1 2010-06-28T00:08:39 *** ValentinJaniaut
2 2010-06-28T00:10:16 *** valeuf
3 2010-06-28T00:10:16 *** ValentinJaniaut
4 2010-06-28T00:38:24 *** ValentinJaniaut
5 2010-06-28T00:39:17 *** valeuf
6 2010-06-28T00:39:18 *** ValentinJaniaut
7 2010-06-28T00:45:33 *** ValentinJaniaut
8 2010-06-28T00:46:29 *** valeuf
9 2010-06-28T00:46:29 *** ValentinJaniaut
10 2010-06-28T00:48:15 *** ValentinJaniaut
11 2010-06-28T00:50:32 *** ValentinJaniaut_
12 2010-06-28T00:51:19 *** valeuf
13 2010-06-28T00:53:08 *** ValentinJaniaut
14 2010-06-28T02:04:07 <diofeher> re
15 2010-06-28T02:04:16 <diofeher> dreimark: what book of jquery are you reading?
16 2010-06-28T05:02:57 *** kursor
17 2010-06-28T06:11:15 *** kursor
18 2010-06-28T07:16:15 <CIA-98> Diogenes Augusto <diofeher@gmail.com> default * 8474:9a6925ac12c1 2.0-dev-jinja2-plugins-dfeher/MoinMoin/ (4 files in 2 dirs): [templating] fixed a lot of html errors and turned ui_lang_attr() in attributes (ui_lang, ui_dor, content_lang, content_dir)
19 2010-06-28T07:39:01 *** DmitryAndreev
20 2010-06-28T07:50:22 <CIA-98> Diogenes Augusto <diofeher@gmail.com> default * 8475:30947fa94823 2.0-dev-jinja2-plugins-dfeher/MoinMoin/ (templates/footer.html theme/jinja.py): [templating] fixed trail inconsistency and shouldShowPageInfo merged with pageinfo again
21 2010-06-28T08:38:10 *** ei-grad
22 2010-06-28T08:44:32 <dreimark> diofeher: by Ralph Steyer
23 2010-06-28T08:45:24 <dreimark> moin
24 2010-06-28T08:47:00 <dreimark> unfortunately he is using php
25 2010-06-28T08:50:12 <ThomasWaldmann> moin
26 2010-06-28T08:55:28 *** gwork
27 2010-06-28T08:55:58 <dreimark> http://tablesorter.com/docs/ sounds handy
28 2010-06-28T08:57:18 <dreimark> http://www.blogrammierer.de/jquery-tablesorter-plugin-erweiterung-fur-das-deutsche-datumsformat-dd-mm-yy-hhmm/ german date format
29 2010-06-28T08:58:09 <dreimark> and other extensions
30 2010-06-28T08:58:37 *** ei-grad
31 2010-06-28T08:58:55 <ThomasWaldmann> if one just not takes that german format, but iso, sorting gets much easier...
32 2010-06-28T10:32:18 <ValentinJaniaut_> moin
33 2010-06-28T10:44:39 *** ValentinJaniaut_
34 2010-06-28T10:44:42 *** valeuf
35 2010-06-28T10:58:46 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8475:d3cea939badd 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/html_in.py: Fix bug with non-existing base_url in HTML_IN converter
36 2010-06-28T10:58:48 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8476:11eb014610b8 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/ (_tests/test_html_in.py html_in.py): Automatically surround input with <html> tag for HTML_IN converter
37 2010-06-28T12:06:39 <waldi> "HTML5 also defines a text/html-sandboxed media type for documents using the HTML syntax. This can be used when hosting untrusted content."
38 2010-06-28T12:06:56 <waldi> hmm, sometimes reading the drafts is not that bad
39 2010-06-28T12:21:54 <waldi> okay, elementtree does not even remotely fullfill the html-output rules finally fixed in html5
40 2010-06-28T12:31:14 <ThomasWaldmann> http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28HTML_5%29
41 2010-06-28T12:38:16 <ThomasWaldmann> style attr did survive for html5
42 2010-06-28T12:48:11 *** DmitryAndreev
43 2010-06-28T12:48:12 <ThomasWaldmann> is there also a xml-sandboxed media type for xml docs?
44 2010-06-28T13:41:09 <valeuf> waldi: is there a reason to surround the output of the HTML_OUT converter with <div> ta
45 2010-06-28T13:41:11 <valeuf> *tag
46 2010-06-28T13:46:30 <CIA-98> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 8381:47555f1414e2 2.0-dev/MoinMoin/storage/backends/fs19.py: (log message trimmed)
47 2010-06-28T13:46:30 <CIA-98> fs19 backend: create item uuids, use them in EDIT_LOG_USERID metadata
48 2010-06-28T13:46:30 <CIA-98> Using uuid.uuid4().hex to create uuids for all items (pages, attachments, users)
49 2010-06-28T13:46:30 <CIA-98> and a sqlalchemy based mapping table to map names or old userids to new uuids.
50 2010-06-28T13:46:30 <CIA-98> In the end, we store uuid into item metadata (and also into an Item instance
51 2010-06-28T13:46:31 <CIA-98> attribute .uuid).
