2008-03-21T00:00:11 <lanius> maybe some getLayerCount method in the layered backend?
2008-03-21T00:00:11 <lanius> maybe some getLayerCount method in the layered backend?
2008-03-21T00:00:17 <johill> useless
2008-03-21T00:00:17 <johill> useless
2008-03-21T00:00:20 <lanius> why
2008-03-21T00:00:20 <lanius> why
2008-03-21T00:00:21 <johill> then you can't search for that
2008-03-21T00:00:21 <johill> then you can't search for that
2008-03-21T00:00:31 <lanius> yes
2008-03-21T00:00:31 <johill> you need to be able to express
2008-03-21T00:00:31 <johill> you need to be able to express
2008-03-21T00:00:31 <lanius> yes
2008-03-21T00:00:38 <johill> "give me a list of pages not from the lowest layer"
2008-03-21T00:00:38 <johill> "give me a list of pages not from the lowest layer"
2008-03-21T00:00:41 <lanius> but it belongs on the layer level
2008-03-21T00:00:42 <lanius> but it belongs on the layer level
2008-03-21T00:00:47 <lanius> and only there
2008-03-21T00:00:47 <lanius> and only there
2008-03-21T00:01:23 <johill> ok I can agree with that
2008-03-21T00:01:23 <johill> ok I can agree with that
2008-03-21T00:01:37 <johill> and remove is_underlay in favour of having an argument to the layer backend marking some as underlay
2008-03-21T00:01:37 <johill> and remove is_underlay in favour of having an argument to the layer backend marking some as underlay
2008-03-21T00:01:50 <johill> maybe layerbackend.add() and layerbackend.addUnderlay()
2008-03-21T00:01:50 <johill> maybe layerbackend.add() and layerbackend.addUnderlay()
2008-03-21T00:02:30 <lanius> the second backend should be automatically marked underlay
2008-03-21T00:02:30 <lanius> the second backend should be automatically marked underlay
2008-03-21T00:02:34 <johill> no
2008-03-21T00:02:34 <johill> no
2008-03-21T00:02:46 <johill> no magic here
2008-03-21T00:02:46 <johill> no magic here
2008-03-21T00:02:50 <lanius> it's no magic
2008-03-21T00:02:51 <lanius> it's no magic
2008-03-21T00:02:57 <lanius> there is one backend on top of the other
2008-03-21T00:02:58 <lanius> there is one backend on top of the other
2008-03-21T00:03:01 <johill> what if I want four layers
2008-03-21T00:03:01 <johill> what if I want four layers
2008-03-21T00:03:02 <lanius> and the others are "underlay"
2008-03-21T00:03:03 <lanius> and the others are "underlay"
2008-03-21T00:03:20 <johill> technically
2008-03-21T00:03:20 <johill> technically
2008-03-21T00:03:26 <lanius> or remove the idea of the one and only underlay
2008-03-21T00:03:27 <lanius> or remove the idea of the one and only underlay
2008-03-21T00:03:27 <johill> but not how that term is used in Moin
2008-03-21T00:03:27 <johill> but not how that term is used in Moin
2008-03-21T00:03:34 <lanius> have to look into the code again where this isUnderlayPage is used
2008-03-21T00:03:35 <lanius> have to look into the code again where this isUnderlayPage is used
2008-03-21T00:03:38 <lanius> maybe we can remove it at all
2008-03-21T00:03:39 <lanius> maybe we can remove it at all
2008-03-21T00:03:40 <johill> mostly search
2008-03-21T00:03:40 <johill> mostly search
2008-03-21T00:04:38 <johill> so anyway
2008-03-21T00:04:38 <johill> so anyway
2008-03-21T00:04:45 <johill> the layer can either be metadata
2008-03-21T00:04:45 <johill> the layer can either be metadata
2008-03-21T00:04:49 <johill> or we need a special search term for it
2008-03-21T00:04:49 <johill> or we need a special search term for it
2008-03-21T00:05:07 <lanius> why is it important where a page lives?
2008-03-21T00:05:07 <lanius> why is it important where a page lives?
2008-03-21T00:05:19 <johill> to not search through help pages when you don't want to
2008-03-21T00:05:20 <johill> to not search through help pages when you don't want to
2008-03-21T00:05:40 <lanius> ok special search termn in our new filter stuff which is executed by the layer backend
2008-03-21T00:05:40 <lanius> ok special search termn in our new filter stuff which is executed by the layer backend
2008-03-21T00:06:02 <johill> yeah we need to design that carefully
2008-03-21T00:06:02 <johill> yeah we need to design that carefully
2008-03-21T00:06:04 <johill> mabye something like
2008-03-21T00:06:04 <johill> mabye something like
2008-03-21T00:06:14 <johill> backend.execute_search_term(searchobject)
2008-03-21T00:06:14 <johill> backend.execute_search_term(searchobject)
2008-03-21T00:06:33 <johill> which itself does searchobject.execute() if it doesn't know about the search object
2008-03-21T00:06:33 <johill> which itself does searchobject.execute() if it doesn't know about the search object
2008-03-21T00:07:10 <johill> but that's a bit backwards
2008-03-21T00:07:10 <johill> but that's a bit backwards
2008-03-21T00:07:14 <johill> we have
2008-03-21T00:07:14 <johill> we have
2008-03-21T00:07:19 <johill> backend.list_items(searchterm)
2008-03-21T00:07:19 <johill> backend.list_items(searchterm)
2008-03-21T00:07:30 <lanius> which could be replaced by the searchobject
2008-03-21T00:07:30 <lanius> which could be replaced by the searchobject
2008-03-21T00:07:33 <johill> where searchterm is something like MoinMoin.search.AndExpression,...
2008-03-21T00:07:33 <johill> where searchterm is something like MoinMoin.search.AndExpression,...
2008-03-21T00:07:45 <johill> yeah
2008-03-21T00:07:45 <johill> yeah
2008-03-21T00:07:58 <johill> I'm using that interchangeably, I don't want a parser down at that level
2008-03-21T00:07:58 <johill> I'm using that interchangeably, I don't want a parser down at that level
2008-03-21T00:08:20 <lanius> ok
2008-03-21T00:08:21 <lanius> ok
2008-03-21T00:08:26 <johill> so say you get an AND()
2008-03-21T00:08:26 <johill> so say you get an AND()
2008-03-21T00:08:42 <johill> and have an sql backend
2008-03-21T00:08:42 <johill> and have an sql backend
2008-03-21T00:09:02 <johill> then you look through the AND-terms to see if any is a TitleMatch-term
2008-03-21T00:09:02 <johill> then you look through the AND-terms to see if any is a TitleMatch-term
2008-03-21T00:09:30 <johill> remove that and execute it directly
2008-03-21T00:09:30 <johill> remove that and execute it directly
2008-03-21T00:09:44 <lanius> yes
2008-03-21T00:09:44 <lanius> yes
2008-03-21T00:10:11 <johill> the remaining expression you execute for each item
2008-03-21T00:10:11 <johill> the remaining expression you execute for each item
2008-03-21T00:10:15 <lanius> yes
2008-03-21T00:10:15 <lanius> yes
2008-03-21T00:10:18 <johill> now
2008-03-21T00:10:18 <johill> now
2008-03-21T00:10:34 <johill> normally each expression has its own implementation in any case
2008-03-21T00:10:34 <johill> normally each expression has its own implementation in any case
2008-03-21T00:10:48 <johill> e.g. AND will just do that, TitleMatch will actually do the title match etc
2008-03-21T00:10:48 <johill> e.g. AND will just do that, TitleMatch will actually do the title match etc
2008-03-21T00:11:08 <lanius> right
2008-03-21T00:11:08 <lanius> right
2008-03-21T00:11:09 <johill> but what would LayerMatch do?
