1 2011-03-13T00:05:36 <MattMaker> should the "expected data" in MoinMoin/storage/_tests/test_serialization.py maybe import and make use of the macros used elsewhere for the same terms? (UUID, NAME, MTIME...) or perhaps it *is* better for the test to fail if those change
2 2011-03-13T00:10:13 <ThomasWaldmann> yes, they should. otoh, won't change too often.
3 2011-03-13T00:16:21 <ThomasWaldmann> and that test is rather fragile anyway, see order of keys
4 2011-03-13T00:28:00 <ThomasWaldmann> MattMaker: are you using windows right now?
5 2011-03-13T00:28:44 <MattMaker> OS X, at the present moment
6 2011-03-13T00:29:51 <ThomasWaldmann> how many failing tests are there when starting py.test from the toplevel dir?
7 2011-03-13T00:31:23 <MattMaker> I am embarrassed to admit, I have not yet fully learned how to use the tests
8 2011-03-13T00:31:35 <MattMaker> I will try it
9 2011-03-13T00:31:48 <ThomasWaldmann> enter the virtualenv
10 2011-03-13T00:31:50 <ThomasWaldmann> py.test
11 2011-03-13T00:31:59 <ThomasWaldmann> that's all :)
12 2011-03-13T00:32:30 <MattMaker> hmm, I will need to do a pull so as to answer the question with the proper data
13 2011-03-13T00:32:33 <MattMaker> doing that now
14 2011-03-13T00:34:53 <MattMaker> btw I see that your timestamp patch fixed the places where MTIME needed to be int(), and it all LGTM
15 2011-03-13T00:38:36 <ThomasWaldmann> LGTM?
16 2011-03-13T00:39:24 <ThomasWaldmann> btw, i am still thinking about whether it should be int or float or int (but with usecs)
17 2011-03-13T00:39:28 <MattMaker> btw, a proposal based on my hacking on fs19: everywhere in all the moin backends when a "default" timestamp is figured, don't use time.time, which is guaranteed to be 100% wrong in sort order relative to other revs. instead, base the default on revno. That gives you a consistent dummy timestamp which conveniently also sorts in order of revision number, and which sorts to "earlier than the other revisions that do have a proper times
18 2011-03-13T00:39:28 <MattMaker> rather than appearing to be the most recent (since time.time is now)
19 2011-03-13T00:40:02 <MattMaker> LGTM = Looks Good To Me
20 2011-03-13T00:40:06 <ThomasWaldmann> ah
21 2011-03-13T00:40:23 <MattMaker> I saw that acronym being used in other open source projects
22 2011-03-13T00:40:34 <MattMaker> for code reviews
23 2011-03-13T00:41:43 <ThomasWaldmann> well, it only should use time.time() in indexing.py, when there is no MTIME meta data entry
24 2011-03-13T00:41:47 <MattMaker> ThomasWaldmann: btw, i am still thinking about whether it should be int or float or int (but with usecs) <== in that case, perhaps it should always be accessed through a moin library function
25 2011-03-13T00:42:00 <ThomasWaldmann> (and there really should be one!)
26 2011-03-13T00:42:04 <MattMaker> in the library function you could change the implementation in one place
27 2011-03-13T00:42:53 <MattMaker> ok. but in indexing.py I still suggest revno rather than time.time as a dummy timestamp
28 2011-03-13T00:43:43 <MattMaker> it preserves the information that the real timestamp was not known, yet preserves the relative order of the unknown timestamps
29 2011-03-13T00:44:28 <ThomasWaldmann> but not relative to the ones with timestamp
30 2011-03-13T00:44:50 <ThomasWaldmann> and iirc the fs19 code tries to do previous_rev.timestamp + 1
31 2011-03-13T00:44:57 <MattMaker> (another alternative would be to search for other revs, find the range of possible timestamps based on the previous known revision and the following known revision... that was complex and buggy, so I stopped trying to do it)
32 2011-03-13T00:45:46 <MattMaker> yes it does, but only in one spot, and by means of a recursion I am not trusting of
33 2011-03-13T00:46:27 <MattMaker> and fails in the case where the missing rev is revision 0
34 2011-03-13T00:46:54 <MattMaker> (and my wiki happens to have two such pages)
35 2011-03-13T00:48:34 <ThomasWaldmann> well, I'ld rather like something that creates monotonic timestamps, not 2 separate ranges
36 2011-03-13T00:49:13 <MattMaker> btw perhaps moin2's setup.py should support a usage like "python setup.py test"
37 2011-03-13T00:50:14 <MattMaker> ThomasWaldmann ... but you can still sort by revno, without introducing falsely confident data
38 2011-03-13T00:51:44 <MattMaker> however, yes, you could make it such that unknowns are always set to the previous known rev's timestamp + 1, and in the case of rev 0, the following known timestamp +1, and if it's the only rev, then time.time is okay
39 2011-03-13T00:52:19 <MattMaker> although perhaps that should be previous timestamp + revno
40 2011-03-13T00:52:38 <ThomasWaldmann> well, create a patch that works with your data
41 2011-03-13T00:52:45 <MattMaker> in case you have a sequence of missing timestamps and want to have timestamp sort in order no matter what
42 2011-03-13T00:53:14 <MattMaker> I will create a patch.
