1 2011-05-19T00:00:32 <ThomasWaldmann> indexing.py:223+
2 2011-05-19T00:13:46 *** sinha
3 2011-05-19T00:23:13 *** RogerHaase
4 2011-05-19T00:25:03 *** grzywacz
5 2011-05-19T00:58:13 *** m4k3r
6 2011-05-19T01:04:55 *** m4k3r
7 2011-05-19T01:08:13 *** m4k3r
8 2011-05-19T05:06:26 *** Marchael
9 2011-05-19T05:25:42 *** ThePing
10 2011-05-19T05:25:42 *** ThePing
11 2011-05-19T05:28:56 <Marchael> moin
12 2011-05-19T06:12:10 <Marchael> in MoinMoin.security.ContentACL in "How ACL is processed" says what "The first matching ACL will tell if the user has access to that resource or not".
13 2011-05-19T06:12:10 <Marchael> Is that mean what if user has read access, but user group has read,write access, then user will have only read access?
14 2011-05-19T07:51:31 <dreimark> Marchael: no
15 2011-05-19T07:52:03 <dreimark> it is always the whole set of rights
16 2011-05-19T07:52:24 <dreimark> but it can happen that you have a user in many groups
17 2011-05-19T07:54:26 <dreimark> and then the order in the acl line is important
18 2011-05-19T07:54:57 <Marchael> dreimark: in MoinMoin.security.ContentACL. cfg == flask.current_app.cfg ?
19 2011-05-19T07:55:00 <dreimark> if at first a group matches with only read rights
20 2011-05-19T07:55:12 <dreimark> then the user has only read rights
21 2011-05-19T07:55:30 <Marchael> I want to use ContentACL for ACL query parsing
22 2011-05-19T07:55:35 <dreimark> or only all the rights given for the group
23 2011-05-19T07:56:47 <dreimark> then no other acl entry is processed
24 2011-05-19T08:00:02 <dreimark> Marchael: i don't understand
25 2011-05-19T08:00:07 <dreimark> why do you want to parse
26 2011-05-19T08:00:16 <dreimark> and not use the existing methods?
27 2011-05-19T08:00:47 <Marchael> whoosh can't tokenize moin acl
28 2011-05-19T08:01:39 <Marchael> so ThomasWaldmann recomend me to use MoinMoin.security.ContentACL for parsing
29 2011-05-19T08:04:04 <dreimark> Marchael: http://hg.moinmo.in/moin/2.0/file/f6e023768f35/MoinMoin/security/_tests/test_security.py#l221
30 2011-05-19T08:30:28 *** ronny
31 2011-05-19T08:43:12 <dreimark> bbl
32 2011-05-19T08:48:03 *** gwork
33 2011-05-19T09:02:49 *** Marchael
34 2011-05-19T09:10:46 <dreimark> ew
35 2011-05-19T09:10:53 <dreimark> re
36 2011-05-19T10:26:09 *** ronny
37 2011-05-19T10:28:14 *** Marchael
38 2011-05-19T11:33:29 *** pkumar
39 2011-05-19T11:41:58 <pkumar> ThomasWaldmann: on EP
40 2011-05-19T11:42:53 <pkumar> my idea was that when we are implementing the tests for the new framework
41 2011-05-19T11:43:08 <pkumar> we'll have it in pytest-2
42 2011-05-19T11:43:49 <pkumar> so, may be we'll have to put less effort while implementing the tests for new framework
43 2011-05-19T12:22:55 <ThomasWaldmann> hi pkumar
44 2011-05-19T12:23:10 <pkumar> hi
45 2011-05-19T12:24:00 <ThomasWaldmann> my thought was just that first fixing all tests for py.test2 and then rewriting some of them for a new framework / for more sanity / to be easier is more work than when doing it in other order / at once.
46 2011-05-19T12:29:08 *** pkumar
47 2011-05-19T12:30:15 *** pkumar
48 2011-05-19T12:30:20 <pkumar> re
49 2011-05-19T12:31:29 <pkumar> just had a network error.
50 2011-05-19T12:33:41 <ThomasWaldmann> my thought was just that first fixing all tests for py.test2 and then rewriting some of them for a new framework / for more sanity / to be easier is more work than when doing it in other order / at once.
