IRC log of #zope3-dev for Tuesday, 2008-12-30

*** nyo1 has joined #zope3-dev00:03
*** nyo has quit IRC00:06
*** nyo1 has quit IRC00:08
*** basti has joined #zope3-dev00:17
*** gc_basti has quit IRC00:35
*** jhauser has quit IRC00:38
*** redir is now known as redir|afk00:42
*** aaronv has quit IRC00:45
*** sunew has quit IRC00:45
*** aaronv has joined #zope3-dev00:46
*** sp0cksbeard has quit IRC00:53
*** jpcw has quit IRC01:01
*** lucielejard has quit IRC01:01
*** basti has quit IRC01:01
*** jamur2 has quit IRC01:10
*** yotaff has quit IRC01:16
*** ignas has quit IRC01:21
*** zagy_ has quit IRC01:22
*** davisagli has quit IRC01:27
*** davisagli has joined #zope3-dev01:28
*** lurkymclurkleton has quit IRC01:32
*** yotaff has joined #zope3-dev01:41
*** malthe|bbl is now known as malthe|Zzz02:03
*** davisagli_ has joined #zope3-dev02:07
*** davisagli has quit IRC02:07
*** davisagli_ is now known as davisagli02:08
*** davisagli_ has joined #zope3-dev02:15
*** davisagli has quit IRC02:21
*** davisagli_ is now known as davisagli02:21
*** tarek is now known as tarek|away02:21
*** timte has quit IRC02:26
*** davisagli has quit IRC02:38
*** pelle_ has quit IRC02:38
*** rmarianski has quit IRC02:40
*** aaronv has quit IRC02:48
*** kaeru has joined #zope3-dev02:53
*** davisagli has joined #zope3-dev03:03
*** lisppaste6 has quit IRC03:19
*** lisppaste6 has joined #zope3-dev03:27
*** fmail has joined #zope3-dev03:34
*** fmail has left #zope3-dev03:34
*** quodt has quit IRC03:38
*** harobed has quit IRC03:48
*** natea_ has joined #zope3-dev04:34
*** greenman has quit IRC04:44
*** jfkw has quit IRC06:02
*** greenman has joined #zope3-dev06:07
*** binseer has joined #zope3-dev06:32
*** natea__ has joined #zope3-dev06:37
*** natea_ has quit IRC06:42
*** natea_ has joined #zope3-dev06:43
*** natea___ has joined #zope3-dev06:45
*** natea__ has quit IRC07:02
*** natea_ has quit IRC07:07
*** baijum has joined #zope3-dev07:08
*** JaRoel|4D has quit IRC08:33
*** zagy has joined #zope3-dev09:08
*** zagy_ has joined #zope3-dev09:12
*** mintsauce has joined #zope3-dev09:24
*** zagy has quit IRC09:24
*** JaRoel|4D has joined #zope3-dev09:32
*** JaRoel|4_ has joined #zope3-dev09:35
*** ktwilight_ has joined #zope3-dev09:43
*** fairwinds has quit IRC09:45
*** JaRoel|4_ has quit IRC09:46
*** JaRoel|4_ has joined #zope3-dev09:47
*** JaRoel|4D has quit IRC09:56
*** ktwilight has quit IRC09:56
*** davisagli has left #zope3-dev10:00
*** pelle_ has joined #zope3-dev10:00
*** Theuni has joined #zope3-dev10:16
*** afd__ has joined #zope3-dev10:17
*** MJ has joined #zope3-dev10:38
*** Theuni has quit IRC10:38
*** zagy has joined #zope3-dev11:17
*** tarek|away is now known as tarek11:26
*** nyo has joined #zope3-dev11:28
*** mintsauce has quit IRC11:32
*** tarek is now known as tarek|away11:32
*** zagy_ has quit IRC11:38
*** jhauser has joined #zope3-dev11:49
*** binseer has quit IRC11:59
*** binseer has joined #zope3-dev12:15
*** malthe|Zzz is now known as malthe12:19
*** afd__ has quit IRC12:34
*** afd__ has joined #zope3-dev12:36
*** romanofs1i has joined #zope3-dev12:40
*** romanofs1i has quit IRC12:44
*** afd__ has quit IRC12:44
*** romanofski has quit IRC12:45
*** romanofski has joined #zope3-dev12:46
nyomcdonc: ayt?12:46
*** tarek|away has quit IRC12:46
mcdoncnyo: hey12:46
*** afd__ has joined #zope3-dev12:48
nyomcdonc: need to decide on IIndexSort interface. Should the sort method return an _iterable_ (like generator) or require them to return a finite sequence. I wonder if there can be any problems (performance or programming ones) with returning sequences from the ``sort`` method of some indexes.12:48
mcdoncnyo: the interface should probably make as few promises about this as possible.. just that it be iterable12:49
nyomcdonc: currently, for example, the z.a.catalog needs a sequence, so it has to manually do the list(index.sort(...)) thing, but we could just require the sort method to return a sequence12:49
