59 *** puneet
89 2013-04-11T13:26:06 <dreimark> it is not needed for modifications, but for the revert too
90 2013-04-11T13:27:04 * sharky93 just looked at a diff
91 2013-04-11T13:27:34 <sharky93> so you're referring towards rev1=" "&rev2=" "
92 2013-04-11T13:27:59 *** puneet
93 2013-04-11T13:28:09 <dreimark> http://test.moinmo.in/%2Brevert/%2B20fdffc4cfc747ed8fa4089283dbcd80/TEE
94 2013-04-11T13:29:27 <sharky93> dreimark, I dont understand
95 2013-04-11T13:33:37 <sharky93> dreimark, I was referring to something like this http://test.moinmo.in/%2Bdiff/TEE?rev1=a92705791a904975a18e8839d51c1ec9&rev2=1da1992911654abc850c9fec50a2c559
96 2013-04-11T13:35:26 <dreimark> the second example is for doing a modification by reverting
97 2013-04-11T13:35:47 <dreimark> the other is for showing other query parameters
98 2013-04-11T13:38:17 <sharky93> hmm, ok
99 2013-04-11T13:38:51 <ashutosh1461> dreimark, why don't we have a field for DATA id here,https://bitbucket.org/thomaswaldmann/moin-2.0/src/88660eb0ca43f98d603348084e8aca9c697e8c75/MoinMoin/storage/middleware/indexing.py?at=default#cl-262
100 2013-04-11T13:39:23 <ashutosh1461> DATAID*
101 2013-04-11T13:40:02 <ashutosh1461> and if we don't have a field for DATAID, it is not possible for us to parse a whoosh query searching for DATAID. right?
102 2013-04-11T13:41:53 <ThomasWaldmann> "leave blanks before and after urls"
103 2013-04-11T13:42:03 <sharky93> can you modify this URL http://test.moinmo.in/+get/+d329cce3fde64ba18ac308550596ba82/png by adding &w and &h parameters to actually rescale the image on the server side ?
104 2013-04-11T13:43:57 <ThomasWaldmann> ashutosh1461: likely you are the first one needing it
105 2013-04-11T13:44:32 <ThomasWaldmann> so, please make a separate changeset just adding the field
106 2013-04-11T13:46:22 <ashutosh1461> fine, and did you try reproducing the renaming bug (rename 1->2 with any comment , and then rename 2->3 (without entering any comment) then check history/3)?
107 2013-04-11T13:48:52 <sharky93> dreimark, can you modify this URL http://test.moinmo.in/+get/+d329cce3fde64ba18ac308550596ba82/png by adding &w and &h parameters to actually rescale the image on the server side ?
108 2013-04-11T13:49:25 *** macobo_
109 2013-04-11T13:54:10 *** macobo_
110 2013-04-11T13:57:32 *** RogerHaase
111 2013-04-11T14:01:03 <dreimark> sharky93: to which line of code do you refer to
112 2013-04-11T14:01:45 <sharky93> dreimark, I mean by just modifying the URL
113 2013-04-11T14:03:27 <dreimark> where in the code is that defined?
114 2013-04-11T14:03:28 <sharky93> is it possible ?
115 2013-04-11T14:04:24 *** dwcramer
116 2013-04-11T14:04:46 <sharky93> dreimark, what are you referring to ?
117 2013-04-11T14:05:53 *** allure
118 2013-04-11T14:05:53 *** allure
119 2013-04-11T14:05:59 <RogerHaase> sharky93: I don't think any code exists (yet) to make that work. Deciding what the query string should be and creating the code is part of 340
120 2013-04-11T14:10:36 <ThomasWaldmann> no, we have code, but maybe PIL is not installed there
121 2013-04-11T14:11:05 <ThomasWaldmann> (and PIL/pillow is required for resizing the stuff)
122 2013-04-11T14:12:00 *** allure
123 2013-04-11T14:12:47 <ThomasWaldmann> http://hg.moinmo.in/moin/2.0/file/88660eb0ca43/MoinMoin/items/content.py#l675 < there!