51 2011-05-19T12:34:35 <pkumar> ThomasWaldmann: How do we find best suited testing framework for an application
52 2011-05-19T12:38:10 <pkumar> xorAxAx: ^
53 2011-05-19T12:38:39 <xorAxAx> pkumar: i dont understand the question
54 2011-05-19T12:39:25 <pkumar> I mean if we need to decide a framework for MoinMoin tests
55 2011-05-19T12:39:50 <pkumar> what's the way or process of framework selection?
56 2011-05-19T12:39:50 <xorAxAx> well, depends on the test type
57 2011-05-19T12:40:01 <xorAxAx> for integration like tests, you need some lib like paste fixture
58 2011-05-19T12:41:31 <pkumar> What is lib like paste fixture? Can you please elaborate?
59 2011-05-19T12:42:15 <xorAxAx> paste fixture is such a lib
60 2011-05-19T12:42:21 <xorAxAx> please inform yourself
61 2011-05-19T12:42:53 <pkumar> okay, I'll google it :)
62 2011-05-19T12:44:13 <pkumar> ThomasWaldmann: Should I change the work plan for the first week to finalizing of testing framework?
63 2011-05-19T12:44:30 <ThomasWaldmann> talk to your mentors about it
64 2011-05-19T12:44:31 <pkumar> in EP
65 2011-05-19T12:44:38 <pkumar> okay
66 2011-05-19T12:44:58 <pkumar> XorAxAx: What do you suggest about it?
67 2011-05-19T12:45:21 <pkumar> waldi: ^
68 2011-05-19T12:45:30 <xorAxAx> pkumar: no, that should be done before coding starts
69 2011-05-19T12:46:26 <pkumar> We need to decide the testing framework before coding and stick to the current plan?
70 2011-05-19T12:46:33 <pkumar> is it?
71 2011-05-19T12:46:41 *** Marchael
72 2011-05-19T12:47:22 *** greg_f
73 2011-05-19T12:47:23 <xorAxAx> what do you mean by framework here?
74 2011-05-19T12:47:29 <xorAxAx> the framework is py.test 2
75 2011-05-19T12:49:20 <pkumar> yes, it is. But I also meant about the current organization of tests in MoinMoin
76 2011-05-19T12:49:37 <ThomasWaldmann> it might be not ONLY pytest2, see what we have in conftest.py now
77 2011-05-19T12:53:46 <xorAxAx> pkumar: yes, make a plan and discuss it with us
78 2011-05-19T12:56:33 <pkumar> ThomasWaldmann: plan about framework?
79 2011-05-19T12:56:52 <pkumar> xorAxAx:^
80 2011-05-19T12:58:58 <pkumar> ThomasWaldmann: I didn't get 'what we have in conftest.py now' part. Can you please explain?
81 2011-05-19T13:00:27 <ThomasWaldmann> less MoinMoin/conftest.py
82 2011-05-19T13:01:23 <pkumar> okay, you mean the conftest.py code is reduced which makes it more flexible
83 2011-05-19T13:01:35 <pkumar> for other frameworks also
84 2011-05-19T13:01:55 <pkumar> isn't it?
85 2011-05-19T13:05:06 <ThomasWaldmann> what i mean is that you have to think about a framework that is at least as good as that code, preferably better :)
86 2011-05-19T13:06:27 <pkumar> got it
87 2011-05-19T13:06:30 <pkumar> :)
88 2011-05-19T13:12:27 *** pkumar
89 2011-05-19T13:24:38 *** pkumar
90 2011-05-19T13:25:51 *** pkumar
91 2011-05-19T13:26:35 *** pkumar
92 2011-05-19T13:52:19 *** pkumar
93 2011-05-19T13:54:32 *** Marchael
94 2011-05-19T14:06:53 *** Marchael
95 2011-05-19T14:35:57 *** m4k3r
96 2011-05-19T15:02:56 *** Marchael
97 2011-05-19T15:09:52 *** ronny
98 2011-05-19T15:48:48 *** ronny
99 2011-05-19T16:34:49 *** RogerHaase
100 2011-05-19T17:01:20 *** Marchael
101 2011-05-19T17:15:49 <RogerHaase> dreimark: ThomasWaldmann: I removed leading { for element type test on http://codereview.appspot.com/4535086/ - anything else that needs fixing before I push?
