1 2011-05-16T00:18:05 *** grzywacz
2 2011-05-16T02:15:13 *** cosmos_thought
3 2011-05-16T02:15:18 <cosmos_thought> moinmoin!
4 2011-05-16T02:15:23 <cosmos_thought> :)
5 2011-05-16T02:15:29 <cosmos_thought> im in for nightshift :D
6 2011-05-16T03:06:50 *** MattMaker
7 2011-05-16T03:58:41 *** cosmos_thought_
8 2011-05-16T04:01:47 *** cosmos_thought
9 2011-05-16T05:32:47 *** Marchael
10 2011-05-16T06:23:43 *** cosmos_thought_
11 2011-05-16T08:54:27 *** gwork
12 2011-05-16T09:38:42 <ThomasWaldmann> moin
13 2011-05-16T10:04:48 *** m4k3r
14 2011-05-16T10:16:52 *** MattMaker
15 2011-05-16T10:21:16 *** m4k3r
16 2011-05-16T10:28:12 *** m4k3r
17 2011-05-16T10:32:44 *** m4k3r
18 2011-05-16T12:56:09 <dreimark> ronny: see ML special grant
19 2011-05-16T13:00:31 <ronny> dreimark: hu?
20 2011-05-16T13:03:22 <dreimark> We are happy to announce that EuroPython is offering a grant with 3 free tickets reserved to students and 3 free tickets reserved to unemployed programmers, including hotel accommodation.
21 2011-05-16T13:03:26 <dreimark> ronny: ^
22 2011-05-16T13:03:54 <dreimark> xorAxAx: waldi ^
23 2011-05-16T13:04:15 <dreimark> http://mail.python.org/pipermail/europython/2011-May/007927.html
24 2011-05-16T13:05:36 <xorAxAx> i am busy in the EP timespan
25 2011-05-16T13:07:03 <ronny> hmm, lets see what to write down
26 2011-05-16T13:07:16 *** greg_f
27 2011-05-16T13:15:50 *** Marchael
28 2011-05-16T14:10:38 *** Marchael
29 2011-05-16T15:26:02 *** Marchael
30 2011-05-16T15:45:23 *** Marchael
31 2011-05-16T16:42:00 *** ronny
32 2011-05-16T16:46:33 *** ronny
33 2011-05-16T17:03:18 *** gwork
34 2011-05-16T17:52:23 *** RogerHaase
35 2011-05-16T18:30:56 *** Marchael
36 2011-05-16T19:18:55 <dreimark> .oO 7days left before coding starts
37 2011-05-16T19:19:40 <dreimark> and still lots to discuss, document for some of out students
38 2011-05-16T19:24:38 <xorAxAx> yep
39 2011-05-16T19:45:15 *** greg_f
40 2011-05-16T20:03:18 *** grzywacz
41 2011-05-16T20:26:41 *** ronny
42 2011-05-16T20:27:19 <RogerHaase> ThomasWaldmann: dreimark: please review http://codereview.appspot.com/4526043/
43 2011-05-16T20:27:34 *** ronny
44 2011-05-16T21:24:39 *** RogerHaase
45 2011-05-16T21:25:42 *** RogerHaase
46 2011-05-16T21:26:30 <ThomasWaldmann> re
47 2011-05-16T21:44:41 <ThomasWaldmann> RogerHaase: looking at it now
48 2011-05-16T21:48:59 <ThomasWaldmann> thanks for adding comments
49 2011-05-16T21:50:49 <ThomasWaldmann> RogerHaase: btw, in case you work under windows, please make sure your editor does not convert files to crlf line endings, nor introduces tab characters
50 2011-05-16T21:51:43 *** m4k3r
51 2011-05-16T21:56:32 <RogerHaase> ThomasWaldmann: sorry. I saw your note of a few days ago. I thought I had scite set to lf only, but obviously not for css.