124 2013-04-11T14:17:06 *** MaikuMori
125 2013-04-11T14:19:06 *** bruno__
126 2013-04-11T14:19:06 *** MaikuMori
127 2013-04-11T14:31:31 <sharky93> ah!, some relief :D
128 2013-04-11T14:32:32 *** SteveMcIntyre
129 2013-04-11T14:32:41 *** SteveMcIntyre
130 2013-04-11T14:33:17 *** SteveMcIntyre
131 2013-04-11T14:33:25 *** SteveMcIntyre
132 2013-04-11T14:39:30 *** rihnapstor
133 2013-04-11T14:40:08 <ThomasWaldmann> sharky93: http://test.moinmo.in/+get/png?w=30&t=3
134 2013-04-11T14:40:51 <sharky93> ThomasWaldmann, awesome!
135 2013-04-11T14:43:40 * ThomasWaldmann upgrades that now to current code, so expect some brokenness there
136 2013-04-11T14:49:04 *** kuzman0v
137 2013-04-11T15:04:27 *** spy
138 2013-04-11T15:04:27 *** spy
139 2013-04-11T15:08:35 * ThomasWaldmann has made the test.moinmo.in defunct :(
140 2013-04-11T15:10:53 <dreimark> i see
141 2013-04-11T15:14:40 <dreimark> http://www.google-melange.com/gsoc/connect/google/gsoc2013/moinmoin
142 2013-04-11T16:02:53 <sunu> ThomasWaldmann: dreimark Could you take a look at my proposal and leave some comments on how to improve it ? http://moinmo.in/TarashishMishra/GSOC-2013
143 2013-04-11T16:14:05 *** puneet
144 2013-04-11T16:17:52 <dreimark> sunu: reading
145 2013-04-11T16:24:33 <dreimark> sunu: one thing so i don't forget it later. there are some evaluations, please define a goal until midterm or state of the project
146 2013-04-11T16:25:17 <dreimark> also a milestone time base is a good idea. so you can plan what needs to be done until some day
147 2013-04-11T16:26:03 <sunu> dreimark: I plan to do that once I make sure I'm going on the right way with the idea. So I am waiting for a review first :)
148 2013-04-11T16:26:45 <dreimark> yeah
149 2013-04-11T16:32:10 *** rihnapstor
150 2013-04-11T16:42:19 <ashutosh1461> ThomasWaldmann, it is getting too much complex if we try to introduce a new DATAID filed into the meta,
151 2013-04-11T16:43:53 <ashutosh1461> i think that the earlier based on ITEMID was better.
152 2013-04-11T16:45:21 <ThomasWaldmann> ashutosh1461: explain why it is "getting much too complex" and btw: your itemid based code was not working correctly.
153 2013-04-11T16:49:51 <sharky93> ThomasWaldmann, please review https://codereview.appspot.com/8660044/
154 2013-04-11T16:51:17 *** dwcramer
155 2013-04-11T16:51:44 *** j4nu5
156 2013-04-11T16:54:47 <sharky93> ThomasWaldmann, should the comment here go ? https://bitbucket.org/thomaswaldmann/moin-2.0/src/88660eb0ca43f98d603348084e8aca9c697e8c75/MoinMoin/converter/html_out.py?at=default#cl-90 now that there is minimal or no support for IE8 ?
157 2013-04-11T16:56:43 <dreimark> sharky93: can you propose a unit test for the change in the code review
158 2013-04-11T16:57:20 <j4nu5> hi! i wanted to participate in gsoc 2013 under moinmoin ... "Change API and UI for itemids/revids" and "Trash and Purge Policy" caught my eye ... i am comfortable with python but new to moinmoin .. can someone suggest a good starting point for getting more info about these ideas ... thanks!