102 2011-05-19T17:23:55 * ThomasWaldmann looks
103 2011-05-19T17:25:44 <ThomasWaldmann> if str(elem.tag).endswith('page'): why that sort of check?
104 2011-05-19T17:26:04 <ThomasWaldmann> (what values can elem.tag have?)
105 2011-05-19T17:28:28 <ThomasWaldmann> RogerHaase: ^^
106 2011-05-19T17:29:03 <RogerHaase> ThomasWaldman: if item being transcluded is a page, then it starts with a DIV, so wrapping it in a SPAN causes html error. but wrapping in Span is correct for images.
107 2011-05-19T17:29:37 <ThomasWaldmann> yes, i've seen why you are doing this changeset
108 2011-05-19T17:29:53 <RogerHaase> elem.tag can end in any kind of object AFAIK
109 2011-05-19T17:30:17 <ThomasWaldmann> but that doesn't explain the "endswith('page') check?
110 2011-05-19T17:31:15 <ThomasWaldmann> (also I am unclear about that ">" change in the css, why it is needed)
111 2011-05-19T17:31:25 <RogerHaase> I expected that to be a better way of testing an element to see what kind of object it represents... but str(elem.tag) is the only way I found.
112 2011-05-19T17:32:44 <ThomasWaldmann> so, they values of str(elem.tag) are...?
113 2011-05-19T17:32:48 <ThomasWaldmann> the*
114 2011-05-19T17:34:05 <RogerHaase> with my prior changes, this page with multiple transclusions http://test.moinmo.in/transclusion, shows overlays for all transclusions. With the > added, at most 3 overlays are visible at one time.
115 2011-05-19T17:35:36 <ThomasWaldmann> hmm, don't we want those overlays for all transclusions?
116 2011-05-19T17:36:05 <RogerHaase> yes, but only when you mouse over the transcluded item
117 2011-05-19T17:36:49 <ThomasWaldmann> I don't think I get it.
118 2011-05-19T17:37:05 <RogerHaase> elem.tag = MoinMoin.util.tree.Name(u'{http://moinmo.in/namespaces/page}page')
119 2011-05-19T17:37:47 *** gwork
120 2011-05-19T17:38:30 <RogerHaase> I think page can be replaced by any valid item type.
121 2011-05-19T17:40:11 * ThomasWaldmann is confused and has to read some src
122 2011-05-19T17:40:29 <RogerHaase> The page (http://test.moinmo.in/transclusion) is presently an Easter Egg hunt. It shows only 2 of the 14(?) transcluded items.
123 2011-05-19T17:41:30 <RogerHaase> With my changes, all of the transcluded items will be shown when the corresponding item is mouseovered
124 2011-05-19T17:42:54 <RogerHaase> clarification: it shows the overlays for only 2...
125 2011-05-19T17:44:23 <ThomasWaldmann> if body.tag.uri == moin_page and body.tag.name == 'list-item-body':
126 2011-05-19T17:44:34 <ThomasWaldmann> are you searching for a check similar to that one ^^ ?
127 2011-05-19T17:46:03 <RogerHaase> I think so, checking my traceback...
128 2011-05-19T17:46:13 <ThomasWaldmann> waldi: please help with that, that stuff always confuses me
129 2011-05-19T17:46:45 <ThomasWaldmann> and maybe we start with first having an exact definition of what we need to do :)
130 2011-05-19T17:47:20 <RogerHaase> elem.tag.uri = u'http://moinmo.in/namespaces/page' -- much nicer!
131 2011-05-19T17:47:49 <ThomasWaldmann> be aware that that is just the namespace
132 2011-05-19T17:48:38 <RogerHaase> elem.tag.name = u'page' - nicer still
133 2011-05-19T17:50:30 <RogerHaase> AFAIK, I only need to distinguigh transcluded pages from everything else here -- so isn't testing elem.tag.name the way to do it?
134 2011-05-19T17:51:22 <waldi> RogerHaase: not sure what you are talking about
135 2011-05-19T17:51:45 <waldi> you never need to distiguish them, they'll only show up with a different base uri
136 2011-05-19T17:51:49 <RogerHaase> I know that feeling :_)
137 2011-05-19T17:54:14 <RogerHaase> In this case we are trying to wrap the element with either SPANs or DIVs as appropriate so we can make an overlay link to the transcluded item page.