mcdoncwhy does z.a.catalog need a sequence?12:50
nyomcdonc: because it needs to know length12:50
mcdoncone thing i thought of12:51
nyomcdonc: the ResultSet object that catalog's searchResults returns has the __len__ object12:51
mcdoncand this would make it possible to not need the length12:51
mcdoncinstead of throwing away docids that dont have a value in the sort index...12:51
mcdoncmaybe the IIndexSort interface can take a flag12:51
mcdoncsort_unknown_bottom=True (bad name)12:52
mcdoncbut essentially a flag that tells the sort method how to sort docids that dont have a value in the sort index12:52
mcdoncwhether they should be at the bottom or the top12:53
mcdoncso the length of the iterable you get back from the sort method would be guaranteed to be the length of the set you passed in to sort12:53
mcdonc(and then you dont need to recompute length, you'd have it already)12:53
nyomcdonc: I don't think it's possible anyway. it will break ResultSet. it relies on the sequence nature of uid list12:54
*** jpcw has joined #zope3-dev12:54
nyomcdonc: if there will be a generator, it will work only once for example12:54
nyo(iteration, i mean)12:55
mcdoncthat's mostly resultset's problem i think12:55
mcdoncrandom access into the result set is a higher-level thing12:55
mcdoncand unwinding the generator if its necessary is probably the right thing12:56
mcdoncrandom access into a result set is sort of nonsensical for most apps i think12:56
nyomcdonc: so you think we need to declare it just as iterable as it's now?12:57
mcdoncnyo: i think so... at least avoid promising that it will be a fully materialized random-acessible thing... does that make sense?12:57
nyomcdonc: yep12:58
mcdoncwhat do you think of the dont-throw-away-values thing12:58
nyomcdonc: appending docids that is not in the index to the top/bottom can be quite hard to implement12:58
nyoin some cases )12:58
mcdoncmaybe yeah12:58
mcdonci havent done it yet12:59
mcdoncit sounds good ;-)12:59
nyoalso, I don't think there will be much cases where it's needed12:59
mcdonci think it'll always be bottom anyway12:59
mcdoncif it were to be implemented12:59
mcdonci dont want random stuff at the top of my result set12:59
mcdoncso i think it's fine to avoid it for now in the interface13:00
nyomcdonc: btw, even FieldIndex can't do that nicely when using set intersections13:00
nyomcdonc: ok13:00
mcdoncthanks for taking charge of this ;-)13:01
nyomcdonc: I'll commit keywordindex backward-incompatibility breakage now (normalization thing, we talked about), can you review?13:01
*** greenman has quit IRC13:07
*** harobed has joined #zope3-dev13:10
*** pelle_ has quit IRC13:11
nyomcdonc: Done. I also wrote an explanation in CHANGES.txt, please review that as well.13:12
nyomcdonc: I'm afk for some time13:12
mcdoncnyo: ok... will do13:13
*** toutpt has joined #zope3-dev13:16
mcdoncnyo: would you be willing to reverse the arg order of limit and reverse in IIndexSort?  it'd help me not need to subclass to service existing code13:31
nyomcdonc: sort(docids, reverse=False, limit=None)? of course. that will do more sense even:)13:41
*** tarek|away has joined #zope3-dev13:42
mcdoncnyo: yes, great...13:42
nyomcdonc: done13:43
mcdoncthank you!13:43
nyoso I'm going to release it. others seems to either don't have any objections or are not interested (so they also don't have objections).13:44
mcdoncnyo: go for it13:44
mcdoncthis is factored really nicely... its great we can both use the same index impls but different catalog impls13:46
nyomcdonc: yeah, they can also be used just separately or with z3c.indexer13:47
* mcdonc reads
mcdoncgood idea...13:50
mcdoncyes, i suspect that being able to spell the ordering of which indexes to intersect in what order would help performance too... z.a.catalog and repoze.catalog dont allow for this13:51
nyomcdonc: yeah, it also does index-on-commit so it does less indexing-unindexing when using event-based indexing13:52