2008-03-21T00:11:10 <johill> but what would LayerMatch do?
2008-03-21T00:11:33 <lanius> only call the list_item methods from the speicifed layers
2008-03-21T00:11:33 <lanius> only call the list_item methods from the speicifed layers
2008-03-21T00:11:40 <johill> no no
2008-03-21T00:11:40 <johill> no no
2008-03-21T00:11:54 <johill> each search-expression is executed for a single object
2008-03-21T00:11:54 <johill> each search-expression is executed for a single object
2008-03-21T00:12:01 <johill> after the backend has removed t hose it can execute
2008-03-21T00:12:01 <johill> after the backend has removed t hose it can execute
2008-03-21T00:12:25 <lanius> but the item itself does not know to which backend it belongs
2008-03-21T00:12:25 <lanius> but the item itself does not know to which backend it belongs
2008-03-21T00:12:33 <johill> excatly the problem
2008-03-21T00:12:33 <johill> excatly the problem
2008-03-21T00:12:44 <lanius> so why not match it like i said in this case
2008-03-21T00:12:44 <lanius> so why not match it like i said in this case
2008-03-21T00:13:00 <johill> match it like you said?
2008-03-21T00:13:00 <johill> match it like you said?
2008-03-21T00:13:08 <johill> there are two cases
2008-03-21T00:13:08 <johill> there are two cases
2008-03-21T00:13:12 <lanius> only call the list_item methods from the speicifed layers
2008-03-21T00:13:13 <lanius> only call the list_item methods from the speicifed layers
2008-03-21T00:13:18 <johill> if the LayerMatch is easy to find in the whole expression
2008-03-21T00:13:18 <johill> if the LayerMatch is easy to find in the whole expression
2008-03-21T00:13:30 <johill> then the LayerBackend can just remove it from the expression and do exactly that
2008-03-21T00:13:30 <johill> then the LayerBackend can just remove it from the expression and do exactly that
2008-03-21T00:14:00 <johill> but if it's, say, Xor(TitleMatch('Page.*'), LayerMatch(1)) that won't work
2008-03-21T00:14:00 <johill> but if it's, say, Xor(TitleMatch('Page.*'), LayerMatch(1)) that won't work
2008-03-21T00:14:08 <lanius> hm, right
2008-03-21T00:14:08 <lanius> hm, right
2008-03-21T00:14:15 <johill> I suppose we just pass the backend
2008-03-21T00:14:16 <johill> I suppose we just pass the backend
2008-03-21T00:14:20 <johill> and the item
2008-03-21T00:14:20 <johill> and the item
2008-03-21T00:14:30 <johill> then the LayerMatch can do something like if (isinstancde(backend, LayerBackend))
2008-03-21T00:14:30 <johill> then the LayerMatch can do something like if (isinstancde(backend, LayerBackend))
2008-03-21T00:15:04 <lanius> that could be done
2008-03-21T00:15:04 <lanius> that could be done
2008-03-21T00:15:13 <johill> only sane way I guess
2008-03-21T00:15:24 <johill> but a special case
2008-03-21T00:15:27 <johill> only sane way I guess
2008-03-21T00:15:27 <johill> but a special case
2008-03-21T00:15:38 <lanius> no other idea from me yet
2008-03-21T00:15:38 <lanius> no other idea from me yet
2008-03-21T00:16:20 <johill> no, sounds about right
2008-03-21T00:16:20 <johill> no, sounds about right
2008-03-21T00:16:30 <lanius> but all filters are special cases for their own
2008-03-21T00:16:30 <lanius> but all filters are special cases for their own
2008-03-21T00:16:35 <lanius> so i don't see a problem there
2008-03-21T00:16:36 <lanius> so i don't see a problem there
2008-03-21T00:16:50 <johill> well, most operate only on the item not the backend
2008-03-21T00:16:51 <johill> well, most operate only on the item not the backend
2008-03-21T00:16:59 <lanius> most is not a rule ;)
2008-03-21T00:16:59 <lanius> most is not a rule ;)
2008-03-21T00:17:13 <johill> yeah
2008-03-21T00:17:13 <johill> yeah
2008-03-21T00:17:28 <johill> but I wouldn't want to have xapian specific code in the search expression objects
2008-03-21T00:17:28 <johill> but I wouldn't want to have xapian specific code in the search expression objects
2008-03-21T00:17:31 <johill> like it is the case now
2008-03-21T00:17:31 <johill> like it is the case now
2008-03-21T00:17:48 <johill> that would be smarter handled by walking the expression tree in the xapian backend and handling what can be handled
2008-03-21T00:17:48 <johill> that would be smarter handled by walking the expression tree in the xapian backend and handling what can be handled
2008-03-21T00:17:49 <lanius> what is xapian specific in there now?
2008-03-21T00:17:50 <lanius> what is xapian specific in there now?
2008-03-21T00:18:02 <johill> over half the code
2008-03-21T00:18:02 <johill> over half the code
2008-03-21T00:18:15 <johill> each search expression can generate its own xapian equivalent
2008-03-21T00:18:15 <johill> each search expression can generate its own xapian equivalent
2008-03-21T00:18:24 <johill> rather than having one place that does that for all expression objects
2008-03-21T00:18:24 <johill> rather than having one place that does that for all expression objects
2008-03-21T00:18:25 <lanius> can
2008-03-21T00:18:25 <lanius> can
2008-03-21T00:18:34 <johill> yeah that's bogus
2008-03-21T00:18:34 <johill> yeah that's bogus
2008-03-21T00:18:40 <xorAxAx> visitor pattern is oldskoool
2008-03-21T00:18:40 <xorAxAx> visitor pattern is oldskoool
2008-03-21T00:18:41 <lanius> but it also can use other mechanisms
2008-03-21T00:18:41 <lanius> but it also can use other mechanisms
2008-03-21T00:19:03 <johill> lanius: yeah but think if you want any other indexer too then you end up .... don't want to think about it
2008-03-21T00:19:06 <johill> xorAxAx: yeah....
2008-03-21T00:19:06 <xorAxAx> better do the aspect oriented chaos insertion with copy&paste
2008-03-21T00:19:10 <johill> lanius: yeah but think if you want any other indexer too then you end up .... don't want to think about it
2008-03-21T00:19:10 <johill> xorAxAx: yeah....
2008-03-21T00:19:10 <xorAxAx> better do the aspect oriented chaos insertion with copy&paste
2008-03-21T00:19:27 <johill> I th ink I have a good plan now :)
2008-03-21T00:19:27 <johill> I th ink I have a good plan now :)
2008-03-21T00:19:58 <xorAxAx> and in the end moin will contain a linux kernel with fulltext search and plan9 file storage
2008-03-21T00:19:58 <xorAxAx> and in the end moin will contain a linux kernel with fulltext search and plan9 file storage
2008-03-21T00:20:03 <johill> anybody know how to set up xapian?
2008-03-21T00:20:03 <johill> anybody know how to set up xapian?