138 2011-05-19T17:55:30 <RogerHaase> page items always have embedded DIVs so the overlay must be a DIV as well.
139 2011-05-19T17:59:43 <waldi> that is the difference between inline-part and page(?)
140 2011-05-19T18:02:16 <RogerHaase> waldi: I think the answer is yes. Transcluded images are frequently embedded in paragraphs, so images must be wrapped in SPANs.
141 2011-05-19T18:02:44 <waldi> images are considered replaced/inline items
142 2011-05-19T18:06:09 <RogerHaase> waldi: now I don't understand. In this code: http://codereview.appspot.com/4535086/diff/4001/MoinMoin/converter/html_out.py we need to test the elem type.
143 2011-05-19T18:06:59 <RogerHaase> waldi: My current understanding is page elements need to be wrapped in DIV, everything else is to be wrapped in SPAN.
144 2011-05-19T18:09:40 <waldi> not sure where this code comes from, it is not mine
145 2011-05-19T18:14:09 <RogerHaase> waldi: I think the only issue is how to best test to see if the elem is a page: I think the answer is to test elem.tag.name for the value u'page'
146 2011-05-19T18:44:46 *** Marchael
147 2011-05-19T18:45:59 <ThomasWaldmann> waldi: still there?
148 2011-05-19T18:48:54 <waldi> yes
149 2011-05-19T18:50:45 <ThomasWaldmann> so, is RogerHaase correct or do you have better suggestions?
150 2011-05-19T18:52:54 <RogerHaase> screenshot of multiple overlays: http://test.moinmo.in/RevisedTransclusion
151 2011-05-19T18:53:49 <ThomasWaldmann> is this because you use the whole object to show it?
152 2011-05-19T18:54:07 <ThomasWaldmann> looks a bit strange, imho
153 2011-05-19T18:54:23 *** franklin
154 2011-05-19T18:55:15 <RogerHaase> yes. I am thinking the overlay needs work to make it less obnoxious and look more like a link (e.g. some kind of blue thing).
155 2011-05-19T18:55:20 *** franklin_
156 2011-05-19T18:55:37 <waldi> RogerHaase: page must only occur in block context, so it is always a div in the output. however this is taken care by the recursion
157 2011-05-19T18:56:55 *** xorAxAx
158 2011-05-19T18:57:33 <RogerHaase> waldi: right. I am thinking the question and answer are now so obvious that the question is confusing.
159 2011-05-19T19:03:06 <RogerHaase> ThomasWaldmann: re image looking strange: 3 layers of transclusions is also strange, in most cases there will only be one overlay link showing.
160 2011-05-19T19:03:26 *** xorAxAx
161 2011-05-19T19:03:26 *** xorAxAx
162 2011-05-19T19:14:37 <RogerHaase> i updated http://codereview.appspot.com/4535086/ with revised test.
163 2011-05-19T19:21:47 <ThomasWaldmann> be careful, you don't check the namespace
164 2011-05-19T19:22:15 <ThomasWaldmann> not sure if it matters, though, at that stage
165 2011-05-19T19:29:39 *** sinha
166 2011-05-19T19:40:46 *** greg_f
167 2011-05-19T19:44:26 *** sinha
168 2011-05-19T19:51:51 <RogerHaase> ThomasWaldmann: Tried to transclude some python code -- it also requires that it be wrapped in DIVs. Looks like I need to create one of every kind of object and make a list of div vs. span.
169 2011-05-19T19:52:11 *** Marchael
170 2011-05-19T19:56:51 <ThomasWaldmann> i feel this should be solved at another place
171 2011-05-19T19:57:33 <ThomasWaldmann> in the dom-tree, it should be somehow clear what's blocklevel and wants a div and what's inline and wants a span
172 2011-05-19T19:59:49 <RogerHaase> Yes, that would make it an easy test here. Are there other places that need to know blocklevel vs. inline?
173 2011-05-19T20:00:30 <RogerHaase> gtg, will work on this more over weekend
174 2011-05-19T20:01:03 <ThomasWaldmann> i guess wald knows best where that stuff is documented
175 2011-05-19T20:01:20 *** RogerHaase
176 2011-05-19T21:02:07 *** ronny
177 2011-05-19T21:20:38 *** ronny
178 2011-05-19T23:19:21 *** xorAxAx
179