43 2011-03-13T00:53:40 <MattMaker> I just tested my current patches against the head and found them to still be working : )
44 2011-03-13T00:54:58 <MattMaker> I can gather them into a paste if you want to preview them, but half of the diff is extraneous debugging stuff and also possibly some redundant changes that will be removed before I offer it as a formal patch
45 2011-03-13T00:55:42 <ThomasWaldmann> put them into a pastebin, so i can see :)
46 2011-03-13T00:59:30 <MattMaker> okay, warts and all
47 2011-03-13T00:59:32 <MattMaker> http://paste.pocoo.org/show/352680/
48 2011-03-13T01:00:17 <MattMaker> it does some dubious and not yet complete tricks in order to coax my fs19 data into a fs2
49 2011-03-13T01:02:29 <MattMaker> there will be some more tightening up of the specs on the SQLAlchemy declarations, because really that is a place where you want to fail aggressively to catch bugs in the code
50 2011-03-13T01:02:33 <MattMaker> things like nullable=False
51 2011-03-13T01:03:25 <MattMaker> note that the above diff is before your recent timestamp changes
52 2011-03-13T01:05:09 <ThomasWaldmann> well, that stuff looks like many things mixed up
53 2011-03-13T01:05:24 <ThomasWaldmann> can you create clean changesets, one thing at a time?
54 2011-03-13T01:07:01 <MattMaker> of course
55 2011-03-13T01:07:09 <MattMaker> but you wanted to see it :)
56 2011-03-13T01:08:24 <MattMaker> I don't yet think all of my changes are good or needed; I am extracting clean changesets now that I have something working for my wiki import
57 2011-03-13T01:08:31 <ThomasWaldmann> btw, the os.path.getmtime() is not that bad, assuming that one does not have false fs timestamps it'll give the correct value
58 2011-03-13T01:09:02 *** Pranav_rcmas
59 2011-03-13T01:09:06 <MattMaker> but (surprise
60 !) I have some false fs timestamps
61 2011-03-13T01:09:09 <MattMaker> :)
62 2011-03-13T01:10:39 <MattMaker> it can happen for example when doing a naive copy of one's wiki across filesystems and forgetting to preserve fs mtimes
63 2011-03-13T01:10:55 <ThomasWaldmann> sure
64 2011-03-13T01:11:59 <ThomasWaldmann> btw, all that uuid stuff is rather incomplete/unfinished yet
65 2011-03-13T01:12:17 <MattMaker> (at least, I assume that's what I did wrong. but I figured, better that I'd make it work without that)
66 2011-03-13T01:12:39 <MattMaker> *nod* I understand the uuid stuff is incomplete
67 2011-03-13T01:12:56 <ThomasWaldmann> btw, please:
68 2011-03-13T01:12:58 <MattMaker> my current changes are mostly just to get my fs19 wiki imported
69 2011-03-13T01:12:59 <ThomasWaldmann> # comment
70 2011-03-13T01:13:00 <ThomasWaldmann> not
71 2011-03-13T01:13:03 <ThomasWaldmann> #comment
72 2011-03-13T01:13:09 <MattMaker> oops
73 2011-03-13T01:13:17 <MattMaker> ok. I do try to do that
74 2011-03-13T01:13:23 <ThomasWaldmann> (for stuff you want to keep)
75 2011-03-13T01:13:36 <ThomasWaldmann> except for code commented out
76 2011-03-13T01:13:36 <MattMaker> right, some of these comments are not at all for keeping
77 2011-03-13T01:13:42 <MattMaker> :)
78 2011-03-13T01:15:53 <ThomasWaldmann> maybe it is easiest to extract some simple fixes and do them first
79 2011-03-13T01:16:35 <MattMaker> oh, I do. I keep several repos to make diffs clearer
80 2011-03-13T01:17:07 <MattMaker> the reason these are all mashed together, is to work around problems that appear separately in each of the three steps for importing from fs19
81 2011-03-13T01:17:52 <MattMaker> and which depend on each other, and also depend on hacks from one phase coming into the next
82 2011-03-13T01:18:39 <MattMaker> the idea being, just get my data moved over, and then go back and see what the changes mean, and make clean patches
83 2011-03-13T01:21:02 <MattMaker> I haven't tried the xml.py commandline command yet, but it might be handy if the moin server accepted a config setting that means "convert and serialize this fs19, restart and load it into an fs20, and finally restart with the final fs20"
84 2011-03-13T01:21:09 <MattMaker> *fs2