52 2010-06-28T13:46:32 <CIA-98> Best usage is to use backend.serialize, first for FsUserBackend, then for
53 2010-06-28T13:56:51 <ThomasWaldmann> valeuf: is it id or class "content"?
54 2010-06-28T13:57:15 <valeuf> heuh, I am not sure I get your point
55 2010-06-28T13:57:24 <ThomasWaldmann> (13:41) < valeuf> waldi: is there a reason to surround the output of the HTML_OUT converter with <div> ta
56 2010-06-28T13:57:52 <valeuf> Actually, I think I know why there is div :p
57 2010-06-28T13:58:04 <valeuf> the <page> tag is converted into <div>
58 2010-06-28T13:58:07 <ThomasWaldmann> (because we usually have the content wrapped into such a div)
59 2010-06-28T13:58:16 <ThomasWaldmann> also, you need some root element for the stuff
60 2010-06-28T13:59:09 <valeuf> ThomasWaldmann: yep, well actually it is not a so big deal, because we won't run cycle roundtrip, so one div more is acceptable
61 2010-06-28T14:10:20 <ThomasWaldmann> when roundtripping, the amount of enclosing divs must not change
62 2010-06-28T14:12:54 <xorAxAx> yep
63 2010-06-28T14:17:26 <ThomasWaldmann> xorAxAx: do you have an idea about how we could proceed with duplicate names in user profiles?
64 2010-06-28T14:17:46 <xorAxAx> why do they exist?
65 2010-06-28T14:17:58 <valeuf> ThomasWaldmann: so maybe I should add something to detect surrounded <div> and not convert it ?
66 2010-06-28T14:18:31 <xorAxAx> i thought that the names are unique
67 2010-06-28T14:18:44 <xorAxAx> valeuf: no, you should think about page<->div
68 2010-06-28T14:18:54 <ThomasWaldmann> xorAxAx: likely because ancient moins did not check. so if a user lost his password, he created another profile with same name.
69 2010-06-28T14:19:11 <xorAxAx> or because name2id is broken
70 2010-06-28T14:19:33 <valeuf> xorAxAx: only if the div surround the whole html code, otherwise I cannot convert it as a <page>
71 2010-06-28T14:19:39 <valeuf> xorAxAx: but I think I see what you mean
72 2010-06-28T14:19:49 <valeuf> I finish to write my roundtrip tests, and I will try it
73 2010-06-28T14:20:00 <xorAxAx> ThomasWaldmann: well, rename both and notify the users by mail
74 2010-06-28T14:20:06 <ThomasWaldmann> the problem is that if you have n profiles with same name, you can't be sure if they are really same identity
75 2010-06-28T14:21:28 <ThomasWaldmann> the problem arises when converting user items with fs19
76 2010-06-28T14:21:58 <ThomasWaldmann> as long as one keeps some uuid (either old or new) as item name, no problem
77 2010-06-28T14:22:22 <ThomasWaldmann> but if one wants to use user name as item name, it is not 1:1
78 2010-06-28T14:23:27 <ThomasWaldmann> notifying users might also not be that helpful, because most accounts likely got recreated because of access to old one was lost
79 2010-06-28T14:25:05 <ThomasWaldmann> (forgotten password, email became invalid, ...)
80 2010-06-28T14:37:55 * ThomasWaldmann implements refcounting for userids (for references from edit-log) - then we'll know which user accounts we really need
81 2010-06-28T14:43:34 <ThomasWaldmann> (on old wikis, there are likely lots of crap accounts for spam users)
82 2010-06-28T14:49:48 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8477:062c2b239d6d 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in_out.py: Add roundtrip tests for the HTML-DOM converters
83 2010-06-28T14:49:50 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8478:c20b7349dfd7 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/html_in.py: Consider div as a possible root element for the html, so avoid div multiplication in roundtrip conversion
84 2010-06-28T15:04:24 <ThomasWaldmann> http://paste.pocoo.org/show/230884/ refcount output
85 2010-06-28T15:05:00 <ThomasWaldmann> dreimark: ^^ do you save that often or did you use some script under your user id?
86 2010-06-28T15:05:14 <valeuf> dreimark: cheater :p
87 2010-06-28T15:09:32 <CIA-98> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 8382:baac971b6413 2.0-dev/MoinMoin/storage/backends/fs19.py: fs19: count how often a user(id) is referenced from edit-log
88 2010-06-28T15:16:57 <dreimark> ThomasWaldmann: that was ab to get the amount of requests for restarting the wsgi script
89 2010-06-28T15:17:20 <dreimark> at the time we have had that exploit in the bug report
90 2010-06-28T15:17:28 <dreimark> and eSyr's fix
91 2010-06-28T15:18:46 <ThomasWaldmann> dreimark: you made EDITs with ab?