52 2011-05-16T22:02:57 <ThomasWaldmann> i tried that hg eolextension, but somehow it behaved strange for me. so looks like we need to care manually.
53 2011-05-16T22:04:06 <TheSheep> you can just check for eol and drop the push if it's wrong
54 2011-05-16T22:04:11 <ThomasWaldmann> btw, ignore that "global comment". i though i have to use the "mail" functionality to get my first comment out, but obviously that was sent immediately.
55 2011-05-16T22:04:16 <TheSheep> that's better than converting it behind the scenes
56 2011-05-16T22:05:02 <ThomasWaldmann> TheSheep: I'ld like to have something reasonable for eol and tabs
57 2011-05-16T22:05:13 <ThomasWaldmann> but i didn't find anything
58 2011-05-16T22:05:23 <TheSheep> ThomasWaldmann: it's easy to write
59 2011-05-16T22:05:43 <ThomasWaldmann> (and locally, something pre-commit would be nice to avoid unneccessary work)
60 2011-05-16T22:06:01 <RogerHaase> it is better to have crlf for the few windows .bat files (else many windows editors display everything as one line)
61 2011-05-16T22:06:25 <ThomasWaldmann> yeah, it is not totally trivial
62 2011-05-16T22:06:50 <ThomasWaldmann> i thought eolextension was doing it on the right level, but it behaved rather strange
63 2011-05-16T22:07:03 <ThomasWaldmann> and strangely different depending on its version
64 2011-05-16T22:07:58 <TheSheep> anything that changes your data behind your back is just wrong
65 2011-05-16T22:08:10 <ThomasWaldmann> TheSheep: also, i don't mind if there is a 0x09 in a .png files :)
66 2011-05-16T22:08:25 <TheSheep> ThomasWaldmann: merurial recognizes binary files
67 2011-05-16T22:09:08 <ThomasWaldmann> hopefully by a better algorithm than i saw recently somewhere else :)
68 2011-05-16T22:09:23 <TheSheep> btw, I have a version of key-value storage for mercurial that doesn't depend on the filesystem
69 2011-05-16T22:09:44 <ThomasWaldmann> for hg backend?
70 2011-05-16T22:09:53 <TheSheep> possibly
71 2011-05-16T22:10:14 <RogerHaase> ThomasWaldmann: re. Why does it change +get to +modify? -- it is changing both "+get" to "" and "+modify" to ""
72 2011-05-16T22:10:37 <ThomasWaldmann> TheSheep: how does it work currently? and how can it not depend on filesystem?
73 2011-05-16T22:11:18 <ThomasWaldmann> RogerHaase: oops %). well, then the question is "why do we have +modify?".
74 2011-05-16T22:12:21 <TheSheep> ThomasWaldmann: it talks to mercurial repo directly, not through the working copy
75 2011-05-16T22:12:32 <RogerHaase> I didn't understand the +modify either, but I left it in
76 2011-05-16T22:12:40 <TheSheep> ThomasWaldmann: so as long as you don't update your repo, you can have any filenames you like in there
77 2011-05-16T22:13:31 <TheSheep> ThomasWaldmann: of course, with some filenames you won't be able to update your working copy on some filesystems
78 2011-05-16T22:13:44 <TheSheep> ThomasWaldmann: but that doesn't stop the storage from working
79 2011-05-16T22:13:47 <ThomasWaldmann> TheSheep: ah, ok
80 2011-05-16T22:14:11 <ThomasWaldmann> RogerHaase: ok
81 2011-05-16T22:15:05 <ThomasWaldmann> TheSheep: i recently thought about having some kiss frontend to mercurial to have something like a web dropbox thing
82 2011-05-16T22:15:46 <ThomasWaldmann> (much more simple and stupid than hgweb, with browser drag&drop, revctrl for dummies somehow)
83 2011-05-16T22:16:09 <TheSheep> ThomasWaldmann: sounds fun
84 2011-05-16T22:16:27 <TheSheep> ThomasWaldmann: I will see if I can make one of our interns do that
85 2011-05-16T22:16:28 <ThomasWaldmann> but I am not sure anymore it can be done, some stuff just isn't that easy
86 2011-05-16T22:16:45 <TheSheep> you can do an 80% solution
87 2011-05-16T22:16:54 <TheSheep> and just ignore the remaining 20%
88 2011-05-16T22:16:59 <ThomasWaldmann> problem is it should be r/w
89 2011-05-16T22:17:11 <ThomasWaldmann> and w means conflicts
90 2011-05-16T22:17:33 <TheSheep> for binaries you just want to error on conflicts
91 2011-05-16T22:17:41 <ThomasWaldmann> also, you lose important metadata when not using hg client, like parent revision
92 2011-05-16T22:17:43 <TheSheep> because you can just upload the file again
93 2011-05-16T22:18:01 <TheSheep> not at all
94 2011-05-16T22:18:12 <ThomasWaldmann> how so?