159 2013-04-11T16:57:43 <dreimark> ThomasWaldmann: ^
160 2013-04-11T16:57:45 <dreimark> hi j4nu5
161 2013-04-11T16:57:59 <sharky93> dreimark, im not very well acquainted with writing tests, could you suggest a way to begin
162 2013-04-11T16:58:23 <dreimark> sharky93: we have selenium tests too, so you can write a browser test too
163 2013-04-11T16:58:52 <j4nu5> dreimark: hi
164 2013-04-11T16:59:11 <dreimark> sharky93: look at http://hg.moinmo.in/moin/2.0/file/88660eb0ca43/_ui_tests
165 2013-04-11T16:59:21 * sharky93 looks
166 2013-04-11T16:59:46 <j4nu5> dreimark: its not too late to start for gsoc, is it?
167 2013-04-11T16:59:58 <dreimark> j4nu5: you are in time
168 2013-04-11T17:00:14 <ashutosh1461> ThomasWaldmann, i will show it to you in a cr, and in which case the itemid code was not working properly ?
169 2013-04-11T17:02:34 <dreimark> j4nu5: ThomasWaldmann currently sets up the test wiki we provide for the current code base, so i can't show there something but you can fetch the current state
170 2013-04-11T17:02:54 <dreimark> http://moinmo.in/MoinMoin2.0
171 2013-04-11T17:03:31 <dreimark> have you used moin1 we changed lots there
172 2013-04-11T17:04:10 <dreimark> one important change was not to rely on names, we have now uuids and the name is just an attribute in meta
173 2013-04-11T17:04:35 *** rihnapstor
174 2013-04-11T17:05:06 <ashutosh1461> ThomasWaldmann, firstly I cannot find any case in which it goes into the else loop here, https://bitbucket.org/thomaswaldmann/moin-2.0/src/88660eb0ca43f98d603348084e8aca9c697e8c75/MoinMoin/storage/backends/stores.py?at=default#cl-126 as DATAID here is not permanently stored in the meta field so it would never go in the else loop, so why is it there?
175 2013-04-11T17:05:19 *** greg_f
176 2013-04-11T17:05:44 *** puneet
177 2013-04-11T17:06:15 <dreimark> j4nu5: since we have the possibility for multiple names of an item
178 2013-04-11T17:06:24 * ThomasWaldmann has to leave now, will do reviews later...
179 2013-04-11T17:07:25 <dreimark> some items can have same name, so identifying the right one on the name is not possible
180 2013-04-11T17:07:44 <dreimark> if you think on acls for example
181 2013-04-11T17:10:54 <j4nu5> dreimark: hmm ... so aren't "Improve namespaces and multiple names support" and "Change API and UI for itemids/revids" conflicting?
182 2013-04-11T17:15:07 *** j4nu5
183 2013-04-11T17:18:18 *** badescunicu
184 2013-04-11T17:37:10 <ThomasWaldmann> re
185 2013-04-11T17:40:44 <ThomasWaldmann> sharky93: guess we don't need to consider IE8 any more as MS deploys newer IE on Win 7 (and XP is soon out of support anyway)
186 2013-04-11T17:41:04 <ThomasWaldmann> so, if it works GOOD on current browsers, I am fine with it
187 2013-04-11T17:41:37 <ThomasWaldmann> GOOD means that images are shown normally (no malfunctions, no frames and sliders displayed erroneously, etc.)
188 2013-04-11T17:42:38 <sharky93> ThomasWaldmann, yes no such issues, RogerHaase ^^
189 2013-04-11T17:48:31 *** desophos
190 2013-04-11T18:04:39 *** dwcramer
191 2013-04-11T18:07:46 *** badescunicu
192 2013-04-11T18:15:34 <ThomasWaldmann> ashutosh1461: there is no else in line 126
193 2013-04-11T18:17:34 <dreimark> there should a unit test added for this change
194 2013-04-11T18:18:41 <ThomasWaldmann> ashutosh1461: if you mean 148, iirc it happens for deserialization
195 2013-04-11T18:18:52 <ashutosh1461> ThomasWaldmann, i was talking about the if else loop in the line 126 block, but the problem is clear now, thanks :)
196 2013-04-11T18:20:47 <ThomasWaldmann> if else is no "loop"
197 2013-04-11T18:22:06 <ashutosh1461> s/loop/block
198 2013-04-11T19:10:58 <ThomasWaldmann> ashutosh1461: btw, i merged your stuff
199 2013-04-11T19:11:21 <ThomasWaldmann> http://hg.moinmo.in/moin/2.0/rev/d366d1af5e1e but there you didn't give the issue number, nor did you say what exactly was the bug you have fixed
200 2013-04-11T19:12:39 <ThomasWaldmann> looks like it is that: https://bitbucket.org/thomaswaldmann/moin-2.0/issue/352/linking-transclusion-of-subitems
201 2013-04-11T19:14:35 <ashutosh1461> yes
202 2013-04-11T19:15:29 <ashutosh1461> ThomasWaldmann, also take a look at https://codereview.appspot.com/8612044/ , i talked with Roger he was able to reproduce the problem.
