MoinMoin: MoinMoinChat/Logs/moin-dev/2011-09-27

   1 2011-09-27T06:13:59  *** MattMaker has quit IRC
   2 2011-09-27T06:14:08  *** MattMaker has joined #moin-dev
   3 2011-09-27T07:47:29  *** raignarok_ has joined #moin-dev
   4 2011-09-27T07:53:41  *** raignarok_ has quit IRC
   5 2011-09-27T08:24:59  <dreimark> moin
   6 2011-09-27T08:25:03  <dreimark> ThomasWaldmann: http://codereview.appspot.com/5129052/
   7 2011-09-27T08:25:07  <dreimark> bbl
   8 2011-09-27T08:27:06  <dreimark> ronny: ^^
   9 2011-09-27T08:43:18  <ronny> dreimark: i suppose there is need for a filter term for that, ThomasWaldmann should have a better grasp on how to do that with whoosh
  10 2011-09-27T09:58:03  <dreimark> i had another idea for that bookmark, not sure if ng can do that, but can we use a revision id instead of the date ?
  11 2011-09-27T09:58:33  <dreimark> another related question is if we can perform a search until this revision?
  12 2011-09-27T09:59:18  <dreimark> this probably can be used to solve the pagination problem of changed content within search
  13 2011-09-27T09:59:32  <ronny> dreimark: im not sure what you mean now
  14 2011-09-27T09:59:58  <dreimark> the pagination code in history needs the whole information and splits it afterwards
  15 2011-09-27T10:00:54  <dreimark> by this you know the length of day entries and days
  16 2011-09-27T10:01:36  <dreimark> if you change something from a older day then a lot of lines can change
  17 2011-09-27T10:02:18  <dreimark> so the solution for that is to perform a search until the revision in previoous search used
  18 2011-09-27T10:02:31  <dreimark> so no updates in that "old" request
  19 2011-09-27T10:02:51  <ronny> i dont get it at all, im missing some context
  20 2011-09-27T10:06:30  * dreimark tries another way around
  21 2011-09-27T10:06:57  <dreimark> ronny: if I do hg log, i get a numerical id: hex id
  22 2011-09-27T10:07:13  <dreimark> the numerical id describes a linearity of following commits
  23 2011-09-27T10:07:28  <dreimark> do we have in ng something like this numerical id
  24 2011-09-27T10:07:59  <ronny>  dreimark no
  25 2011-09-27T10:08:21  <ronny> dreimark: the numeric id's are a implementation detail of revlog
  26 2011-09-27T10:21:03  <ThomasWaldmann> moin
  27 2011-09-27T10:24:01  <ThomasWaldmann> dreimark: using a revid for the bookmark is a bit pointless, becaus revids have no order
  28 2011-09-27T10:25:21  <ThomasWaldmann> and using a datetime, we can do a range search with whoosh
  29 2011-09-27T10:25:31  <ThomasWaldmann> on the MTIME field
  30 2011-09-27T10:25:51  <dreimark> hg has a commit order
  31 2011-09-27T10:26:34  <ThomasWaldmann> we don't use hg, not even a revlog
  32 2011-09-27T10:26:37  <dreimark> range search solves it too
  33 2011-09-27T10:27:47  <ThomasWaldmann> what you need for g.h. is something like search(all_revs_index, RangeTerm(bookmark, now), sortedby=MTIME)
  34 2011-09-27T10:28:46  <ThomasWaldmann> what you need for g.h. is something like search(all_revs_index, RangeTerm(MTIME, (bookmark, now)), sortedby=MTIME)
  35 2011-09-27T10:28:49  <dreimark> hmm, i also need remember time of first RangeTerm call
  36 2011-09-27T10:29:22  <dreimark> if we use the range for paginating results
  37 2011-09-27T10:29:37  <ThomasWaldmann> and all that paging stuff needs more thought
  38 2011-09-27T10:30:31  <dreimark> how can i exclude user items from my query?