85 2011-03-13T01:21:45 <MattMaker> or alternatively, a commandline script that does all that
86 2011-03-13T01:22:00 <ThomasWaldmann> well, there is save_xml to write stuff out
87 2011-03-13T01:22:35 <ThomasWaldmann> and the xml command now can serialize directly from a moin 1.9 data_dir path
88 2011-03-13T01:22:49 <MattMaker> (currently I do it all by editing wikiconfig_local.py and restarting, twice)
89 2011-03-13T01:23:11 <MattMaker> I saw that xml change and intend to try it yes
90 2011-03-13T01:25:55 <MattMaker> another finding: the "best" hack/fix for those annoying "warning: no previously-included files matching" sorts of errors during setup.py is ... well, it's pretty humorous, really:
91 2011-03-13T01:26:50 <MattMaker> pip *already* is filtering the stdout/stderr of the commands it is running, for exactly this purpose of hiding the definitely-not-relevant errors
92 2011-03-13T01:27:08 <MattMaker> so regexes for those errors can just be added to that :)
93 2011-03-13T01:27:22 *** Pranav_rcmas
94 2011-03-13T01:27:47 <MattMaker> rather, it doesn't filter them out, but increases the log severity level you must use in order to see them
95 2011-03-13T01:28:27 <MattMaker> currently, the "warning: no previously-included files matching" are warnings, and they should not be, at least not for an end user of distutils/setup.py
96 2011-03-13T01:28:42 <MattMaker> it is a flaw in distutils in my opinion
97 2011-03-13T01:29:52 <ThomasWaldmann> yeah, sounds like
98 2011-03-13T01:29:53 <MattMaker> it gets confusing to talk about, sometimes. I am not surprised that pip does something so quirky, so hack-y, to work around it.
99 2011-03-13T01:31:01 <MattMaker> not to mention, you have to have four different regexes because these over-enthusiastic distutils file-globbing errors appear in multiple places, with slightly different wordings %)
100 2011-03-13T01:31:51 <ThomasWaldmann> it's all a big hack :|
101 2011-03-13T01:32:04 <ThomasWaldmann> hopefully distutils2 gets it right
102 2011-03-13T01:32:40 <MattMaker> I also made a monkeypatch that reaches in and corrects the bad warnings situation... but pip thwarts that, because it repeatedly re-imports setuptools and calls all the other setup.py separately and repeatedly
103 2011-03-13T01:33:15 <MattMaker> but, the monkeypatch approach works so long as you are using the setup.py directly
104 2011-03-13T01:33:18 <ThomasWaldmann> %)
105 2011-03-13T01:33:20 <MattMaker> yes
106 2011-03-13T01:33:24 <MattMaker> it is all a big hack
107 2011-03-13T01:34:02 <MattMaker> distutils is not pretty inside, when you go looking into the code
108 2011-03-13T01:35:22 <MattMaker> yes, I hope that distutils2 doesn't simply break old behavior, re-implement some features, and fail to support some needed use cases that the old code did support
109 2011-03-13T01:35:39 <MattMaker> I also hope that moin2 does not do that :D
110 2011-03-13T01:38:34 <ThomasWaldmann> well, some stuff will break
111 2011-03-13T01:38:45 <MattMaker> breaking is fine
112 2011-03-13T01:38:52 <MattMaker> making impossible, not so fine
113 2011-03-13T01:38:54 <ThomasWaldmann> just because it was a dirty hack and is solved now differently
114 2011-03-13T01:38:56 <MattMaker> that's all I meant
115 2011-03-13T01:38:58 *** Pranav_rcmas
116 2011-03-13T01:39:02 <MattMaker> yes, agreed
117 2011-03-13T01:40:28 <MattMaker> ok, I am running the py.test
118 2011-03-13T01:41:55 <MattMaker> ouch: http://paste.pocoo.org/show/352695/
119 2011-03-13T01:42:23 <MattMaker> oh, I still need to do a pull -u in this one, 1 sec
120 2011-03-13T01:43:32 <ThomasWaldmann> try with fs, not fs2
121 2011-03-13T01:43:37 <ThomasWaldmann> see issue tracker
122 2011-03-13T01:46:19 * ThomasWaldmann looks at that right now
123 2011-03-13T01:46:35 <MattMaker> ah, I see
124 2011-03-13T01:46:58 <MattMaker> hmm, index.db is SQLite
125 2011-03-13T01:48:11 <MattMaker> in cases like this, is the test wrong to throw an exception rather than marking it failed and continuing onward?