203 2013-04-11T19:19:55 <RogerHaase> well, I am thinking there are more issues with comments, when you rename an item, history gets lost rather than following new name; if you make a comment when deleting item, it never shows anywhere, even if item with same name is recreated
204 2013-04-11T19:33:37 * ThomasWaldmann must go, brb
205 2013-04-11T19:39:00 *** SrijanM
206 2013-04-11T20:07:24 <ThomasWaldmann> RogerHaase: history does not get lost
207 2013-04-11T20:07:47 <ThomasWaldmann> it's just our current UI which does only lookup by name, not lookup by itemid
208 2013-04-11T20:08:51 <ThomasWaldmann> and i think both methods are correct, it just depends on what one wants
209 2013-04-11T20:10:33 <ThomasWaldmann> for delete and recreate with same name: that will be a new itemid, so it is not the same item any more. the old (deleted) one will show up in a yet-to-code "trash" view, that shows all items with no name.
210 2013-04-11T20:11:57 <ThomasWaldmann> initially i wanted that trash view to be for admin only, but considering user wants to undelete, it makes even more sense to show it to any user.
211 2013-04-11T20:12:34 <RogerHaase> just noticed that if you have itemA with some history and rename to itemB, then rename itemC with its own history to itemA, the new itemA has old itemA's history. If you delete an item with history and create a new item with same name, the old history reappears.
212 2013-04-11T20:13:16 <ThomasWaldmann> the current history lookup is by-name. so are not looking up by item-identity.
213 2013-04-11T20:13:37 <ThomasWaldmann> +you
214 2013-04-11T20:17:34 <ashutosh1461> ThomasWaldmann, i did work upon what you said about searching according to dataid, but i to emphasize on searching by item id because :- ( first see this error message which comes on destroying a renamed item, http://goo.gl/HHoS9 , this is because of this line http://goo.gl/Ig0CC )
215 2013-04-11T20:17:50 <dreimark> the name is no primary key. that search gives just results not matching
216 2013-04-11T20:19:42 <ashutosh1461> i want to * and is there any problem in my code ?
217 2013-04-11T20:23:10 <ThomasWaldmann> ashutosh1461: i looked at the links you posted, but i still don't get your point. what's the problem with dataid?
218 2013-04-11T20:23:48 <ThomasWaldmann> and, did you make the drawing as i suggested?
219 2013-04-11T20:24:43 <ashutosh1461> yes, i have a clear picture of what the problem is.
220 2013-04-11T20:26:16 <ashutosh1461> see, what happens is after destruction of a revision, for a particular revid => we delete it's data id, then search for another (latest_backend_revision on the basis of searching by itemid) then make that as our primary revision
221 2013-04-11T20:27:46 <dreimark> bbl
222 2013-04-11T20:28:25 <ashutosh1461> so say an item has 3 revs (<rev1,dataid1>,<rev2,dataid2>,<rev3,dataid3>), when we try to delete rev1 , first dataid1 and rev1 are deleted, and rev2 becomes the main revision of the Item which comes by retrieving data from dataid2
223 2013-04-11T20:29:32 <ThomasWaldmann> so?
224 2013-04-11T20:30:17 <ashutosh1461> but in the case of rename the scenario is like this (<rev1,dataid1>,<rev2,dataid1>) so first dataid1 gets deleted and then when we try to find another revision in (latest_backend_revision) based on item-id search we get rev2 and try to retrieve the data using dataid1 which already has been deleted, this is causing the main problem.