2008-03-21T00:20:12 <ThomasWaldmann> hehe
2008-03-21T00:20:12 <ThomasWaldmann> hehe
2008-03-21T00:20:13 <johill> don't forget a full 802.11 stack
2008-03-21T00:20:13 <johill> don't forget a full 802.11 stack
2008-03-21T00:20:20 <johill> my specialty ;)
2008-03-21T00:20:20 <johill> my specialty ;)
2008-03-21T00:20:32 <lanius> why not just boot moin instead of any other crap os
2008-03-21T00:20:33 <lanius> why not just boot moin instead of any other crap os
2008-03-21T00:20:53 <ThomasWaldmann> unununium os :)
2008-03-21T00:20:53 <ThomasWaldmann> unununium os :)
2008-03-21T00:21:19 * johill orders RAM upgrade
2008-03-21T00:21:19 * johill orders RAM upgrade
2008-03-21T00:21:31 <lanius> there is nothing that can replace RAM but more RAM?
2008-03-21T00:21:31 <lanius> there is nothing that can replace RAM but more RAM?
2008-03-21T00:21:59 <johill> hrm this is weird
2008-03-21T00:21:59 <johill> hrm this is weird
2008-03-21T00:22:10 <johill> why can a search term be negated when there's a NOT expression too??
2008-03-21T00:22:10 <johill> why can a search term be negated when there's a NOT expression too??
2008-03-21T00:22:32 <xorAxAx> optimisation?
2008-03-21T00:22:32 <xorAxAx> optimisation?
2008-03-21T00:22:43 <johill> oh wait there is no Not expression
2008-03-21T00:22:43 <johill> oh wait there is no Not expression
2008-03-21T00:22:45 <johill> ewww
2008-03-21T00:22:45 <johill> ewww
2008-03-21T00:22:55 <xorAxAx> well, makes perfect sense
2008-03-21T00:22:55 <xorAxAx> well, makes perfect sense
2008-03-21T00:23:10 <lanius> ^^
2008-03-21T00:23:10 <lanius> ^^
2008-03-21T00:23:15 <johill> yeah, copy&paste the code into each possible expression
2008-03-21T00:23:15 <johill> yeah, copy&paste the code into each possible expression
2008-03-21T00:23:34 <johill> good thinking
2008-03-21T00:23:34 <johill> good thinking
2008-03-21T00:24:47 * xorAxAx &
2008-03-21T00:24:47 * xorAxAx &
2008-03-21T00:25:15 <johill> fg xorAxAx
2008-03-21T00:25:15 <johill> fg xorAxAx
2008-03-21T00:25:17 <johill> phh
2008-03-21T00:25:17 <johill> phh
2008-03-21T00:29:03 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3379:0ffcad924a92 1.7/MoinMoin/i18n/ (34 files): updated i18n (copied from 1.6 + a few fixes)
2008-03-21T00:29:03 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3379:0ffcad924a92 1.7/MoinMoin/i18n/ (34 files): updated i18n (copied from 1.6 + a few fixes)
2008-03-21T00:29:04 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3380:6d51aad79f59 1.7/wiki/ (underlay.tar underlay.tar.bz2): updated underlay (copied from 1.6), remove underlay.tar.bz2, use underlay.tar from now on
2008-03-21T00:29:04 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3380:6d51aad79f59 1.7/wiki/ (underlay.tar underlay.tar.bz2): updated underlay (copied from 1.6), remove underlay.tar.bz2, use underlay.tar from now on
2008-03-21T00:29:05 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3381:c8296a0c9799 1.7/Makefile: updated Makefile for 1.7 underlay (pathes, use master17.moinmo.in, use .tar)
2008-03-21T00:29:05 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3381:c8296a0c9799 1.7/Makefile: updated Makefile for 1.7 underlay (pathes, use master17.moinmo.in, use .tar)
2008-03-21T00:29:06 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3382:d0cc11c9c821 1.7/ (14 files in 7 dirs): merged main
2008-03-21T00:29:06 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3382:d0cc11c9c821 1.7/ (14 files in 7 dirs): merged main
2008-03-21T00:40:23 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3383:779f2ab6cb91 1.7/MoinMoin/macro/_tests/test_EmbedObject.py: test_EmbedObject: add a test to check all major mimetype
2008-03-21T00:40:23 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3383:779f2ab6cb91 1.7/MoinMoin/macro/_tests/test_EmbedObject.py: test_EmbedObject: add a test to check all major mimetype
2008-03-21T00:40:24 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3384:475f8094a65f 1.7/MoinMoin/macro/Hits.py: macro.Hits: removed order reverse
2008-03-21T00:40:24 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3384:475f8094a65f 1.7/MoinMoin/macro/Hits.py: macro.Hits: removed order reverse
2008-03-21T00:40:25 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3385:339628170ebf 1.7/ (37 files in 3 dirs): merged main
2008-03-21T00:40:25 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3385:339628170ebf 1.7/ (37 files in 3 dirs): merged main
2008-03-21T00:48:17 <johill> ThomasWaldmann: is underlay.tar in any way more efficient?
2008-03-21T00:48:17 <johill> ThomasWaldmann: is underlay.tar in any way more efficient?
2008-03-21T00:50:21 <johill> lanius: still around?
2008-03-21T00:50:21 <johill> lanius: still around?
2008-03-21T00:54:04 <TheSheep> johill: it's still a text file, so diffs are a little better
2008-03-21T00:54:04 <TheSheep> johill: it's still a text file, so diffs are a little better
2008-03-21T00:54:20 <TheSheep> johill: and changing one part of it doesn't change the whole file
2008-03-21T00:54:20 <TheSheep> johill: and changing one part of it doesn't change the whole file
2008-03-21T00:54:23 <johill> does hg store diffs?
2008-03-21T00:54:23 <johill> does hg store diffs?
2008-03-21T00:54:31 * johill is just too used to git
2008-03-21T00:54:31 * johill is just too used to git
2008-03-21T00:54:33 <TheSheep> johill: for text files, yes
2008-03-21T00:54:33 <TheSheep> johill: for text files, yes
2008-03-21T00:54:39 <dreimark> johill: it is because of diffs
2008-03-21T00:54:39 <dreimark> johill: it is because of diffs
2008-03-21T00:54:48 <johill> but a tar contains lots of binary NULs
2008-03-21T00:54:48 <johill> but a tar contains lots of binary NULs
2008-03-21T00:54:51 <johill> so it's not a text file
2008-03-21T00:54:52 <johill> so it's not a text file
2008-03-21T00:56:32 <johill> this layering sucks
2008-03-21T00:56:33 <johill> this layering sucks
2008-03-21T00:58:18 <ThomasWaldmann> johill: with bz2 bdiff has no chance, with a tar it should be better
2008-03-21T00:58:18 <ThomasWaldmann> johill: with bz2 bdiff has no chance, with a tar it should be better
2008-03-21T00:59:11 <ThomasWaldmann> (and I am currently testing if i can kill the underlay.tar.bz2 including history using hg convert)
2008-03-21T00:59:11 <ThomasWaldmann> (and I am currently testing if i can kill the underlay.tar.bz2 including history using hg convert)
2008-03-21T01:00:07 <johill> wouldn't that change all revision IDs?
2008-03-21T01:00:15 <johill> wouldn't that change all revision IDs?
2008-03-21T01:01:16 <ThomasWaldmann> convert still runs :)
2008-03-21T01:01:16 <ThomasWaldmann> convert still runs :)
2008-03-21T01:05:12 <dreimark> ThomasWaldmann: can we exchange the logo on cia.vc?