  39 2011-09-27T10:33:27  <ThomasWaldmann> try without doing that
  40 2011-09-27T10:35:14  <dreimark> it crashed yesterday because of missing mtime
  41 2011-09-27T10:37:15  <ThomasWaldmann> well, add mtime to user profile data :)
  42 2011-09-27T10:38:37  <dreimark> ah, ok :)
  43 2011-09-27T10:38:59  <dreimark> I asked that yesterday too - and the answer was different
  44 2011-09-27T10:40:04  <dreimark> i guess i misunderstand that
  45 2011-09-27T11:38:41  *** greg_f has joined #moin-dev
  46 2011-09-27T12:37:54  <CIA-65> Thomas Waldmann <tw AT waldmann-edv DOT de> storage-ng * 894:93ae039cb363 2.0/MoinMoin/ (9 files in 5 dirs):
  47 2011-09-27T12:37:54  <CIA-65> fixed indexable and diff views, use revids not revnos for diffs and urls
  48 2011-09-27T12:37:54  <CIA-65> added revid attr and read/seek/tell methods to PseudoRevision
  49 2011-09-27T12:54:38  <CIA-65> Thomas Waldmann <tw AT waldmann-edv DOT de> storage-ng * 895:119cb9b352af 2.0/MoinMoin/user.py: add MTIME to user item revision metadata
  50 2011-09-27T12:54:40  <ThomasWaldmann> dreimark: ^^
  51 2011-09-27T12:56:16  *** mkerrin has joined #moin-dev
  52 2011-09-27T14:08:50  <dreimark> ThomasWaldmann: thanks
  53 2011-09-27T14:10:03  * dreimark is busy with some different computer stuff currently, will continue in some hours with the global_history tests
  54 2011-09-27T18:52:48  *** mkerrin has left #moin-dev
  55 2011-09-27T19:06:10  *** greg_f has quit IRC
  56 2011-09-27T20:02:59  <ThomasWaldmann> re
  57 2011-09-27T21:59:19  <dreimark> something has broken diff http://paste.pocoo.org/show/483494/
  58 2011-09-27T22:00:32  <CIA-65> Thomas Waldmann <tw AT waldmann-edv DOT de> storage-ng * 896:7a01ee71dc4c 2.0/MoinMoin/ (items/__init__.py templates/diff_text.html): diff view: remove revision navigation (does not work like this, was based on revno integers)
  59 2011-09-27T22:02:17  <ThomasWaldmann> you need to add a : in the url in front of revids
  60 2011-09-27T22:02:39  <ThomasWaldmann> that's a router issue i wanted to discuss with ronny
  61 2011-09-27T22:03:52  <ThomasWaldmann> dreimark: ^^
  62 2011-09-27T22:03:55  <dreimark> ok
  63 2011-09-27T22:05:40  <dreimark> currently i have user items in the history, and they have always the same time as the bookmark function
  64 2011-09-27T22:05:52  <dreimark> so they are always visible
  65 2011-09-27T22:28:33  <ThomasWaldmann> just ignore that for now
  66 2011-09-27T22:34:29  <CIA-65> Reimar Bauer <rb.proj AT googlemail DOT com> storage-ng * 897:42662a463f0b 2.0/MoinMoin/ (apps/frontend/views.py templates/global_history.html): fix global_history view
  67 2011-09-27T22:40:03  * ThomasWaldmann tries
  68 2011-09-27T22:41:36  * dreimark fixes diff on the rihgt place
  69 2011-09-27T22:42:04  <ThomasWaldmann> bug in 778
  70 2011-09-27T22:42:23  <ThomasWaldmann> %d format does not work any more as it is no integer
  71 2011-09-27T22:43:34  <dreimark> true
  72 2011-09-27T22:44:05  <dreimark> i fix that, in a moment
  73 2011-09-27T22:53:18  <CIA-65> Reimar Bauer <rb.proj AT googlemail DOT com> storage-ng * 898:32c57b84512f 2.0/MoinMoin/apps/frontend/views.py: moved prefix of ':' for diffs to diff
  74 2011-09-27T22:53:19  <CIA-65> Reimar Bauer <rb.proj AT googlemail DOT com> storage-ng * 899:c2a3e35cdc2f 2.0/MoinMoin/apps/frontend/views.py: global_history revno for comment fixed
  75 2011-09-27T22:54:10  <ThomasWaldmann> ehrm
  76 2011-09-27T22:54:21  <ThomasWaldmann> that : is rather a bug than a feature
  77 2011-09-27T22:54:42  <ThomasWaldmann> but just let it be like it is now, we need to fix that later
  78 2011-09-27T22:56:12  <dreimark> i know, it is better to kill it only on one place
  79 2011-09-27T23:03:28  <ThomasWaldmann> better now :)
  80 2011-09-27T23:09:40  * dreimark looks at the download from index view
  81 2011-09-27T23:18:42  <CIA-65> Thomas Waldmann <tw AT waldmann-edv DOT de> storage-ng * 900:24d7fb1b59c9 2.0/MoinMoin/apps/frontend/views.py: fix tags
  82 2011-09-27T23:43:03  <ThomasWaldmann> hmm, history does now a lookup based on the name
  83 2011-09-27T23:44:09  <ThomasWaldmann> that means a rename breaks the history display
  84 2011-09-27T23:44:40  <ThomasWaldmann> "breaks" in the sense of same-item history
  85 2011-09-27T23:44:56  <ThomasWaldmann> it would show other items that had same name, though
  86 2011-09-27T23:45:33  <ThomasWaldmann> if we would do a itemid lookup and then another lookup based on the itemid, we would get same-item history, including renames