126 2011-03-13T01:50:37 <ThomasWaldmann> well, running out of file handles isn't nice
127 2011-03-13T01:51:04 <MattMaker> using fseventer on OS X to watch the filesystem, the handles are under /tmp
128 2011-03-13T02:05:43 <MattMaker> the last file accessed before the problem is data/userprofiles/index.db
129 2011-03-13T02:06:29 <MattMaker> after that, just a ton of open() with no filename
130 2011-03-13T02:08:55 <ThomasWaldmann> it is all index.db
131 2011-03-13T02:09:09 <ThomasWaldmann> likely caused by the way conftest.py does stuff
132 2011-03-13T02:13:49 <MattMaker> http://dev.haiku-os.org/ticket/3063
133 2011-03-13T02:15:32 <MattMaker> hmm, ulimit -n on my OS X = 256
134 2011-03-13T02:16:19 <ThomasWaldmann> maybe just continue with your stuff, maybe I can fix this one
135 2011-03-13T02:16:39 <MattMaker> ok
136 2011-03-13T02:35:01 *** Pranav_rcmas2
137 2011-03-13T02:37:34 *** Pranav_rcmas
138 2011-03-13T03:27:37 *** Pranav_rcmas2
139 2011-03-13T03:39:21 *** Pranav_rcmas2
140 2011-03-13T03:41:40 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 113:8135eeb9ab9a 2.0/MoinMoin/storage/backends/fs2.py: fs2: add some missing close() calls for ResultProxies
141 2011-03-13T03:42:32 *** cosmo__
142 2011-03-13T03:45:23 *** cosmo
143 2011-03-13T03:48:02 <ThomasWaldmann> hmm, the close() call fixed might be right, but it doesn't fix the fd leak with fs2 configured when running tests
144 2011-03-13T03:49:00 <ThomasWaldmann> fixes*
145 2011-03-13T03:49:58 <ThomasWaldmann> i suspect it is either MoinMoin/conftest.py (which is rather strage) or the engine bound to metadata.bind (is there some need to "unbind" or "close" it?)
146 2011-03-13T03:51:45 <ThomasWaldmann> (the latter in MoinMoin/storage/backends/fs2.py)
147 2011-03-13T03:56:52 *** Pranav_rcmas2
148 2011-03-13T04:31:42 <MattMaker> hmm, thing that jumps out at me in the logs is that there are 81 before_wikis and only 38 after_wikis
149 2011-03-13T05:21:34 <MattMaker> no, that wasn't it, I just ran a test with fs instead of fs2 and the ratio of before_wiki to after_wiki was 0.5, so I guess that's normal
150 2011-03-13T05:27:20 *** MattMaker
151 2011-03-13T07:22:01 <Marchael> moin
152 2011-03-13T07:24:06 <Marchael> what exactly I need extract from http://pypi.python.org/pypi?%3Aaction=index ?
153 2011-03-13T07:24:50 <Marchael> I mean there are many projects like zope, z3ext, z3c etc
154 2011-03-13T07:26:24 <Marchael> also that list is huge and if you give me some recomendations how to find, I`ll be happy :)
155 2011-03-13T07:33:30 *** m4k3r
156 2011-03-13T07:47:21 *** m4k3r
157 2011-03-13T07:59:10 <Marchael> 14185 projects, if I`m not mistake :)
158 2011-03-13T08:20:14 <dreimark> SSDs can have the same timestamp - in the past we had also on different filesystems lots of trouble with obtaining on timestamp.