92 2010-06-28T15:19:23 * ThomasWaldmann reruns refcounter, last run was interrupted due to db lock
93 2010-06-28T15:19:25 <dreimark> no - no edits
94 2010-06-28T15:19:49 <dreimark> that is wrong - it should only view
95 2010-06-28T15:20:29 * ThomasWaldmann likes sqlalchemy with sqlite, makes stuff easy :)
96 2010-06-28T15:37:54 <ThomasWaldmann> http://paste.pocoo.org/show/230900/ < corrected refcounts (now it was a complete run)
97 2010-06-28T15:44:33 <CIA-98> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 8383:26da1af34568 2.0-dev/MoinMoin/storage/backends/fs19.py:
98 2010-06-28T15:44:33 <CIA-98> fs19 backends: 2nd param is idx_path now
99 2010-06-28T15:44:33 <CIA-98> FsPageBackend(data_dir, idx_path, ...)
100 2010-06-28T15:44:33 <CIA-98> FsUserBackend(user_dir, idx_path, ...)
101 2010-06-28T15:44:33 <CIA-98> So you can easily use /dev/shm for idx_path (moinmo.in wiki pages/users
102 2010-06-28T15:44:34 <CIA-98> create about 2MB index.db).
103 2010-06-28T15:49:43 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8479:abb3fd5b924a 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/ (_tests/test_html_in.py html_in.py): Fix typo in span attribute : baseline-shift instead of base-line-shift
104 2010-06-28T15:49:45 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8480:a8bcdedbbac3 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in.py: Update test_html_in with div as a possible root element
105 2010-06-28T15:49:46 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8481:891864797686 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in_out.py: Add span tests for the roundtrip HTML-DOM-HTML test
106 2010-06-28T15:49:51 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8482:58b7583ff9b1 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in_out.py: Add html_element attr from span in thet tests for the roundtrip HTML-DOM-HTML test
107 2010-06-28T15:49:51 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8483:ab6576296ae9 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/html_out.py: Add support for missing span attribute in the HTML_OUT converter
108 2010-06-28T15:49:51 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8484:2602296c434c 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in_out.py: Unskip tests for the roundtrip html conversions
109 2010-06-28T15:49:52 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8485:091a7c51bf0a 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in_out.py: Update test for the code and list in the roundtrip HTML test
110 2010-06-28T15:49:54 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8486:f2b242113d9b 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/html_out.py: Add support for the different type of list in the HTML out convert
111 2010-06-28T15:55:53 <ThomasWaldmann> $ sqlite3 index.db 'select name, refcount from users where refcount=0;'|wc -l
112 2010-06-28T15:55:53 <ThomasWaldmann> 4675
113 2010-06-28T15:55:58 <ThomasWaldmann> ^^ !
114 2010-06-28T15:56:31 <ThomasWaldmann> $ sqlite3 index.db 'select name, refcount from users where refcount>0;'|wc -l
115 2010-06-28T15:56:31 <ThomasWaldmann> 2506
116 2010-06-28T16:01:09 *** RogerHaase
117 2010-06-28T16:05:47 <ThomasWaldmann> RogerHaase: http://paste.pocoo.org/show/230900/ you are in the top-10 :D
118 2010-06-28T16:11:28 <RogerHaase> Moin! Is that good or bad? What is the list?
119 2010-06-28T16:27:30 <ThomasWaldmann> that's your edit/save count
120 2010-06-28T16:27:57 <RogerHaase> I knocked JürgenHermann out of the top 10 :)
121 2010-06-28T16:28:31 <ThomasWaldmann> hehe, yeah :)
122 2010-06-28T16:28:37 <TheSheep> we need to knock out Reimar ;)
123 2010-06-28T16:29:08 <RogerHaase> On the auto-scroll feature request page, do you want the old stuff moved to a discussion page or just deleted?
124 2010-06-28T16:29:39 <ThomasWaldmann> if it is irrelevant for the current state of the code: delete it
125 2010-06-28T16:29:48 <ThomasWaldmann> keep interesting infos
126 2010-06-28T16:30:21 <ThomasWaldmann> (also delete old attached diffs, we have stuff in VCS now)
127 2010-06-28T16:30:37 <RogerHaase> OK, Give me a day or two. I am back from a short vacation and behind on everything.
128 2010-06-28T16:30:48 <ThomasWaldmann> sure
129 2010-06-28T16:31:20 *** DmitryAndreev_
130 2010-06-28T16:31:38 <ThomasWaldmann> btw, that edit-count is a side-effect of a cleanup procedure I write for 1.9 > 2.0 data migration
131 2010-06-28T16:32:14 <ThomasWaldmann> we maybe could get rid of all user profiles with edit count 0
132 2010-06-28T16:33:19 <ThomasWaldmann> (at least they would not be missed history-wise)
133 2010-06-28T16:34:15 <ei-grad> moin
134 2010-06-28T16:34:48 * ei-grad had ~1500 users in ldap directory of university
135 2010-06-28T16:35:03 *** DmitryAndreev
136 2010-06-28T16:35:06 <diofeher> moin
137 2010-06-28T16:35:17 <diofeher> dreimark: this book is only in german?
138 2010-06-28T16:44:06 <ThomasWaldmann> ei-grad: had? did you delete them? X)
139 2010-06-28T16:44:14 <dreimark> re
140 2010-06-28T16:45:48 <ei-grad> no, i just don't administrating it any more...)