2008-03-21T01:05:12 <dreimark> ThomasWaldmann: can we exchange the logo on cia.vc?
2008-03-21T01:05:25 <dreimark> it is currently I thin out faviocon
2008-03-21T01:05:25 <dreimark> it is currently I thin out faviocon
2008-03-21T01:08:43 <dreimark> good night
2008-03-21T01:08:44 <dreimark> good night
2008-03-21T01:08:54 <ThomasWaldmann> gn dreimark
2008-03-21T01:08:54 <ThomasWaldmann> gn dreimark
2008-03-21T01:15:54 <johill> I'm convinced the storage abstraction design I proposed would have made much more sense :(
2008-03-21T01:26:45 <ThomasWaldmann> $ du -sk moin-1.6 moin-1.6-convert_no_underlaybz2
2008-03-21T01:26:45 <ThomasWaldmann> $ du -sk moin-1.6 moin-1.6-convert_no_underlaybz2
2008-03-21T01:26:48 <ThomasWaldmann> 156632 moin-1.6
2008-03-21T01:26:49 <ThomasWaldmann> 156632 moin-1.6
2008-03-21T01:26:51 <ThomasWaldmann> 68480 moin-1.6-convert_no_underlaybz2
2008-03-21T01:26:51 <ThomasWaldmann> 68480 moin-1.6-convert_no_underlaybz2
2008-03-21T01:30:32 <ThomasWaldmann> and yes, all revids change
2008-03-21T01:30:32 <ThomasWaldmann> and yes, all revids change
2008-03-21T01:30:47 <ThomasWaldmann> (but tags are fixed automatically)
2008-03-21T01:30:48 <ThomasWaldmann> (but tags are fixed automatically)
2008-03-21T01:33:27 <TheSheep> ThomasWaldmann: btw, about naming the 'subitems', I've encountered the non-page elements refered to as 'assets' in CMSes
2008-03-21T01:33:27 <TheSheep> ThomasWaldmann: btw, about naming the 'subitems', I've encountered the non-page elements refered to as 'assets' in CMSes
2008-03-21T01:33:59 <ThomasWaldmann> (it creates a .hg/shamap with the mapping)
2008-03-21T01:33:59 <ThomasWaldmann> (it creates a .hg/shamap with the mapping)
2008-03-21T01:34:58 <ThomasWaldmann> TheSheep: if all items are basically the same, we can name them all items
2008-03-21T01:34:58 <ThomasWaldmann> TheSheep: if all items are basically the same, we can name them all items
2008-03-21T01:35:00 <johill> nah, changing revids sucks
2008-03-21T01:35:01 <johill> nah, changing revids sucks
2008-03-21T01:35:21 <lanius> johill: yes
2008-03-21T01:35:22 <lanius> johill: yes
2008-03-21T01:35:35 <johill> lanius: ah, I had almost given up
2008-03-21T01:35:35 <johill> lanius: ah, I had almost given up
2008-03-21T01:35:36 <johill> let me upload a file
2008-03-21T01:35:37 <johill> let me upload a file
2008-03-21T01:35:52 <ThomasWaldmann> johill: the old urls pointing to 1.6 and 1.7 repo will still be valid
2008-03-21T01:35:52 <ThomasWaldmann> johill: the old urls pointing to 1.6 and 1.7 repo will still be valid
2008-03-21T01:35:59 <johill> http://johannes.sipsolutions.net/files/term.py.txt
2008-03-21T01:35:59 <johill> http://johannes.sipsolutions.net/files/term.py.txt
2008-03-21T01:36:12 <ThomasWaldmann> if i do that, it will be when creating 1.8 repo
2008-03-21T01:36:12 <ThomasWaldmann> if i do that, it will be when creating 1.8 repo
2008-03-21T01:36:14 <johill> lanius: that passes backend, pagename (should be itemname) and metadata
2008-03-21T01:36:14 <johill> lanius: that passes backend, pagename (should be itemname) and metadata
2008-03-21T01:36:24 <johill> ThomasWaldmann: will merging work?
2008-03-21T01:36:25 <johill> ThomasWaldmann: will merging work?
2008-03-21T01:37:35 <ThomasWaldmann> good question
2008-03-21T01:37:35 <ThomasWaldmann> good question
2008-03-21T01:37:44 <johill> I doubt it
2008-03-21T01:37:44 <johill> I doubt it
2008-03-21T01:37:53 <TheSheep> parents are different
2008-03-21T01:37:53 <TheSheep> parents are different
2008-03-21T01:39:12 <johill> lanius: what do you think about passing an Item() instance?
2008-03-21T01:39:13 <johill> lanius: what do you think about passing an Item() instance?
2008-03-21T01:39:27 <johill> that means that the backend is no longer insulated from that but I never understood why you did that
2008-03-21T01:39:27 <johill> that means that the backend is no longer insulated from that but I never understood why you did that
2008-03-21T01:39:38 <lanius> insulated from what
2008-03-21T01:39:39 <lanius> insulated from what
2008-03-21T01:39:43 <ThomasWaldmann> abort: repository is unrelated
2008-03-21T01:39:43 <ThomasWaldmann> abort: repository is unrelated
2008-03-21T01:40:01 <johill> knowing about stuff from external.py
2008-03-21T01:40:01 <johill> knowing about stuff from external.py
2008-03-21T01:40:06 <johill> ThomasWaldmann: thought so
2008-03-21T01:40:07 <johill> ThomasWaldmann: thought so
2008-03-21T01:40:23 <lanius> johill: the backend should not know about stuff from external
2008-03-21T01:40:24 <lanius> johill: the backend should not know about stuff from external
2008-03-21T01:40:41 <lanius> there are just a few things for implementing the moin 1.6 backend
2008-03-21T01:40:41 <lanius> there are just a few things for implementing the moin 1.6 backend
2008-03-21T01:40:58 <lanius> but that should be an exception
2008-03-21T01:40:58 <lanius> but that should be an exception
2008-03-21T01:40:59 <johill> lanius: ok then the search function needs to be passed the backend, item name and metadata dict
2008-03-21T01:40:59 <johill> lanius: ok then the search function needs to be passed the backend, item name and metadata dict
2008-03-21T01:41:07 <lanius> yes
2008-03-21T01:41:07 <lanius> yes
2008-03-21T01:41:09 <johill> but then you can't have a search for older revisions
2008-03-21T01:41:09 <johill> but then you can't have a search for older revisions
2008-03-21T01:41:20 <johill> and you can't easily implement full-text search
2008-03-21T01:41:20 <johill> and you can't easily implement full-text search
2008-03-21T01:41:51 <lanius> what about passing each expression only the data it really needs
2008-03-21T01:41:52 <lanius> what about passing each expression only the data it really needs
2008-03-21T01:41:59 <lanius> then you are flexible in extending it later on
2008-03-21T01:42:00 <lanius> then you are flexible in extending it later on
2008-03-21T01:42:13 <johill> how would you do that?
2008-03-21T01:42:13 <johill> how would you do that?
2008-03-21T01:42:21 <lanius> evaluate without parameters
2008-03-21T01:42:21 <lanius> evaluate without parameters
2008-03-21T01:42:34 <johill> and then?
2008-03-21T01:42:34 <johill> and then?