159 2011-03-13T08:20:23 <dreimark> btw. moin
160 2011-03-13T08:26:49 <dreimark> Marchael: the problem is we should know which of these projects will become incompatible with us after we had changed our license
161 2011-03-13T08:27:14 <dreimark> the pypi index of a project just tells it is GPL
162 2011-03-13T08:27:30 <dreimark> my idea was based on it tells exactly the licesense
163 2011-03-13T08:27:59 <dreimark> because then you need only to write a parser which creates a list from the base url
164 2011-03-13T08:28:08 <dreimark> project and license
165 2011-03-13T08:28:39 <dreimark> but that did not work
166 2011-03-13T08:29:10 <dreimark> I don't want that you look at all those projects unless we find a way to get easy the information
167 2011-03-13T08:29:45 <Marchael> :)
168 2011-03-13T08:30:44 <Marchael> I`m uploading pages of projects one by one, but slow speed i sprevents me from that
169 2011-03-13T08:31:15 <Marchael> I started half hours ago and uploading 650 pages
170 2011-03-13T08:32:06 <Marchael> Hm, I`m hope what I can upload all pages in 7-8 hours :)
171 2011-03-13T08:35:26 *** m4k3r
172 2011-03-13T08:48:52 <Marchael> just 1k pages!
173 2011-03-13T08:51:46 <dreimark> but the page tells only # License: GPL
174 2011-03-13T08:51:55 *** Pranav_rcmas1
175 2011-03-13T08:52:06 <dreimark> how do you get which exactly it is?
176 2011-03-13T08:52:27 <Marchael> wget -i url_list
177 2011-03-13T08:58:32 <dreimark> i meant the license?
178 2011-03-13T08:59:23 <dreimark> e.g. http://linkchecker.sourceforge.net/
179 2011-03-13T08:59:39 <dreimark> links on its homepage to http://www.gnu.org/licenses/gpl-2.0.html
180 2011-03-13T09:00:29 <dreimark> while on pypi OSI Approved :: GNU General Public License (GPL)
181 2011-03-13T09:27:43 *** Pranav_rcmas1
182 2011-03-13T10:33:58 *** greg_f
183 2011-03-13T11:45:43 <dreimark> bbl
184 2011-03-13T12:13:04 *** grzywacz
185 2011-03-13T12:24:55 *** Pranav_rcmas
186 2011-03-13T13:04:56 *** Marchael
187 2011-03-13T13:10:17 *** Marchael
188 2011-03-13T13:14:41 *** cosmo__
189 2011-03-13T13:16:59 <Marchael> I hope this what you want http://marchael.kb.csu.ru/public/pl.txt
190 2011-03-13T13:17:52 <Marchael> sorry what I can`t post it to the paste because size is too large :)
191 2011-03-13T13:19:48 <Marchael> this list might be incomplete, because several projects have no license line.
192 2011-03-13T13:22:26 <Marchael> oops, just forgot to add apache license projects
193 2011-03-13T13:23:32 *** MattMaker
194 2011-03-13T13:28:12 *** Pranav_rcmas
195 2011-03-13T13:51:56 <Marchael> http://marchael.kb.csu.ru/public/pl.html
196 2011-03-13T13:53:01 <Marchael> ThomasWaldmann: ^
197 2011-03-13T13:58:26 *** MattMaker
198 2011-03-13T13:58:35 *** MattMaker
199 2011-03-13T14:06:09 *** MattMaker
200 2011-03-13T14:06:22 *** MattMaker
201 2011-03-13T14:08:06 *** MattMaker
202 2011-03-13T14:08:22 *** MattMaker
203 2011-03-13T14:15:27 *** Pranav_rcmas
204 2011-03-13T14:23:18 *** MattMaker
205 2011-03-13T14:23:22 *** MattMaker
206 2011-03-13T14:24:00 <ThomasWaldmann> Marchael: how is that list useful?
207 2011-03-13T14:29:25 <Marchael> that list contains projects and license under it published. Of course it might be incomplete.