225 2013-04-11T20:31:01 <ThomasWaldmann> yeah, that is why we shall not kill data objects that are still referred
226 2013-04-11T20:31:16 <ThomasWaldmann> (and that's exactly what I am saying all the time :)
227 2013-04-11T20:32:57 <ashutosh1461> exactly, and what i am trying to say that, the latest_backend_revision searching is based on item-id, so we can restrict the destruction of items by saying that no need to destroy the data if any previous revision with the same data-id exists
228 2013-04-11T20:33:47 <ashutosh1461> or easier of them all, we can just copy the data while renaming creating a new data id ? much simpler ?
229 2013-04-11T20:35:12 <ThomasWaldmann> that would solve it as long as we don't share dataids with other items
230 2013-04-11T20:35:30 <ThomasWaldmann> but if you lookup by dataid as i say, it solves it for all cases
231 2013-04-11T20:36:29 <ThomasWaldmann> and as dataid is THE reference to data objects, it's quite near-fetched to use that for refcounting
232 2013-04-11T20:37:43 <ThomasWaldmann> (we currently use UUIDs for the dataid, but just imagine we use the data content hash for that, then same data would automatically deduplicate but we would need global refcounting)
233 2013-04-11T20:40:51 <ThomasWaldmann> (renaming is rather irrelevant for that btw, names are just attributes of revisions)
234 2013-04-11T20:41:25 <kuzman0v> is there any change for working on bootstrap implementation in moin, as for gsoc project?
235 2013-04-11T20:41:29 <ThomasWaldmann> (and there can be changing names, 0 or many names, that all does not influence the existence of a revision)
236 2013-04-11T20:41:58 <ThomasWaldmann> kuzman0v: bootstrap is apache license 2, which is not compatible with gpl v2
237 2013-04-11T20:42:02 <kuzman0v> replacing the two themes with new one, one with top bar and one with a sidebar
238 2013-04-11T20:42:07 <ashutosh1461> ThomasWaldmann, got it
239 2013-04-11T20:42:43 <ThomasWaldmann> kuzman0v: other than that, I'ld like a bootstrap theme. they have an issue open about relicensing on their tracker, but it is not done yet.
240 2013-04-11T20:44:47 <ThomasWaldmann> kuzman0v: https://github.com/twitter/bootstrap/issues/2054
241 2013-04-11T20:50:51 <ashutosh1461> ThomasWaldmann, last question in https://bitbucket.org/thomaswaldmann/moin-2.0/src/88660eb0ca43f98d603348084e8aca9c697e8c75/MoinMoin/storage/backends/stores.py?at=default#cl-126 what is the significance of if-else block ? i think the condition of the if-else should be different somehow.
242 2013-04-11T20:51:56 <ThomasWaldmann> 130..147 is the usual case "we save a new revision"
243 2013-04-11T20:52:40 <ThomasWaldmann> it saves the data first, computes the hash and creates a new uuid, then puts hash and uuid into meta and saves that, too
244 2013-04-11T20:52:52 <ashutosh1461> fine and what is the else for ?
245 2013-04-11T20:53:11 <ThomasWaldmann> if we serialize (save) revisions, we save meta and data for them
246 2013-04-11T20:53:25 <ThomasWaldmann> the meta will have a dataid in it
247 2013-04-11T20:53:30 <ThomasWaldmann> and a hash
248 2013-04-11T20:53:41 <ThomasWaldmann> so we just use that
249 2013-04-11T20:53:56 <ThomasWaldmann> (when unserializing / loading)
250 2013-04-11T20:56:10 *** desophos
251 2013-04-11T21:00:02 <kuzman0v> ThomasWaldmann, I see :/ bad.
252 2013-04-11T21:00:20 <kuzman0v> Ok. I will see about the UI imrpovements project and w rite a proposal about that one
253 2013-04-11T21:04:03 <ThomasWaldmann> kuzman0v: yeah. especially if one considers that most projects that choose AL2 want to be rather liberal and that this incompatibility is rarely known or intended from the projects.