141 2010-06-28T16:46:01 <dreimark> ThomasWaldmann: it increased for me and I have no idea why I have such much edits
142 2010-06-28T16:46:12 <dreimark> diofeher: yes
143 2010-06-28T16:47:07 <dreimark> 36080
144 2010-06-28T16:47:17 <dreimark> and second run 68440
145 2010-06-28T16:48:25 <dreimark> b0rked
146 2010-06-28T16:49:50 <ThomasWaldmann> the first run was incomplete, so increase is expected.
147 2010-06-28T16:50:13 <ThomasWaldmann> maybe you really saved that often :)
148 2010-06-28T16:51:12 <ThomasWaldmann> Ctrl-C is broken somehow
149 2010-06-28T16:53:21 *** kursor
150 2010-06-28T17:00:06 <dreimark> PageHits gets a timeout on MM
151 2010-06-28T17:01:31 <dreimark> second run works
152 2010-06-28T17:01:40 <ThomasWaldmann> i think about extending storage api
153 2010-06-28T17:02:31 <ThomasWaldmann> if we give all items a never-changing uuid (and they also have a name, which may change per revision when it gets renamed),
154 2010-06-28T17:03:09 <ThomasWaldmann> the storage api maybe should accept name= and uuid= in the api
155 2010-06-28T17:04:47 <ThomasWaldmann> for page-like items, we usually have the name, so we'ld call it with name=u"foo",
156 2010-06-28T17:05:43 <ThomasWaldmann> but e.g. for users, we usually dealt with a user id (-> uuid in moin2), so we could get the user item with uuid="123456"
157 2010-06-28T17:08:15 <ThomasWaldmann> if we have the user's name (e.g. on login), we can just give the name
158 2010-06-28T17:08:36 <ThomasWaldmann> (assuming that user item name == user name in moin2)
159 2010-06-28T17:10:13 <ThomasWaldmann> we could even use this in urls somehow. if you want a permanent link to something that is resistant to renames, you could link to /$uuid/123456
160 2010-06-28T17:14:36 <ThomasWaldmann> (answering that request, it could lookup the name of that item and do a redirect to the current name, so you'ld get a nice url)
161 2010-06-28T17:20:40 <ThomasWaldmann> with a little "link-to" helper for creating wiki-internal links like [[$uuid/123456|ItemName]], we could even have rename-resistant internal links
162 2010-06-28T17:21:17 <TheSheep> but imagine the hell when you switch backends or migrate that to new version in the future
163 2010-06-28T17:21:53 <ThomasWaldmann> you mean if uuid would change?
164 2010-06-28T17:23:11 <ThomasWaldmann> (uuid would be mandatory item metadata and must not change)
165 2010-06-28T17:24:39 <ThomasWaldmann> btw, the backend would not need to be able to work with both name AND uuid, i think it could use whatever it likes
166 2010-06-28T17:24:56 <ThomasWaldmann> (i mean the final, storage backend)
167 2010-06-28T17:25:29 <ThomasWaldmann> the router middleware would use an index to find out whatever was not given
168 2010-06-28T17:26:49 <ThomasWaldmann> (that index for important metadata is already there btw)
169 2010-06-28T17:28:36 <ThomasWaldmann> TheSheep: you understand what i mean?
170 2010-06-28T17:28:37 <TheSheep> that routing middleware is getting fatter every day ;)
171 2010-06-28T17:28:40 <TheSheep> yes
172 2010-06-28T17:28:51 <TheSheep> it sounds like a good idea, but I'm suspicious :)
173 2010-06-28T17:30:34 <ThomasWaldmann> i recently discussed with waldi about userid-named profiles in 1.9 -> username-named items in 2.0
174 2010-06-28T17:31:05 <ThomasWaldmann> those items could have the profile data as item metadata (like now) or even as rev metadata
175 2010-06-28T17:31:21 <ThomasWaldmann> data would be the homepage of the user
176 2010-06-28T17:32:08 <ThomasWaldmann> widgets used to edit rev metadata of user could be also useful for other item's rev metadata
177 2010-06-28T17:32:23 <TheSheep> as long as there are no clashes
178 2010-06-28T17:32:36 <ThomasWaldmann> http://moinmo.in/ThomasWaldmann/Moin2BrainDump took some notes there
179 2010-06-28T17:32:54 <ThomasWaldmann> yes, username duplicates and pages named like users are a problem
180 2010-06-28T17:33:19 <TheSheep> I meant clashes in the metadata key names
181 2010-06-28T17:33:27 <TheSheep> but that can be dealt with prefixes
182 2010-06-28T17:33:51 <TheSheep> I have an impression that you are slowly building zodb ;)
183 2010-06-28T17:34:53 <ThomasWaldmann> i only had a short look at that and no practical experience
184 2010-06-28T17:35:03 <TheSheep> me neither
185 2010-06-28T17:35:12 <TheSheep> just a hunch, you know
186 2010-06-28T17:35:33 <TheSheep> I mean, there must be a database like that already, it's too generic a thing to not exist
187 2010-06-28T17:35:40 <ThomasWaldmann> btw, the little fatter "router with index" can shortcut some operations
188 2010-06-28T17:36:08 <ThomasWaldmann> e.g. the item.exists() can be answered from index
189 2010-06-28T17:36:21 <TheSheep> yeah, it's about the most common opration
190 2010-06-28T17:36:24 <TheSheep> operation
191 2010-06-28T17:37:21 <TheSheep> btw, does the link-making code need to know the item's mime type?