208 2011-03-13T14:31:13 *** MattMaker
209 2011-03-13T14:31:21 *** MattMaker
210 2011-03-13T14:31:52 <Marchael> > dreimark: because then you need only to write a parser which creates a list from the base url
211 2011-03-13T14:31:52 <Marchael> If I understand right, dreimark ask me to do list of projects with their license
212 2011-03-13T14:38:25 *** Pranav_rcmas
213 2011-03-13T14:39:05 *** MattMaker
214 2011-03-13T14:39:22 *** MattMaker
215 2011-03-13T14:41:14 <ThomasWaldmann> Marchael: as i said already, the data on pypi is not precise enough
216 2011-03-13T14:41:49 <ThomasWaldmann> we need to know which projects are gpl v2 only, because these are problematic if we go v3+
217 2011-03-13T14:42:09 <ThomasWaldmann> while gpl v2 or later is no problem
218 2011-03-13T14:44:26 <ThomasWaldmann> but the data on pypi often just states "gpl" (I doubt they really mean v1) or gpl v2 (has to be doubted also, whether it is v2 or v2+)
219 2011-03-13T14:45:33 <Marchael> http://marchael.kb.csu.ru/public/pl2.html
220 2011-03-13T14:47:12 *** MattMaker
221 2011-03-13T14:47:25 *** MattMaker
222 2011-03-13T14:48:06 <ThomasWaldmann> Marchael: it's hopeless if you base something on flawed/unprecise data
223 2011-03-13T14:48:40 <ThomasWaldmann> e.g. GPL v2.1 - wtf?
224 2011-03-13T14:49:25 <Marchael> I don`t know :)
225 2011-03-13T14:50:59 <Marchael> ThomasWaldmann: so, how can I get precise information about all python GPL v2 projects?
226 2011-03-13T14:51:19 <Marchael> or is it really hopeless?
227 2011-03-13T14:51:37 *** Pranav_rcmas
228 2011-03-13T14:54:36 <ThomasWaldmann> well, one could read their license texts, but that is too much work
229 2011-03-13T14:55:21 *** MattMaker
230 2011-03-13T14:55:27 *** MattMaker
231 2011-03-13T14:55:49 <ThomasWaldmann> the only way is see for pypi is if they would somehow kill all that info, require resubmission with validated and precise info
232 2011-03-13T14:56:03 <ThomasWaldmann> but i don't think this will happen
233 2011-03-13T15:10:23 <Marchael> I spent about 7 hours uploading this pages :(
234 2011-03-13T15:11:56 <Marchael> ok, but if we don`t know precise information about licenses then I can write comment to whoosh proposal
235 2011-03-13T15:19:10 <ThomasWaldmann> uploading?
236 2011-03-13T15:25:30 <Marchael> hm, I mean downloading :)
237 2011-03-13T15:25:36 *** Marchael
238 2011-03-13T15:28:00 *** Marchael
239 2011-03-13T15:48:40 *** Pranav_rcmas2
240 2011-03-13T15:49:55 *** Pranav_rcmas
241 2011-03-13T15:54:19 *** MattMaker
242 2011-03-13T15:54:37 *** MattMaker
243 2011-03-13T16:03:23 *** MattMaker_
244 2011-03-13T16:03:24 *** MattMaker
245 2011-03-13T16:03:25 *** MattMaker_
246 2011-03-13T16:08:22 *** MattMaker
247 2011-03-13T16:08:33 *** MattMaker
248 2011-03-13T16:18:06 <dreimark> re
249 2011-03-13T16:24:25 *** Marchael
250 2011-03-13T16:24:46 *** Marchael
251 2011-03-13T16:26:57 <dreimark> ThomasWaldmann: http://www.gnu.org/licenses/lgpl-2.1.html
252 2011-03-13T16:27:33 <ThomasWaldmann> that's lgpl, not gpl
253 2011-03-13T16:29:03 <dreimark> damn
254 2011-03-13T16:30:02 <dreimark> there was never a gpl2.1
255 2011-03-13T16:34:05 *** m4k3r
256 2011-03-13T16:50:56 <ThomasWaldmann> https://sourceforge.net/tracker/?func=detail&aid=3208983&group_id=66150&atid=513503
257 2011-03-13T16:56:44 <Marchael> nice, but I don`t think what they can fix this bug until March 28...
258 2011-03-13T16:56:58 <ThomasWaldmann> unlikely :)
259 2011-03-13T16:59:58 <ThomasWaldmann> btw, any comments about the new "more" menu link and the browsable docs?