254 2013-04-11T21:04:24 <ThomasWaldmann> there was the same issue with whoosh, but matt was kind enough to relicense it.
255 2013-04-11T21:05:21 <ThomasWaldmann> kuzman0v: if they do the license change, we can maybe have such a theme, but I guess you won't bet on that happening in time before gsoc.
256 2013-04-11T21:11:50 *** desophos
257 2013-04-11T21:29:05 *** dwcramer
258 2013-04-11T21:38:18 *** bruno__
259 2013-04-11T21:39:14 <ashutosh1461> ThomasWaldmann, issue updated, https://codereview.appspot.com/8180044/
260 2013-04-11T22:01:21 <ThomasWaldmann> ashutosh1461: review done
261 2013-04-11T22:06:45 <desophos> hi, i'm interested in GSoC and i'd like to possibly work with moinmoin
262 2013-04-11T22:07:04 <desophos> and i'd like to start getting acquainted with you all :)
263 2013-04-11T22:08:08 <desophos> there are a few project on the ideas page that look interesting to me
264 2013-04-11T22:08:10 <desophos> *projects
265 2013-04-11T22:09:21 <ThomasWaldmann> desophos: hi.
266 2013-04-11T22:09:24 <desophos> so i'll start by seeing if i can make a few commits :)
267 2013-04-11T22:10:09 <ThomasWaldmann> great :) make sure to read the moin2 docs, there are quite some infos about development, installing etc. in there.
268 2013-04-11T22:10:29 <desophos> alright, thanks
269 2013-04-11T22:10:56 <desophos> i'll do that
270 2013-04-11T22:11:47 <ThomasWaldmann> http://discover-devtools.codeschool.com/ < look at it with chrome
271 2013-04-11T22:12:43 <ThomasWaldmann> ah, matt is working on whoosh
272 2013-04-11T22:14:20 *** rihnapstor
273 2013-04-11T22:15:49 <ThomasWaldmann> desophos: about the ideas on that page: please note they are just ideas, not ready-to-take projects. you have to make a project from them, by researching them, adding own ideas, working on a timeline, etc.
274 2013-04-11T22:16:20 <ashutosh1461> ThomasWaldmann, about your comment, i thought the default value to be 1 because, i did not have the exact ideas where the unserializing takes place.
275 2013-04-11T22:17:33 <ThomasWaldmann> ok, but you see that a boolean is the better type for such flags?
276 2013-04-11T22:17:58 <ThomasWaldmann> but before you now change it all to True/False: why do you need it at all?
277 2013-04-11T22:17:59 <desophos> oh, of course ThomasWaldmann; i see that they don't really have implementation details
278 2013-04-11T22:18:06 <desophos> so they need more investigating
279 2013-04-11T22:18:59 <ThomasWaldmann> sunu: matt accepted your whoosh pull request! :)
280 2013-04-11T22:23:53 <ashutosh1461> in the previous code, the if/else situation was check <if DATAID not in meta then make new dataid >, but if we have DATAID in our schema, then the condition makes not much sense(though we can check whether the DATAID is empty or not)
281 2013-04-11T22:24:54 <ashutosh1461> but the i like the serialize check more.
282 2013-04-11T22:26:24 <ThomasWaldmann> well, the problem is that (de)serialization is just one application that needs that behaviour
283 2013-04-11T22:26:31 *** RogerHaase
284 2013-04-11T22:26:42 <ThomasWaldmann> but basically it is rather "we already know precisely what we want"
285 2013-04-11T22:27:16 <ThomasWaldmann> and having it as index field is not really requiring that iirc
286 2013-04-11T22:28:20 <ThomasWaldmann> (maybe check how other index fields are handled that are maybe not there always)
287 2013-04-11T22:29:34 *** ronny
288 2013-04-11T22:30:33 *** dwcramer
289 2013-04-11T22:42:13 <ashutosh1461> ThomasWaldmann, issue updated, https://codereview.appspot.com/8180044/
290 2013-04-11T22:50:24 <ThomasWaldmann> ashutosh1461: done
291 2013-04-11T22:56:26 <ashutosh1461> ThomasWaldmann, don't we need to consider the situation of the form {DATAID: None} ?