192 2010-06-28T17:37:39 <TheSheep> for example to decide whether to use <a>, <img> or <object>
193 2010-06-28T17:37:50 <TheSheep> if so, it's also a common operation
194 2010-06-28T17:37:52 <ThomasWaldmann> a link is always <a>
195 2010-06-28T17:38:14 <ThomasWaldmann> but for transclusion/include kind of stuff, yes, we need to do a lookup
196 2010-06-28T17:38:46 <TheSheep> include also need the content, so it's slow anyways
197 2010-06-28T17:38:48 <ThomasWaldmann> i cleaned that up after 1.5 :)
198 2010-06-28T17:39:28 <TheSheep> and it's not like you are going to have gazillions of inclusions on a page
199 2010-06-28T17:39:30 <TheSheep> hopefully
200 2010-06-28T17:39:42 <ThomasWaldmann> that is also the reason why a link always links to the show action == the rendered view within theme framework
201 2010-06-28T17:40:15 <ThomasWaldmann> because that is just href(itemname) and does not depend on its type
202 2010-06-28T17:41:12 <ThomasWaldmann> i know users are hating this a bit because they always need &do=get for the direct download, but it has a good reason
203 2010-06-28T17:43:32 <ThomasWaldmann> so, what exactly did you mean with "hell when switching backends etc."?
204 2010-06-28T17:43:52 <ThomasWaldmann> it uuid is in item metadata, a backend is expected to store it
205 2010-06-28T17:43:58 <ThomasWaldmann> if...
206 2010-06-28T17:44:11 <TheSheep> well, I'm afraid to give users yet another syntax option, to put it simly
207 2010-06-28T17:44:13 <TheSheep> simply
208 2010-06-28T17:44:26 <TheSheep> becsuee they *will* use it, and then it has to be supported always
209 2010-06-28T17:44:52 <TheSheep> it's easy to support with the way the backends are planned now, but it may be harder if the ideas change
210 2010-06-28T17:45:04 <TheSheep> and this is limiting
211 2010-06-28T17:45:20 <TheSheep> you know what I mean?
212 2010-06-28T17:45:28 <ThomasWaldmann> yes, vaguely
213 2010-06-28T17:45:43 <TheSheep> maybe this particular feature is not actually that bad, but I'd be careful
214 2010-06-28T17:46:01 <ThomasWaldmann> but if backend is just of the "store this thing" type (less than it does now), is it really a problem?
215 2010-06-28T17:46:02 <TheSheep> like with the query parameters in the [||...] syntax
216 2010-06-28T17:46:24 <ThomasWaldmann> i mean the storage api was (is) rather complex:
217 2010-06-28T17:46:36 <ThomasWaldmann> dealing with history, renames, etc.
218 2010-06-28T17:46:56 <ThomasWaldmann> quite some of that is going away due to the index doing it
219 2010-06-28T17:47:58 <ThomasWaldmann> so likely storage backend just needs to be able to iterate over items and revisions and give their metadata
220 2010-06-28T17:48:17 <TheSheep> yeah
221 2010-06-28T17:48:26 <ThomasWaldmann> this will feed index, index is answering searches, history, etc.
222 2010-06-28T17:48:55 <TheSheep> I just think we should wait with exposing the internals in the user interface (wiki syntax in this case)
223 2010-06-28T17:49:09 <TheSheep> you can always add it when it's actually needed
224 2010-06-28T17:49:16 <ThomasWaldmann> ok
225 2010-06-28T17:49:43 <ThomasWaldmann> syntax could be same btw, it is more a naming convention
226 2010-06-28T17:50:07 <ThomasWaldmann> like "if it starts with $ it is a uuid" or so
227 2010-06-28T17:50:12 <TheSheep> it goes well with the character prefix meaning special things
228 2010-06-28T17:50:21 <TheSheep> + for action, $ for uuid :)
229 2010-06-28T17:50:32 <TheSheep> maybe ~ for user page, to avoid clashes
230 2010-06-28T17:51:24 <ThomasWaldmann> except for JohnSmith
231 2010-06-28T17:51:54 <TheSheep> /~JohnSmith
232 2010-06-28T17:52:17 <ThomasWaldmann> i mean all of them :)
233 2010-06-28T17:52:19 <TheSheep> ah, you mean clashes between user pages and other user pages
234 2010-06-28T17:52:59 <TheSheep> there is always JohnySmith ;)
235 2010-06-28T17:53:02 <ThomasWaldmann> otoh, if they wanted some home page, they would have to deal with same conflict
236 2010-06-28T17:54:07 <TheSheep> they could learn to share ;)
237 2010-06-28T17:56:01 <TheSheep> fortunately I'm not likely to ever have this problem with my name