2008-03-21T01:43:00 <lanius> and hm
2008-03-21T01:43:01 <lanius> and hm
2008-03-21T01:43:30 <lanius> not a good idea
2008-03-21T01:43:30 <lanius> not a good idea
2008-03-21T01:43:34 <johill> I don't see why you don't want backends to know about items
2008-03-21T01:43:34 <johill> I don't see why you don't want backends to know about items
2008-03-21T01:43:38 <lanius> but what about passing the backend, pagename and revision
2008-03-21T01:43:38 <lanius> but what about passing the backend, pagename and revision
2008-03-21T01:44:11 <johill> that would work but then the search would have to instantiate an ItemCollection() to get an Item()
2008-03-21T01:44:11 <johill> that would work but then the search would have to instantiate an ItemCollection() to get an Item()
2008-03-21T01:44:18 <johill> seems rather expensive to do all the time
2008-03-21T01:44:18 <johill> seems rather expensive to do all the time
2008-03-21T01:44:20 <lanius> no
2008-03-21T01:44:21 <lanius> no
2008-03-21T01:44:25 <lanius> it can direcly operate on the backend
2008-03-21T01:44:25 <lanius> it can direcly operate on the backend
2008-03-21T01:44:52 <johill> which sort of defeats the purpose of external.py
2008-03-21T01:44:52 <johill> which sort of defeats the purpose of external.py
2008-03-21T01:44:58 <johill> (if you treat search terms as external)
2008-03-21T01:44:58 <johill> (if you treat search terms as external)
2008-03-21T01:45:11 <lanius> i think we have to separate this into two parts
2008-03-21T01:45:12 <lanius> i think we have to separate this into two parts
2008-03-21T01:45:21 <lanius> first: the search term
2008-03-21T01:45:21 <lanius> first: the search term
2008-03-21T01:45:28 <lanius> second: the evaluation
2008-03-21T01:45:28 <lanius> second: the evaluation
2008-03-21T01:45:36 <johill> go on?
2008-03-21T01:45:36 <johill> go on?
2008-03-21T01:45:48 <lanius> because if the evaluation belongs to the search term you cannot optimzie it
2008-03-21T01:45:49 <lanius> because if the evaluation belongs to the search term you cannot optimzie it
2008-03-21T01:46:12 <johill> ah
2008-03-21T01:46:13 <johill> ah
2008-03-21T01:46:16 <johill> I see what you mean
2008-03-21T01:46:16 <johill> I see what you mean
2008-03-21T01:46:18 <johill> no
2008-03-21T01:46:18 <johill> no
2008-03-21T01:46:25 <johill> that's why I define very clear search term classes
2008-03-21T01:46:25 <johill> that's why I define very clear search term classes
2008-03-21T01:46:49 <johill> so the backend can check something like isinstance(t, term.TitleMatch) and pull it out of the expression
2008-03-21T01:46:49 <johill> so the backend can check something like isinstance(t, term.TitleMatch) and pull it out of the expression
2008-03-21T01:47:32 <lanius> yes but still
2008-03-21T01:47:32 <lanius> yes but still
2008-03-21T01:47:39 <lanius> it does some kind of evaluating then
2008-03-21T01:47:40 <lanius> it does some kind of evaluating then
2008-03-21T01:47:54 <lanius> and then again on another level
2008-03-21T01:47:54 <lanius> and then again on another level
2008-03-21T01:47:56 <johill> yeah and the search term's evaluate function won't be called
2008-03-21T01:47:56 <johill> yeah and the search term's evaluate function won't be called
2008-03-21T01:48:05 <lanius> so we don't need them at all
2008-03-21T01:48:05 <lanius> so we don't need them at all
2008-03-21T01:48:07 <johill> because it'll remove it from the expression
2008-03-21T01:48:07 <johill> because it'll remove it from the expression
2008-03-21T01:48:09 <johill> yes we do
2008-03-21T01:48:09 <johill> yes we do
2008-03-21T01:48:17 <johill> because we can't require backends to implement all of it
2008-03-21T01:48:17 <johill> because we can't require backends to implement all of it
2008-03-21T01:48:26 <lanius> righ
2008-03-21T01:48:26 <lanius> righ
2008-03-21T01:48:27 <lanius> t
2008-03-21T01:48:27 <lanius> t
2008-03-21T01:48:33 <lanius> but i would put it somewhere else then
2008-03-21T01:48:34 <lanius> but i would put it somewhere else then
2008-03-21T01:48:45 <johill> so the trivial backend just does for item in items: if filter.evaluate(...): yield item
2008-03-21T01:48:45 <johill> so the trivial backend just does for item in items: if filter.evaluate(...): yield item
2008-03-21T01:48:47 <lanius> maybe in the common backend
2008-03-21T01:48:47 <lanius> maybe in the common backend
2008-03-21T01:48:57 <lanius> but separate data structure and code logic
2008-03-21T01:48:57 <lanius> but separate data structure and code logic
2008-03-21T01:49:05 <johill> no
2008-03-21T01:49:05 <johill> no
2008-03-21T01:49:10 <johill> then you can't easily write new search terms
2008-03-21T01:49:10 <johill> then you can't easily write new search terms
2008-03-21T01:49:24 <lanius> you would have to change two locations
2008-03-21T01:49:25 <lanius> you would have to change two locations
2008-03-21T01:49:41 <johill> while like this you only have to define a new Term class
2008-03-21T01:49:41 <johill> while like this you only have to define a new Term class
2008-03-21T01:49:46 <johill> much nicer
2008-03-21T01:49:46 <johill> much nicer
2008-03-21T01:50:03 <lanius> then again about operating on the backend directly
2008-03-21T01:50:04 <lanius> then again about operating on the backend directly
2008-03-21T01:50:10 <lanius> this is a low level operation
2008-03-21T01:50:11 <lanius> this is a low level operation
2008-03-21T01:50:18 <lanius> but the execution only occurs on this low level
2008-03-21T01:50:18 <lanius> but the execution only occurs on this low level
2008-03-21T01:50:21 <lanius> so why not do it
2008-03-21T01:50:21 <lanius> so why not do it
2008-03-21T01:50:31 <johill> it's not low-level if I write a new Term class in some macro
2008-03-21T01:50:31 <johill> it's not low-level if I write a new Term class in some macro
2008-03-21T01:50:40 <lanius> do you do that?
2008-03-21T01:50:41 <lanius> do you do that?