292 2013-04-11T22:56:56 <ThomasWaldmann> the way you did your check, you are NOT checking that anyway
293 2013-04-11T22:57:26 <ThomasWaldmann> (because .get defaults to None if the key is not in the dict)
294 2013-04-11T22:58:09 <ashutosh1461> if not meta.get(DATAID): => is True in both the cases, 1) if the key is not there or key is there but the value is None
295 2013-04-11T22:58:36 <ThomasWaldmann> yeah
296 2013-04-11T22:58:48 <ThomasWaldmann> or if it is '' or ...
297 2013-04-11T22:59:13 <ThomasWaldmann> btw, there is some validation middleware
298 2013-04-11T22:59:23 <ashutosh1461> but the earlier code if DATAID not in meta: gives True only in the case when the key is not there
299 2013-04-11T22:59:28 <ThomasWaldmann> it's not fully active yet, but sooner or later it will be in power
300 2013-04-11T22:59:54 <ashutosh1461> so what should i do now ? let it be if DATAID not in meta ?
301 2013-04-11T23:00:10 <ThomasWaldmann> yes, because you can't check that simply whether you have a valid dataid or not
302 2013-04-11T23:00:48 <ThomasWaldmann> so it currently is assuming that it is either there and valid (or validated at another place) or not there
303 2013-04-11T23:02:00 <ThomasWaldmann> see MoinMoin.middleware.validation
304 2013-04-11T23:02:48 <ThomasWaldmann> it is currently only emitting warnings to the log, but as soon as it doesn't break anything anymore, it won't let through any invalid stuff to storage any more
305 2013-04-11T23:04:11 <ashutosh1461> ThomasWaldmann, updated, https://codereview.appspot.com/8180044
306 2013-04-11T23:05:59 *** dwcramer
307 2013-04-11T23:09:53 *** kuzman0v
308 2013-04-11T23:14:39 <ashutosh1461> also, found a simple mistake few days back https://codereview.appspot.com/8541046/
309 2013-04-11T23:17:48 *** dwcramer
310 2013-04-11T23:19:18 <ThomasWaldmann> ashutosh1461: review done
311 2013-04-11T23:26:34 *** dwcramer
312 2013-04-11T23:29:55 <ashutosh1461> ThomasWaldmann, yes, i checked the testing part, this should be it https://codereview.appspot.com/8541046/
313 2013-04-11T23:31:27 <ThomasWaldmann> well, you need to check it everywhere. don't assume we have 100% test coverage.
314 2013-04-11T23:32:20 <ThomasWaldmann> guess you meant the other codereview
315 2013-04-11T23:32:33 <ashutosh1461> https://codereview.appspot.com/8180044 *
316 2013-04-11T23:36:46 <ThomasWaldmann> done
317 2013-04-11T23:37:39 <ashutosh1461> looks like i uploaded the wrong one :) , really need some sleep after fixing this one
318 2013-04-11T23:40:57 <ThomasWaldmann> yeah, i already wondered (as it's rather late here, it's much later in india :)
319 2013-04-11T23:41:19 <ashutosh1461> https://codereview.appspot.com/8180044/
320 2013-04-11T23:41:58 <ashutosh1461> its 5:11am here right now, and i have classes from 11 :/
321 2013-04-11T23:44:31 <ThomasWaldmann> ok, fix that little issue, run tests, commit, sleep :)
322 2013-04-11T23:45:16 <ThomasWaldmann> being sleepy or in general not having enough sleep is not good for code quality
323 2013-04-11T23:46:07 <ashutosh1461> one more thing, should i commit this as well? https://codereview.appspot.com/8612044/
324 2013-04-11T23:48:39 <ThomasWaldmann> no, not yet
325 2013-04-11T23:50:22 <ashutosh1461> fine, will talk about it later :)
326 2013-04-11T23:50:29 <ashutosh1461> gn :)
327 2013-04-11T23:50:34 *** ashutosh1461
328