238 2010-06-28T17:56:32 <ThomasWaldmann> well, if profile data is in user item, they would share quite a lot, even the password and identity
239 2010-06-28T17:57:50 <TheSheep> :P
240 2010-06-28T17:58:15 <TheSheep> but there was this problem with login names before too
241 2010-06-28T17:58:29 <TheSheep> and somehow people managed
242 2010-06-28T18:04:57 *** grzywacz
243 2010-06-28T18:08:19 <CIA-98> Dmitry Andreev <klieg.polar@gmail.com> default * 8321:1d7091afe0ce 2.0-dev-wiki-rest-dandreev/MoinMoin/converter2/rst_in.py: ReStructuredText->DOM: footnote
244 2010-06-28T18:08:20 <CIA-98> Dmitry Andreev <klieg.polar@gmail.com> default * 8322:61062429121f 2.0-dev-wiki-rest-dandreev/MoinMoin/converter2/rst_in.py: ReStructuredText->DOM: basic support of tables and references
245 2010-06-28T18:08:22 <CIA-98> Dmitry Andreev <klieg.polar@gmail.com> default * 8323:1fad573a0147 2.0-dev-wiki-rest-dandreev/MoinMoin/converter2/_tests/test_rst_in.py: ReStructuredText->DOM: more tests
246 2010-06-28T18:08:50 *** DmitryAndreev_
247 2010-06-28T18:13:02 <CIA-98> Diogenes Augusto <diofeher@gmail.com> default * 8476:f53171194373 2.0-dev-jinja2-plugins-dfeher/MoinMoin/theme/jinja.py: [templating] dropping language functions in JinjaTheme
248 2010-06-28T18:17:24 <ThomasWaldmann> diofeher: moin :)
249 2010-06-28T18:17:37 <ThomasWaldmann> btw, did you read about what a clean changeset is?
250 2010-06-28T18:21:49 *** DmitryAndreev
251 2010-06-28T18:31:07 <diofeher> ThomasWaldmann hi thomas
252 2010-06-28T18:31:18 <diofeher> ThomasWaldmann: np, where can i read this?
253 2010-06-28T18:35:00 <TheSheep> diofeher: http://moinmo.in/CodingStyle#MoinMoin_Repository_Commit_Style
254 2010-06-28T18:41:30 <ThomasWaldmann> it is mostly about concentrating on one topic and resisting doing unrelated changes while one is working on that
255 2010-06-28T19:06:55 <dreimark> ei-grad: working?
256 2010-06-28T19:08:45 <dreimark> ThomasWaldmann: i think when we have this name conflict because of migration of ancient moin versions. both names should become disabled
257 2010-06-28T19:10:16 <dreimark> if the name isn't uniq name2id makes wrong matches
258 2010-06-28T19:11:05 <dreimark> so this cache should be cleaned too
259 2010-06-28T19:11:21 <dreimark> (renaming both and disable is also a good option)
260 2010-06-28T19:12:00 <ThomasWaldmann> i don't care about old name2id caches
261 2010-06-28T19:12:27 <ThomasWaldmann> the fs19 code directly reads the old data and extracts what it needs
262 2010-06-28T19:17:12 <ThomasWaldmann> and disabling users would not help, because even a disabled user profile would need storing under some specific name (if we use named item)
263 2010-06-28T19:18:14 <dreimark> hmm ok. the problem exists also for 1.9 or?
264 2010-06-28T19:19:37 <ThomasWaldmann> 1.9 uses the old userid for storing the profile
265 2010-06-28T19:21:02 <dreimark> if one migrates to 1.9 can we get by this two userids with not unique email or name?
266 2010-06-28T19:54:18 <ThomasWaldmann> i didn't look at some details like disabled state yet, but i have same names in multiple profiles
267 2010-06-28T19:59:11 *** kursor
268 2010-06-28T19:59:40 *** ei-grad
269 2010-06-28T19:59:43 *** grzywacz
270 2010-06-28T20:00:06 *** valeuf
271 2010-06-28T20:00:40 *** RogerHaase
272 2010-06-28T20:00:50 *** vpv
273 2010-06-28T20:00:58 *** diofeher
274 2010-06-28T20:01:10 *** waldi
275 2010-06-28T20:01:23 *** dennda
276 2010-06-28T20:01:23 *** aa_
277 2010-06-28T20:01:28 *** gwork
278 2010-06-28T20:01:31 *** eSyr
279 2010-06-28T20:01:46 *** xorAxAx
280 2010-06-28T20:01:47 *** TheSheep
281 2010-06-28T20:01:58 *** dreimark
282 2010-06-28T20:01:58 *** blathijs
283 2010-06-28T20:02:15 *** moinBot`
284 2010-06-28T20:05:44 *** moinBot
285 2010-06-28T20:07:37 *** kursor
286 2010-06-28T20:08:57 <ei-grad> dreimark: working...
287 2010-06-28T20:10:22 *** kursor
288 2010-06-28T20:14:02 * ThomasWaldmann will just use the uuid as user item "name" for now (as a first step), this will cause no new problems
289 2010-06-28T21:06:18 <ThomasWaldmann> diofeher: on what are you working currently?