2008-03-21T01:51:03 <johill> you could, if that macro needs something the regular terms can't express
2008-03-21T01:51:03 <johill> you could, if that macro needs something the regular terms can't express
2008-03-21T01:51:08 <lanius> if you do that then operate on an item
2008-03-21T01:51:08 <lanius> if you do that then operate on an item
2008-03-21T01:51:23 <lanius> it's only slow then because the backend doesn't support it
2008-03-21T01:51:23 <lanius> it's only slow then because the backend doesn't support it
2008-03-21T01:51:37 <johill> now the cat bites its tail
2008-03-21T01:51:37 <johill> now the cat bites its tail
2008-03-21T01:51:52 <lanius> no
2008-03-21T01:51:53 <lanius> no
2008-03-21T01:51:53 <johill> I can't make search terms operate on items because tehy're executed by the backend
2008-03-21T01:51:53 <johill> I can't make search terms operate on items because tehy're executed by the backend
2008-03-21T01:51:58 <lanius> yes
2008-03-21T01:51:58 <lanius> yes
2008-03-21T01:52:03 <lanius> but search terms only operate on list_items
2008-03-21T01:52:04 <lanius> but search terms only operate on list_items
2008-03-21T01:52:10 <lanius> which the terms aren't allowed to do then
2008-03-21T01:52:11 <lanius> which the terms aren't allowed to do then
2008-03-21T01:52:18 <johill> within list_items(), yes
2008-03-21T01:52:18 <johill> within list_items(), yes
2008-03-21T01:52:22 <lanius> yes
2008-03-21T01:52:22 <lanius> yes
2008-03-21T01:52:40 <johill> so if my macro wants a list of pages that some_filter_function() returns true for
2008-03-21T01:52:40 <johill> so if my macro wants a list of pages that some_filter_function() returns true for
2008-03-21T01:52:54 <johill> it would createa new Term class that used some_filter_function() as its evaluate
2008-03-21T01:52:54 <johill> it would createa new Term class that used some_filter_function() as its evaluate
2008-03-21T01:53:02 <lanius> call list_items without a filter function
2008-03-21T01:53:02 <lanius> call list_items without a filter function
2008-03-21T01:53:08 <johill> no no
2008-03-21T01:53:08 <johill> no no
2008-03-21T01:53:18 <johill> that defeats the purpose
2008-03-21T01:53:18 <johill> that defeats the purpose
2008-03-21T01:53:38 <lanius> we already said that the general terms can't be very performant
2008-03-21T01:53:39 <lanius> we already said that the general terms can't be very performant
2008-03-21T01:53:43 <lanius> only the optimized ones
2008-03-21T01:53:43 <lanius> only the optimized ones
2008-03-21T01:54:01 <johill> yes but a unified interface has value too
2008-03-21T01:54:01 <johill> yes but a unified interface has value too
2008-03-21T01:54:49 <lanius> then we have to exclude the case with old revisions?
2008-03-21T01:54:49 <lanius> then we have to exclude the case with old revisions?
2008-03-21T01:55:43 <johill> I think overall I haven't understood why you made a distinction between ItemCollection and backend
2008-03-21T01:55:43 <johill> I think overall I haven't understood why you made a distinction between ItemCollection and backend
2008-03-21T01:55:53 <johill> why is each backend not just a special descendent of ItemCollection?
2008-03-21T01:55:53 <johill> why is each backend not just a special descendent of ItemCollection?
2008-03-21T01:56:46 <lanius> because on the one side you have an object oriented interface
2008-03-21T01:56:46 <lanius> because on the one side you have an object oriented interface
2008-03-21T01:56:56 <lanius> which is easier to use
2008-03-21T01:56:57 <lanius> which is easier to use
2008-03-21T01:57:06 <lanius> and on the other side you have a flat interface, which is easier to implement
2008-03-21T01:57:07 <lanius> and on the other side you have a flat interface, which is easier to implement
2008-03-21T01:57:51 <johill> but they could be in a single object
2008-03-21T01:57:51 <johill> but they could be in a single object
2008-03-21T01:58:01 <lanius> how
2008-03-21T01:58:02 <lanius> how
2008-03-21T01:58:10 <johill> with the glue code in the base class
2008-03-21T01:58:10 <johill> with the glue code in the base class
2008-03-21T01:58:53 <lanius> yes, but there is no real difference between glue code in the base class (which would make the inheritence structure only deeper) or in a separate class, is there
2008-03-21T01:58:53 <lanius> yes, but there is no real difference between glue code in the base class (which would make the inheritence structure only deeper) or in a separate class, is there
2008-03-21T01:59:32 <johill> not much
2008-03-21T01:59:32 <johill> not much
2008-03-21T01:59:49 <johill> but it would allow the backend to pass Item instances to the filter term
2008-03-21T01:59:49 <johill> but it would allow the backend to pass Item instances to the filter term
2008-03-21T02:00:37 <lanius> yes
2008-03-21T02:00:37 <lanius> yes
2008-03-21T02:00:49 <lanius> but i would consider the filter low level
2008-03-21T02:00:50 <lanius> but i would consider the filter low level
2008-03-21T02:01:10 <lanius> and we should supply enough filter functions so that a macro developer only has to use those existing terms
2008-03-21T02:01:10 <lanius> and we should supply enough filter functions so that a macro developer only has to use those existing terms
2008-03-21T02:02:06 <johill> ok
2008-03-21T02:02:06 <johill> ok
2008-03-21T02:02:07 <johill> anyway
2008-03-21T02:02:07 <johill> anyway
2008-03-21T02:02:09 <johill> I don't care much
2008-03-21T02:02:09 <johill> I don't care much
2008-03-21T02:02:30 <johill> if need be, we can define a filter Term class that takes the stuff and creates an item before passing it on to the subclass _evaluate
2008-03-21T02:02:30 <johill> if need be, we can define a filter Term class that takes the stuff and creates an item before passing it on to the subclass _evaluate
2008-03-21T02:02:42 <johill> so filter Terms just get to be low-level
2008-03-21T02:02:42 <johill> so filter Terms just get to be low-level
2008-03-21T02:03:55 <lanius> yes
2008-03-21T02:03:55 <lanius> yes
2008-03-21T02:04:46 <johill> oh btw
2008-03-21T02:04:51 <johill> oh btw
2008-03-21T02:04:57 <johill> get_data_backend() says something about error raising
2008-03-21T02:04:57 <johill> get_data_backend() says something about error raising
2008-03-21T02:05:13 <johill> shouldn't that rather be done in the mid-layer, at Item/ItemCollection level?
2008-03-21T02:05:13 <johill> shouldn't that rather be done in the mid-layer, at Item/ItemCollection level?
2008-03-21T02:05:55 <lanius> what exactly do you mean
2008-03-21T02:05:55 <lanius> what exactly do you mean
2008-03-21T02:06:24 <johill> isn't that talking about acl checking?
2008-03-21T02:06:24 <johill> isn't that talking about acl checking?