260 2011-03-13T17:05:07 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 114:f436f37cff0d 2.0/setup.py: setup.py: be more precise with the license info
261 2011-03-13T17:08:38 <dreimark> looks interesting, i looked on test.mm first, and didn't understand
262 2011-03-13T17:10:08 <dreimark> ThomasWaldmann: the documentation link should be moved to tab menu
263 2011-03-13T17:10:20 <dreimark> it is not only for admin purposes
264 2011-03-13T17:11:07 *** Marchael
265 2011-03-13T17:12:06 <dreimark> now we need a way to write these pages by the wiki
266 2011-03-13T17:13:52 <ThomasWaldmann> dreimark: moving it to the menu doesn't work currently
267 2011-03-13T17:14:15 <ThomasWaldmann> because in the menu we only support endpoints without params, but for the docs, we need params
268 2011-03-13T17:15:10 <ThomasWaldmann> could be solved by an endpoint doing a redirect, though
269 2011-03-13T17:16:55 <dreimark> ok
270 2011-03-13T17:18:05 <dreimark> it is now cfg.superusers ?
271 2011-03-13T17:19:30 <dreimark> ThomasWaldmann: http://hg.moinmo.in/moin/2.0/file/f436f37cff0d/docs/admin/upgrade.rst#l68
272 2011-03-13T17:19:40 <dreimark> i am not sure which is wrong
273 2011-03-13T17:21:55 <dreimark> http://static.moinmo.in/m2/html/_modules/MoinMoin/user.html#User.isSuperUser
274 2011-03-13T17:24:47 <dreimark> strange wikiconfig_local seems not to be reloaded on change
275 2011-03-13T17:25:06 <ThomasWaldmann> l68 is wrong, must be superusers
276 2011-03-13T17:25:44 *** Marchael
277 2011-03-13T17:26:52 <dreimark> ok
278 2011-03-13T17:27:18 <dreimark> is autoreload working for you?
279 2011-03-13T17:28:24 <ThomasWaldmann> i guess it is working how it is supposed to work
280 2011-03-13T17:30:03 <dreimark> hmm, i am quite sure i have not had to restart moin process some time ago
281 2011-03-13T17:33:32 <ThomasWaldmann> it only watches files of modules in sys.modules
282 2011-03-13T17:35:48 <CIA-50> Reimar Bauer <rb.proj AT googlemail DOT com> default * 115:e330a698dea3 2.0/docs/admin/upgrade.rst: upgrade.rst: fixed config var
283 2011-03-13T17:39:19 <ThomasWaldmann> it looks at stuff under MoinMoin/, under env/, at wikiconfig.py, at /usr/lib/python...
284 2011-03-13T17:40:53 <dreimark> hmm, usually that is enough
285 2011-03-13T17:41:20 <ThomasWaldmann> it doesn't look at wikiconfig_local.py
286 2011-03-13T17:41:58 *** m4k3r
287 2011-03-13T17:42:29 *** m4k3r
288 2011-03-13T17:44:11 <dreimark> we don't need to change this. I was just wondering
289 2011-03-13T17:45:33 <dreimark> ThomasWaldmann: can we get the query params removed from the url if they are not needed?
290 2011-03-13T17:45:51 <dreimark> we don't need for +admin
291 2011-03-13T17:45:55 <dreimark> http://localhost:8080/+admin/?item_name=Bild6&rev=None
292 2011-03-13T17:46:01 <dreimark> such a complex url
293 2011-03-13T17:49:59 <ThomasWaldmann> the reloading of the wikiconfig_local.py could be fixed, but it is not very easy due to the misc ways that configfile can be given
294 2011-03-13T17:50:53 <ThomasWaldmann> adding item_name and rev is a feature of the way the menu works. not needed everywhere, but at some places.