290 2010-06-28T21:19:52 *** AshishG
291 2010-06-28T21:19:58 <AshishG> 2:0
292 2010-06-28T21:21:02 <AshishG> ThomasWaldmann: you should watch football
293 2010-06-28T21:22:17 <TheSheep> what's the benefit of that?
294 2010-06-28T21:22:57 <AshishG> men after a ball
295 2010-06-28T21:23:07 <AshishG> ok - that's not the best way to put it
296 2010-06-28T21:23:23 <AshishG> TheSheep you don't like football? :(
297 2010-06-28T21:23:51 * TheSheep would rather watch women
298 2010-06-28T21:24:02 <AshishG> hehe
299 2010-06-28T21:24:21 <AshishG> brazillian bums ;)
300 2010-06-28T21:24:30 <AshishG> supporters obviously
301 2010-06-28T21:25:04 <AshishG> TheSheep: which country are you from?
302 2010-06-28T21:27:07 <TheSheep> Poland
303 2010-06-28T21:27:24 <AshishG> hmm
304 2010-06-28T21:32:51 <ThomasWaldmann> there are few things on this planet i care less about than football :)
305 2010-06-28T21:33:21 <ThomasWaldmann> well, maybe some other games involving one ball
306 2010-06-28T21:41:49 <dreimark> isn't it soccer ?
307 2010-06-28T21:42:13 <ThomasWaldmann> doesn't matter, see above :)
308 2010-06-28T21:42:24 <dreimark> oh yeah
309 2010-06-28T21:42:34 *** gwork
310 2010-06-28T21:45:07 <xorAxAx> football is BE, soccer is AE
311 2010-06-28T21:46:01 <dreimark> vuvuzela
312 2010-06-28T21:48:29 * TheSheep prefers less monotonic instruments
313 2010-06-28T21:51:23 *** AshishG
314 2010-06-28T21:52:49 <valeuf> ThomasWaldmann: I am luckier than you, French team suck so much, we are not talking anymore about soccer :p
315 2010-06-28T21:53:01 <valeuf> ThomasWaldmann: It was a big improvement...
316 2010-06-28T21:53:11 <valeuf> ThomasWaldmann: but I guess in germany ... it is not same
317 2010-06-28T21:55:09 *** AshishG
318 2010-06-28T21:59:23 <AshishG> xorAxAx is right
319 2010-06-28T22:00:41 <valeuf> doing roundtrip tests with the HTML converter, make me realizing that few things was completely wrong :s
320 2010-06-28T22:00:55 <valeuf> I hate to realize that I did something stupid...
321 2010-06-28T22:03:11 <dreimark> AshishG: yes there are some other words which differs
322 2010-06-28T22:04:23 <AshishG> oh yeah, they sure do, we here follow (or at least the educations system thinks it does) BE and whenever i go to the US, i hear different words
323 2010-06-28T22:04:25 <AshishG> also the movies
324 2010-06-28T22:04:36 <AshishG> and bush speak - misunderestimate me
325 2010-06-28T22:04:42 <AshishG> :P
326 2010-06-28T22:10:17 <xorAxAx> valeuf: add roundtrip tests
327 2010-06-28T22:10:21 <xorAxAx> valeuf: unittests
328 2010-06-28T22:11:24 <valeuf> xorAxAx: http://hg.moinmo.in/moin/2.0-dev-docbook-html-vjaniaut/rev/062c2b239d6d
329 2010-06-28T22:11:48 <valeuf> xorAxAx: I am working on, but I realized that I did something completely stupid in my unit tests
330 2010-06-28T22:12:45 <valeuf> most of them are wrong, because the " | " means union I thought that it would be equivalent to the "and" logic, but it is a "or" ...
331 2010-06-28T22:13:09 <xorAxAx> ah, i didnt spot that :)
332 2010-06-28T22:13:20 *** grzywacz
333 2010-06-28T22:13:29 <ei-grad> waldi: what did you mean by 'memberOf is an administrative attribute, not accessible by normal users'? are we both talking about MS Active Directory? http://paste.pocoo.org/show/231065/
334 2010-06-28T22:13:32 <valeuf> xorAxAx: so I am rewritting more correctly all my unit tests ...
335 2010-06-28T22:22:34 <diofeher> ThomasWaldmann: i was watching brazil soccer game :)
336 2010-06-28T22:22:50 <diofeher> but i'm refactoring add_msg and msg
337 2010-06-28T22:23:18 *** valeuf
338 2010-06-28T22:24:04 * ei-grad reading http://middleware.internet2.edu/dir/groups/docs/internet2-mace-dir-groups-best-practices-200210.htm , strange I had not found it before
339 2010-06-28T22:24:27 <diofeher> AshishG watching the game?
340 2010-06-28T22:24:32 * ThomasWaldmann will be reviewing the football changesets :P
341 2010-06-28T22:26:05 *** AshishG
342 2010-06-28T22:29:26 <dreimark> diofeher: me is watching cona cabana fans
343 2010-06-28T22:29:40 <dreimark> s/n/b/
344 2010-06-28T22:30:36 <waldi> ei-grad: http://msdn.microsoft.com/en-us/library/ms677099(v=VS.85).aspx
345 2010-06-28T22:30:38 <diofeher> dreimark: cona cabana? what is this?