2008-03-21T02:06:43 <lanius> acl checking has to be done on item level, that is correct
2008-03-21T02:06:43 <lanius> acl checking has to be done on item level, that is correct
2008-03-21T02:08:13 <johill> k
2008-03-21T02:08:13 <johill> k
2008-03-21T02:08:45 <lanius> will try to get some sleep now :-) gn8
2008-03-21T02:08:45 <lanius> will try to get some sleep now :-) gn8
2008-03-21T02:08:58 <johill> night
2008-03-21T02:08:58 <johill> night
2008-03-21T02:25:41 <johill> yay, py.test crashes in its own code
2008-03-21T02:25:41 <johill> yay, py.test crashes in its own code
2008-03-21T03:03:37 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2588:17e38b189d67 1.6/ (2 files in 2 dirs): Fix for FCKeditor dialog boxes with FireFox3 (thanks to Rich Renomeron)
2008-03-21T03:03:37 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2588:17e38b189d67 1.6/ (2 files in 2 dirs): Fix for FCKeditor dialog boxes with FireFox3 (thanks to Rich Renomeron)
2008-03-21T03:08:34 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3386:26d7550a4b01 1.7/wiki/htdocs/applets/FCKeditor/editor/js/fckeditorcode_gecko_2.js: Fix for FCKeditor dialog boxes with FireFox3 (thanks to Rich Renomeron)
2008-03-21T03:08:36 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3386:26d7550a4b01 1.7/wiki/htdocs/applets/FCKeditor/editor/js/fckeditorcode_gecko_2.js: Fix for FCKeditor dialog boxes with FireFox3 (thanks to Rich Renomeron)
2008-03-21T04:06:55 <johill> zzz now, will hook it up after sleep
2008-03-21T04:06:55 <johill> zzz now, will hook it up after sleep
2008-03-21T04:07:30 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3663:b827d5517f41 1.7-storage-hwendel/MoinMoin/storage/_tests/__init__.py: fix pep8 failure in storage test
2008-03-21T04:07:30 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3663:b827d5517f41 1.7-storage-hwendel/MoinMoin/storage/_tests/__init__.py: fix pep8 failure in storage test
2008-03-21T04:07:31 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3664:46262f283845 1.7-storage-hwendel/MoinMoin/search/ (_tests/test_terms.py term.py): add query language for backend.list_items() and ItemCollection.iterate(), includes tests
2008-03-21T04:07:31 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3664:46262f283845 1.7-storage-hwendel/MoinMoin/search/ (_tests/test_terms.py term.py): add query language for backend.list_items() and ItemCollection.iterate(), includes tests
2008-03-21T04:07:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3665:d62d35f61a45 1.7-storage-hwendel/ (44 files in 7 dirs): merge with main branch
2008-03-21T04:07:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3665:d62d35f61a45 1.7-storage-hwendel/ (44 files in 7 dirs): merge with main branch
2008-03-21T08:11:17 <dreimark> moin
2008-03-21T08:11:17 <dreimark> moin
2008-03-21T08:11:30 * dreimark looks at xmlrpc in 1.7
2008-03-21T08:11:30 * dreimark looks at xmlrpc in 1.7
2008-03-21T08:14:50 <ThomasWaldmann> moin
2008-03-21T08:14:50 <ThomasWaldmann> moin
2008-03-21T08:32:09 <dreimark> seems session can't do a user_obj
2008-03-21T08:32:09 <dreimark> seems session can't do a user_obj
2008-03-21T10:10:23 <dreimark> ha got it
2008-03-21T10:10:23 <dreimark> ha got it
2008-03-21T10:15:53 <dreimark> ThomasWaldmann: please check http://moinmo.in/ReimarBauer?action=AttachFile&do=view&target=xmlrpc_unknown_editor.patch
2008-03-21T10:15:53 <dreimark> ThomasWaldmann: please check http://moinmo.in/ReimarBauer?action=AttachFile&do=view&target=xmlrpc_unknown_editor.patch
2008-03-21T10:21:53 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3387:17bfccaeaf31 1.7/MoinMoin/macro/_tests/test_EmbedObject.py: test_EmbedObject: whitespace fix
2008-03-21T10:21:53 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3387:17bfccaeaf31 1.7/MoinMoin/macro/_tests/test_EmbedObject.py: test_EmbedObject: whitespace fix
2008-03-21T10:32:23 <dreimark> bbl
2008-03-21T10:32:23 <dreimark> bbl
2008-03-21T10:59:02 <dennda> moin (no pun intended)
2008-03-21T10:59:02 <dennda> moin (no pun intended)
2008-03-21T11:33:02 <johill> dreimark: looks about right
2008-03-21T11:33:02 <johill> dreimark: looks about right
2008-03-21T11:42:00 <johill> xapian can only search for words, right?
2008-03-21T11:42:00 <johill> xapian can only search for words, right?
2008-03-21T11:42:04 <johill> not in the middle of words
2008-03-21T11:42:04 <johill> not in the middle of words
2008-03-21T11:46:41 <johill> hm, there is stemming, but that's not quite all
2008-03-21T11:46:41 <johill> hm, there is stemming, but that's not quite all
2008-03-21T11:51:09 <dreimark> johill: s/about/above/ ?
2008-03-21T11:51:09 <dreimark> johill: s/about/above/ ?
2008-03-21T11:51:42 <dreimark> hi dennda
2008-03-21T11:51:42 <dreimark> hi dennda
2008-03-21T11:51:44 <johill> 'above
2008-03-21T11:51:45 <johill> 'above
2008-03-21T11:51:49 <johill> 'above looks about right' :)
2008-03-21T11:51:49 <johill> 'above looks about right' :)
2008-03-21T11:52:03 <johill> I'm wondering whether that should be in request
2008-03-21T11:52:03 <johill> I'm wondering whether that should be in request
2008-03-21T11:52:58 <dreimark> in 1.6 it was done by http://hg.moinmo.in/moin/1.6/file/17e38b189d67/MoinMoin/xmlrpc/__init__.py
2008-03-21T11:52:58 <dreimark> in 1.6 it was done by http://hg.moinmo.in/moin/1.6/file/17e38b189d67/MoinMoin/xmlrpc/__init__.py
2008-03-21T11:53:09 <dreimark> line 598
2008-03-21T11:53:09 <dreimark> line 598
2008-03-21T11:53:31 <dreimark> some session data but that gots removed
2008-03-21T11:53:31 <dreimark> some session data but that gots removed
2008-03-21T11:54:45 <johill> why doesn't hgweb have line anchors like every other interface :)
2008-03-21T11:54:45 <johill> why doesn't hgweb have line anchors like every other interface :)
2008-03-21T11:55:17 <johill> yeah I think what you're doing is right
2008-03-21T11:55:17 <johill> yeah I think what you're doing is right
2008-03-21T11:55:44 <dreimark> may be we should enter a soc task wish list :)
2008-03-21T11:55:44 <dreimark> may be we should enter a soc task wish list :)
2008-03-21T11:56:19 <johill> would ano
2008-03-21T11:56:19 <johill> would ano
2008-03-21T11:56:39 <johill> our current query syntax sucks for xapian
2008-03-21T11:56:39 <johill> our current query syntax sucks for xapian
2008-03-21T11:57:16 <dreimark> is that the reason for "those" bugs ?
2008-03-21T11:57:16 <dreimark> is that the reason for "those" bugs ?
2008-03-21T11:57:23 * dennda is writing his application
2008-03-21T11:57:23 * dennda is writing his application
2008-03-21T11:57:29 <johill> "those" bugs?
2008-03-21T11:57:29 <johill> "those" bugs?