*** tarek|away has quit IRC13:53
mcdoncthat's awesome13:54
mcdonci'm so going to steal that ;-)13:55
nyomcdonc: okay, 3.5.0 is there13:59
mcdoncnyo: cool, thanks13:59
*** tarek|away has joined #zope3-dev14:20
*** Jell-O-Fishi has quit IRC14:28
*** goschtl has joined #zope3-dev14:41
*** tarek|away has quit IRC14:47
*** tarek|away has joined #zope3-dev14:48
*** timte has joined #zope3-dev15:11
*** tarek|away has quit IRC15:16
*** jayaraj has joined #zope3-dev15:21
*** fairwinds has joined #zope3-dev15:22
*** malthe is now known as malthe|away15:42
*** romanofski has quit IRC15:48
goschtlhi has anyone an example of a custom IGroup implementation?15:49
nyoany1 interested in reviewing "nadako-sorting" branch of :)15:51
*** kaeru has quit IRC15:52
*** lurkymclurkleton has joined #zope3-dev15:57
*** lurkymclurkleton has joined #zope3-dev15:58
*** jayaraj has quit IRC16:02
*** jamur2 has joined #zope3-dev16:02
*** jayaraj has joined #zope3-dev16:03
*** sp0cksbeard has joined #zope3-dev16:09
*** binseer has quit IRC16:35
*** sawdog has joined #zope3-dev16:40
*** zagy has quit IRC16:43
*** zagy has joined #zope3-dev16:43
*** redir|afk is now known as redir16:44
*** baijum has quit IRC16:52
*** aaronv has joined #zope3-dev16:56
*** jayaraj has quit IRC16:56
*** aaronv has quit IRC16:57
*** kaeru has joined #zope3-dev16:58
*** zagy has quit IRC16:58
*** zagy has joined #zope3-dev16:58
*** natea___ has quit IRC17:00
*** MJ has quit IRC17:03
*** Theuni has joined #zope3-dev17:11
*** natea_ has joined #zope3-dev17:23
*** reco has joined #zope3-dev17:36
*** natea_ has quit IRC17:57
*** goschtl has quit IRC17:58
*** jukart has joined #zope3-dev18:09
*** jukart has quit IRC18:09
*** nyo has quit IRC18:20
*** iham is now known as iham|away18:23
*** natea_ has joined #zope3-dev18:42
*** JaRoel|4_ has quit IRC18:45
*** JaRoel|4D has joined #zope3-dev18:45
*** davisagli has joined #zope3-dev18:53
*** tarek|away has joined #zope3-dev19:03
*** Theuni has quit IRC19:08
*** tarek|away is now known as tarek19:15
*** dbfrombrc|away has quit IRC19:25
*** JaRoel|4D has quit IRC19:30
*** reco_ has joined #zope3-dev19:33
*** nyo has joined #zope3-dev19:37
*** toutpt has quit IRC19:41
*** tarek has quit IRC19:48
*** reco has quit IRC19:59
*** natea_ has quit IRC19:59
*** reco has joined #zope3-dev20:03
*** JaRoel|4D has joined #zope3-dev20:14
*** reco_ has quit IRC20:20
*** lucielejard has joined #zope3-dev20:27
*** tarek has joined #zope3-dev20:28
*** pelle_ has joined #zope3-dev20:32
*** tarek has quit IRC20:38
*** iham|away has quit IRC20:41
*** jfkw has joined #zope3-dev20:44
*** afd__ has quit IRC20:56
*** yotaff has quit IRC21:13
*** sunew has joined #zope3-dev21:21
*** tarek has joined #zope3-dev21:27
*** tarek has quit IRC21:55
*** greenman has joined #zope3-dev22:00
*** greenman has quit IRC22:04
*** jhauser has quit IRC22:10
*** MrSipan has joined #zope3-dev22:26
*** aaronv has joined #zope3-dev22:29
*** jfkw has quit IRC22:48
*** zagy has quit IRC22:52
*** greenman has joined #zope3-dev22:52
*** redir` has joined #zope3-dev23:16
*** MrSipan has quit IRC23:17
*** pelle_ has quit IRC23:17
*** sp0cksbeard has quit IRC23:17
*** timte has quit IRC23:17
*** harobed has quit IRC23:17
*** BjornT has quit IRC23:17
*** aclark has quit IRC23:17
*** redir has quit IRC23:17
*** aclark has joined #zope3-dev23:18
*** BjornT has joined #zope3-dev23:18
*** wgrant has joined #zope3-dev23:19
*** harobed has joined #zope3-dev23:21
*** pelle_ has joined #zope3-dev23:22
*** MrSipan has joined #zope3-dev23:22
*** davisagli has quit IRC23:22
*** timte has joined #zope3-dev23:23
*** dbfrombrc has joined #zope3-dev23:28
*** zagy has joined #zope3-dev23:28
*** acsr has joined #zope3-dev23:39
*** brandon_rhodes has joined #zope3-dev23:46
*** dbfrombrc has quit IRC23:54
*** acsr_ has quit IRC23:54
*** calimero has joined #zope3-dev23:57

Generated by 2.15.1 by Marius Gedminas - find it at!