346 2010-06-28T22:31:16 *** valeuf
347 2010-06-28T22:31:31 <TheSheep> mercurial has grown key-value pairs support that would be sooo useful 2 years ago
348 2010-06-28T22:31:35 <TheSheep> would have been
349 2010-06-28T22:32:58 <valeuf> wow, power cut is worse than everything. it was just during 10 minutes, but I was so stressful ...
350 2010-06-28T22:36:19 <ei-grad> waldi: don't understand what is the problem to use it in our group backend implementation... %(
351 2010-06-28T22:38:22 <ThomasWaldmann> TheSheep: http://paste.pocoo.org/show/231083/
352 2010-06-28T22:38:27 <waldi> ei-grad: several: you have no access and it is not indexes
353 2010-06-28T22:38:32 <waldi> s/s$/d/
354 2010-06-28T22:39:24 <ei-grad> it is available to read, and we are not going to update it...
355 2010-06-28T22:39:35 <ThomasWaldmann> TheSheep: idea is that router gets one of name / uuid, completes it by lookup or auto-assignment and give both name and uuid to lower levels
356 2010-06-28T22:40:05 <waldi> ei-grad: no, it is not
357 2010-06-28T22:40:09 <ThomasWaldmann> in special cases, like xml deserialization, router might even get name AND uuid
358 2010-06-28T22:41:03 <TheSheep> ThomasWaldmann: what if you give both and they don't resolve to the same item? :D
359 2010-06-28T22:41:41 <ThomasWaldmann> i guess the first check will win then :)
360 2010-06-28T22:43:20 <TheSheep> I think it should delete all data
361 2010-06-28T22:43:22 <TheSheep> in such case
362 2010-06-28T22:43:27 <TheSheep> for security
363 2010-06-28T22:43:39 <ThomasWaldmann> or rather: whatever the real backend does. likely a final backend will either use name XOR uuid.
364 2010-06-28T22:44:07 <ThomasWaldmann> :P
365 2010-06-28T22:46:27 <ei-grad> waldi: what do you mean? it is... i don't understand... http://paste.pocoo.org/show/231088/
366 2010-06-28T22:48:03 <ei-grad> even despite the fact that it is not indexed, will be more correct when the database itself provides the entries of members of the group in the one response, than when we make a many requests to get every member entry...
367 2010-06-28T22:49:24 <ei-grad> in second case, we would rather use GreedyGroup (because in fact it will be like greedy group)
368 2010-06-28T22:49:27 <waldi> ei-grad: you are asking the wrong questions
369 2010-06-28T22:52:47 <dreimark> diofeher: coba
370 2010-06-28T22:57:52 <diofeher> dreimark: copacabana? world cup? didn't understood :P
371 2010-06-28T22:57:57 <dreimark> waldi: when is it readable?
372 2010-06-28T22:58:36 <dreimark> diofeher: i watched tv news they showed people watching soccer at copacabana
373 2010-06-28T23:01:05 <diofeher> aaaah yes hehe
374 2010-06-28T23:02:09 <diofeher> everyone stop to work here when brazilian soccer team have game
375 2010-06-28T23:05:34 <waldi> dreimark: as admin
376 2010-06-28T23:06:26 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8487:e058e064678d 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in.py: Fix error in the tests with XPATH union
377 2010-06-28T23:06:28 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8488:5bfd6eec345d 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/html_in.py: Fix typo mistake in the tables for the HTML_IN converter
378 2010-06-28T23:06:30 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8489:98c90a52f80d 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in.py: Log the output from the HTML_IN converter for the unittest
379 2010-06-28T23:06:31 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8490:8ae433d388ea 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in.py: Clean old and useless comment
380 2010-06-28T23:06:53 <waldi> (okay, and the own object)
381 2010-06-28T23:10:08 <CIA-98> Valentin Janiaut <valentin.janiaut@utbm.fr> default * 8491:91ee3e474362 2.0-dev-docbook-html-vjaniaut/MoinMoin/converter2/_tests/test_html_in_out.py: Replace union by and in the unit tests using xpath
382 2010-06-28T23:17:29 <ei-grad> waldi: User1 can read memberOf field of User2 object... what is wrong? http://paste.pocoo.org/show/231108/ %(
383 2010-06-28T23:37:11 <ei-grad> waldi: may be the right question is should I implement the support of nested groups in the way, which it will work only with Active Directory (and possibly with other directory, if it using memberOf attribute)?
384 2010-06-28T23:40:52 <waldi> ei-grad: the first step is to define the questions you can ask. (and in this case find out that moin is asking the wrong ones.)
385 2010-06-28T23:42:45 <waldi> you should be able to define three basic questions to a database of groups, which can be used to answer anything
386 2010-06-28T23:51:24 *** valeuf
387 2010-06-28T23:52:41 <ThomasWaldmann> fsck, i think i changed on wrong level