95 2011-05-16T22:18:16 <TheSheep> you can include the parent revision in the upload form
96 2011-05-16T22:18:41 <ThomasWaldmann> and which dummy should fill that out? :)
97 2011-05-16T22:18:51 <TheSheep> ajax request?
98 2011-05-16T22:19:05 <TheSheep> get the revision at the moment you started to upload the file
99 2011-05-16T22:19:17 <TheSheep> and if it's the same at the moment the upload finished, you don't have a conflict
100 2011-05-16T22:19:31 <ThomasWaldmann> no
101 2011-05-16T22:20:20 <ThomasWaldmann> if file has a base B, x checks out B, y checks out B, x checks in modified B, y checks in modified B
102 2011-05-16T22:20:35 <ThomasWaldmann> then you need to know that the last one is based on B
103 2011-05-16T22:20:48 <ThomasWaldmann> (not on B')
104 2011-05-16T22:21:06 <ThomasWaldmann> one maybe could put it into filename, but that's a bit ugly
105 2011-05-16T22:21:52 <TheSheep> well, for modifying files you would check them out with hg anyways
106 2011-05-16T22:22:01 <ThomasWaldmann> somehow that "filesystems suck due to missing standardized metadata support" is a recurring theme
107 2011-05-16T22:22:16 <ThomasWaldmann> TheSheep: target group is not developers
108 2011-05-16T22:22:26 <ThomasWaldmann> the want something shiny and easy :)
109 2011-05-16T22:22:43 <TheSheep> ThomasWaldmann: target group can't understand the concept of conflicts
110 2011-05-16T22:22:53 <TheSheep> ThomasWaldmann: besides, you can't merge binaries anyways
111 2011-05-16T22:23:04 <TheSheep> ThomasWaldmann: so it's either error or overwrite
112 2011-05-16T22:23:11 <ThomasWaldmann> yes, but you still need to know whether you have a change or a conflict
113 2011-05-16T22:23:20 <TheSheep> ThomasWaldmann: with overwrite being the thing the users are used to
114 2011-05-16T22:23:33 <TheSheep> you don't need to know that to overwrite
115 2011-05-16T22:23:55 <ThomasWaldmann> yeah, but then you rather silently lose changes
116 2011-05-16T22:24:09 <TheSheep> nope, you still have the history
117 2011-05-16T22:24:17 <ThomasWaldmann> yes, but noone will notice
118 2011-05-16T22:24:30 <TheSheep> you can track which user downloaded which versions, btw
119 2011-05-16T22:24:42 <ThomasWaldmann> except maybe the one who did the lost change, IF he/she is so carefully watching
120 2011-05-16T22:25:08 <ThomasWaldmann> hmm, yes, doing it serverside. than might solve that.
121 2011-05-16T22:25:18 <ThomasWaldmann> if users are logged in.