295 2011-03-13T17:51:51 <dreimark> we should show only what is needed
296 2011-03-13T17:51:59 * dreimark imho
297 2011-03-13T17:56:32 <ThomasWaldmann> we could move the actions menu completely to the template
298 2011-03-13T17:57:09 <ThomasWaldmann> if we do not generate it, but just hardcode what we need in the template, it is more flexible
299 2011-03-13T17:57:33 <dreimark> that would make it also easier for someone to overwrite the template
300 2011-03-13T17:58:21 <ThomasWaldmann> if you overwrite the template, it doesn't matter what was there before
301 2011-03-13T17:59:07 <dreimark> true the "menu" is currently somhow connected to the template already
302 2011-03-13T18:02:24 <dreimark> http://hg.moinmo.in/moin/2.0/file/e330a698dea3/MoinMoin/templates/editbar.html#l88
303 2011-03-13T18:02:53 <dreimark> i thought that it is now only "More"
304 2011-03-13T18:06:07 <ThomasWaldmann> hmm, the url args issue is rather in the dispatch handler
305 2011-03-13T18:06:26 <dreimark> i have only changed that strng in base.html
306 2011-03-13T18:06:28 <ThomasWaldmann> see MoinMoin.apps.frontend.views.dispatch
307 2011-03-13T18:22:40 *** Pranav_rcmas2
308 2011-03-13T18:26:53 <dreimark> http://pycon.blip.tv/ videos of PyCon talks
309 2011-03-13T18:27:22 *** greg_f
310 2011-03-13T18:31:05 <dreimark> ok, that needs there only the endpoint
311 2011-03-13T18:32:33 <dreimark> and for some other views a bit more info
312 2011-03-13T18:33:01 * ThomasWaldmann tries to fix that
313 2011-03-13T18:33:14 <dreimark> thx
314 2011-03-13T18:42:47 *** MattMaker
315 2011-03-13T18:43:07 *** MattMaker
316 2011-03-13T19:03:13 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 116:25723df2d237 2.0/MoinMoin/apps/frontend/views.py: dispatch: do not give unneeded args to endpoints not wanting them
317 2011-03-13T19:16:22 <dreimark> :)
318 2011-03-13T19:36:26 <dreimark> even the mimetyoe needs not to be shown in the url :) much better now
319 2011-03-13T19:38:29 <dreimark> ah, that was already on modify
320 2011-03-13T19:59:46 <ThomasWaldmann> dreimark: see issue tracker for reload issue
321 2011-03-13T20:15:33 <dreimark> works
322 2011-03-13T20:17:01 <dreimark> shoudl we add a hgignore entry
323 2011-03-13T20:17:05 <dreimark> ThomasWaldmann: ^
324 2011-03-13T20:17:19 <dreimark> or maybe wikiconfig_.* ?
325 2011-03-13T20:22:14 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 117:0553dfbec661 2.0/.hgignore: hgignore: ignore wikiconfig_*.py
326 2011-03-13T20:27:49 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 118:c9684bc241db 2.0/wikiconfig.py: add hint to wikiconfig about how to get config autoreload
327 2011-03-13T20:27:51 <dreimark> may be it is even better to have the "global" views in a site bar
328 2011-03-13T20:28:04 <dreimark> and the left over ones on the edit bar
329 2011-03-13T20:28:23 <dreimark> side bar
330 2011-03-13T20:28:57 <ThomasWaldmann> yes, that'll be a topic for the new theme
331 2011-03-13T20:30:53 <dreimark> ricks fixed left theme "hiding sidebar" looks interesting too
332 2011-03-13T20:31:15 <dreimark> http://moinmo.in/ThemeMarket/FixedLeft
333 2011-03-13T20:36:01 <ThomasWaldmann> well, that can be nice, but also cause issues
334 2011-03-13T20:49:11 <dreimark> for the theme changes we should also use buttons which don't look like the existing ones
335 2011-03-13T21:34:35 * ThomasWaldmann cleans up some user profile stuff
336 2011-03-13T21:41:31 *** Marchael
337 2011-03-13T21:44:51 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 119:2390e9f1d1dc 2.0/MoinMoin/ (storage/backends/fs19.py user.py):
338 2011-03-13T21:44:51 <CIA-50> fs19: improve user profile processing, user: use int bookmarks
339 2011-03-13T21:44:51 <CIA-50> rename last_saved to MTIME (we also use MTIME as modification timestamp for
340 2011-03-13T21:44:51 <CIA-50> wiki items).
341 2011-03-13T21:44:51 <CIA-50> convert bookmark timestamps from str/usecs to int/secs
342 2011-03-13T21:44:52 <CIA-50> kill 'show_fancy_links', not used any more
343 2011-03-13T21:44:53 <CIA-50> remove 'email_subscribed_events' / 'jabber_subscribed_events', no support yet
344 2011-03-13T21:55:08 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 120:891c6d1551d4 2.0/MoinMoin/storage/backends/fs19.py: fs19: create int MTIME values
345 2011-03-13T22:13:25 <TheSheep> http://trac.edgewall.org/wiki/TracDev/ComponentArchitecture
346 2011-03-13T22:13:47 * TheSheep is sprinting on trac a little
347 2011-03-13T22:14:01 <TheSheep> I will get to know their plugin system a little better
348 2011-03-13T22:46:45 <dreimark> :)
349 2011-03-13T22:59:48 <dreimark> gn
350 2011-03-13T23:31:39 <CIA-50> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 121:9b1f916249b8 2.0/MoinMoin/storage/backends/fs19.py: fs19: fix uuid type inconsistency (bytes vs. unicode)