122 2011-05-16T22:25:30 <TheSheep> you want logins at least for writing
123 2011-05-16T22:25:44 <ThomasWaldmann> obviously also for downloading
124 2011-05-16T22:26:02 <ThomasWaldmann> but, you still won't know
125 2011-05-16T22:26:12 <ThomasWaldmann> a user could do multiple downloads...
126 2011-05-16T22:26:14 <TheSheep> I think it's an interesting idea and a more or less useful version is doable
127 2011-05-16T22:26:34 <TheSheep> as I said, 80%, not faultproof
128 2011-05-16T22:26:59 <ThomasWaldmann> did you see that jquery.file-upload stuff in moin2?
129 2011-05-16T22:27:11 <TheSheep> no
130 2011-05-16T22:27:17 <TheSheep> I've been extra busy lately
131 2011-05-16T22:27:33 <ThomasWaldmann> click on the right "index" menu item
132 2011-05-16T22:27:40 <ThomasWaldmann> nice drag&drop stuff
133 2011-05-16T22:27:43 <TheSheep> I hope to get a different job in a month or two and then have more time
134 2011-05-16T22:28:03 <ThomasWaldmann> just looks like shit because it didn't get css love yet
135 2011-05-16T22:28:14 <TheSheep> I know :(
136 2011-05-16T22:29:52 <TheSheep> I hope to get back to it at some point
137 2011-05-16T22:30:03 <TheSheep> goodnight for now
138 2011-05-16T22:32:03 <ThomasWaldmann> gn
139 2011-05-16T22:34:26 <RogerHaase> ThomasWaldmann: re. mouseover; on http://test.moinmo.in/mylogo it is not obvious there is a link to the item view; any idea on how to make it more obvious? (other than whole image mouseover)
140 2011-05-16T22:36:06 <ThomasWaldmann> well, one has to know it :)
141 2011-05-16T23:10:59 <RogerHaase> ThomasWaldmann: link back to UL corner: 4526043
142 2011-05-16T23:16:21 <ThomasWaldmann> RogerHaase: ok
143 2011-05-16T23:18:16 <RogerHaase> I will push and add bug report for transcluding wiki page into middle of paragraph issue.
144 2011-05-16T23:22:20 <RogerHaase> ThomasWaldmann: OK to push http://codereview.appspot.com/4432084?
145 2011-05-16T23:26:57 <ThomasWaldmann> i was just wondering about that mailto: test, see my comment there
146 2011-05-16T23:31:07 <ThomasWaldmann> everything else looks good and unsupicious :)
147 2011-05-16T23:31:22 <ThomasWaldmann> and you even fixed those todos :)
148 2011-05-16T23:32:18 <RogerHaase> most mail systems limit special characters to . and _ so the test is rather not very relevant. I could delete both (including the one for matilto://me@foo.bar) or leave for the pytest2 guys to evaluate.
149 2011-05-16T23:33:36 <ThomasWaldmann> waldi: mailto:// doesn't make sense, does it?
150 2011-05-16T23:34:33 *** grzywacz
151 2011-05-16T23:35:02 <ThomasWaldmann> 'mailto:' [RFC 822 Address spec]' ����
152 2011-05-16T23:35:03 <ThomasWaldmann> where '[RFC 822 Address spec]' =�[local-part] "@" [domain] ("," [local-part] "@" [domain])*
153 2011-05-16T23:35:30 <dreimark> re
154 2011-05-16T23:35:50 <ThomasWaldmann> hi dreimark
155 2011-05-16T23:36:08 <RogerHaase> No, mail systems do not accept the // as mail names. A better option, I could fix a test by adding the . and _ (more common characters)
156 2011-05-16T23:37:53 <ThomasWaldmann> yes, have tests for common usecases
157 2011-05-16T23:40:44 <dreimark> good news by TheSheep
158 2011-05-16T23:43:23 <RogerHaase> OK, will fix the test and push (may take a few hours to get to it)
159 2011-05-16T23:43:44 <ThomasWaldmann> great :)