2008-03-21T11:57:53 <dreimark> Advanced Search does not really work
2008-03-21T11:57:53 <dreimark> Advanced Search does not really work
2008-03-21T11:58:27 <johill> maybe
2008-03-21T11:58:27 <johill> maybe
2008-03-21T11:58:34 <johill> but if you enter only Hello
2008-03-21T11:58:34 <johill> but if you enter only Hello
2008-03-21T11:58:44 <johill> it has to find chello too
2008-03-21T11:58:44 <johill> it has to find chello too
2008-03-21T11:58:58 <johill> which is pretty sucky and I think most people wouldn't even want that
2008-03-21T11:58:58 <johill> which is pretty sucky and I think most people wouldn't even want that
2008-03-21T11:59:10 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3388:8ff8bc7dea20 1.7/MoinMoin/xmlrpc/__init__.py: xmlrpc.__init__: fixed missing user xmlrpc_applyAuthToken
2008-03-21T11:59:10 <CIA-39> Reimar Bauer <rb.proj AT googlemail DOT com> default * 3388:8ff8bc7dea20 1.7/MoinMoin/xmlrpc/__init__.py: xmlrpc.__init__: fixed missing user xmlrpc_applyAuthToken
2008-03-21T11:59:23 <ThomasWaldmann> johill: that is not easily possible
2008-03-21T11:59:23 <ThomasWaldmann> johill: that is not easily possible
2008-03-21T11:59:39 <ThomasWaldmann> johill: it supports * at the end
2008-03-21T11:59:39 <ThomasWaldmann> johill: it supports * at the end
2008-03-21T11:59:41 <johill> ThomasWaldmann: that's what I'm saying
2008-03-21T11:59:41 <johill> ThomasWaldmann: that's what I'm saying
2008-03-21T11:59:51 <johill> ThomasWaldmann: so imho we should change the query syntax in moin
2008-03-21T11:59:51 <johill> ThomasWaldmann: so imho we should change the query syntax in moin
2008-03-21T12:00:05 <johill> to make that like '*hello' if wanted
2008-03-21T12:00:05 <johill> to make that like '*hello' if wanted
2008-03-21T12:00:29 <ThomasWaldmann> johill: but if we index(reverse(word)), we can search for reverse(*word)
2008-03-21T12:00:29 <ThomasWaldmann> johill: but if we index(reverse(word)), we can search for reverse(*word)
2008-03-21T12:01:09 <johill> and then we can emulate that
2008-03-21T12:01:09 <johill> and then we can emulate that
2008-03-21T12:01:10 <johill> right
2008-03-21T12:01:10 <johill> right
2008-03-21T12:01:12 <johill> however
2008-03-21T12:01:12 <johill> however
2008-03-21T12:01:38 <johill> that means you need to express 'search for "hello"' as 're:\bhello\b'
2008-03-21T12:01:38 <johill> that means you need to express 'search for "hello"' as 're:\bhello\b'
2008-03-21T12:01:55 <johill> whcih (a) is not intuitive and (b) requires special parsing or, by default, not using the index
2008-03-21T12:01:55 <johill> whcih (a) is not intuitive and (b) requires special parsing or, by default, not using the index
2008-03-21T12:02:23 <ThomasWaldmann> doesnt help much for searching *foo* though
2008-03-21T12:02:23 <ThomasWaldmann> doesnt help much for searching *foo* though
2008-03-21T12:02:58 <johill> yeah but at least then the user explicitly asked for that
2008-03-21T12:03:12 <johill> yeah but at least then the user explicitly asked for that
2008-03-21T12:03:15 <johill> that's all I'm saying, pretty much
2008-03-21T12:03:15 <johill> that's all I'm saying, pretty much
2008-03-21T12:03:29 <johill> google behaves that way too afaik
2008-03-21T12:03:29 <johill> google behaves that way too afaik
2008-03-21T12:07:59 <johill> so it's not like users expect to find 'golden' when searching for 'den'
2008-03-21T12:07:59 <johill> so it's not like users expect to find 'golden' when searching for 'den'
2008-03-21T12:09:16 <ThomasWaldmann> no, but heidelbeerjoghurt when searching for beer :)
2008-03-21T12:09:16 <ThomasWaldmann> no, but heidelbeerjoghurt when searching for beer :)
2008-03-21T12:09:31 * ThomasWaldmann eats one
2008-03-21T12:09:31 * ThomasWaldmann eats one
2008-03-21T12:10:16 <johill> heh
2008-03-21T12:10:16 <johill> heh
2008-03-21T12:10:21 <ThomasWaldmann> not sure how much the stemming helps here
2008-03-21T12:10:21 <ThomasWaldmann> not sure how much the stemming helps here
2008-03-21T12:10:36 <johill> no the point isn't that
2008-03-21T12:10:36 <johill> no the point isn't that
2008-03-21T12:10:55 <johill> the point is that if we make a slight change to the search syntax we can easily express both and optimise the default case better
2008-03-21T12:10:55 <johill> the point is that if we make a slight change to the search syntax we can easily express both and optimise the default case better
2008-03-21T12:11:56 <johill> i.e. if we allow the user to enter "beer" (not finding heidelbeere) and "beer*" (finding beeren") and "*beer*" (finding heidelbeere) then we can optimise the just "beer" case in xapian
2008-03-21T12:11:56 <johill> i.e. if we allow the user to enter "beer" (not finding heidelbeere) and "beer*" (finding beeren") and "*beer*" (finding heidelbeere) then we can optimise the just "beer" case in xapian
2008-03-21T12:12:00 <ThomasWaldmann> can we maybe first fix the bugs and then add changes? (except if the change is related to fixing a bug)
2008-03-21T12:12:00 <ThomasWaldmann> can we maybe first fix the bugs and then add changes? (except if the change is related to fixing a bug)
2008-03-21T12:12:16 <johill> oh I wasn't going to make that change in teh old query parser and search stuff
2008-03-21T12:12:16 <johill> oh I wasn't going to make that change in teh old query parser and search stuff
2008-03-21T12:12:27 <johill> but rather do it as I go to the new finding API
2008-03-21T12:12:28 <johill> but rather do it as I go to the new finding API
2008-03-21T12:13:25 <ThomasWaldmann> see "major feature freeze"
2008-03-21T12:13:25 <ThomasWaldmann> see "major feature freeze"
2008-03-21T12:14:52 <johill> not in the storage branch I hope
2008-03-21T12:14:52 <johill> not in the storage branch I hope
2008-03-21T12:15:45 <ThomasWaldmann> no, storage won't get merged into 1.7 main branch now
2008-03-21T12:15:46 <ThomasWaldmann> no, storage won't get merged into 1.7 main branch now
2008-03-21T12:15:57 <ThomasWaldmann> see that page i keep posting :)
2008-03-21T12:15:57 <ThomasWaldmann> see that page i keep posting :)
2008-03-21T12:17:55 <ThomasWaldmann> but the storage branch should not become the storage-xapian branch :P
2008-03-21T12:17:55 <ThomasWaldmann> but the storage branch should not become the storage-xapian branch :P
2008-03-21T12:19:34 <ThomasWaldmann> dreimark: thanks for fixing the xmlrpc user obj. rather obvious once one has found it. :D
2008-03-21T12:19:34 <ThomasWaldmann> dreimark: thanks for fixing the xmlrpc user obj. rather obvious once one has found it. :D
2008-03-21T12:20:57 <johill> I know
2008-03-21T12:20:57 <johill> I know
2008-03-21T12:21:14 <johill> but xapian is alread
2008-03-21T12:21:14 <johill> but xapian is alread
2008-03-21T12:21:18 <johill> already in it
2008-03-21T12:21:18 <johill> already in it
2008-03-21T12:21:27 <johill> can you mark classes as 'final'?
2008-03-21T12:21:27 <johill> can you mark classes as 'final'?
2008-03-21T12:22:20 <johill> guess not
2008-03-21T12:22:20 <johill> guess not
2008-03-21T12:22:28 <johill> will have to rely on docs then
2008-03-21T12:22:28 <johill> will have to rely on docs then
2008-03-21T12:27:26 <dreimark> ThomasWaldmann: :)
2008-03-21T12:27:26 <dreimark> ThomasWaldmann: :)
2008-03-21T12:27:54 * ThomasWaldmann fixes moin xmlrpc write docs
2008-03-21T12:27:54 * ThomasWaldmann fixes moin xmlrpc write docs
2008-03-21T12:30:08 <ThomasWaldmann> johill: do you thing we should fix the docs or the code there?
2008-03-21T12:30:09 <ThomasWaldmann> johill: do you thing we should fix the docs or the code there?
2008-03-21T12:30:42 <johill> how do the two differ?
2008-03-21T12:30:42 <johill> how do the two differ?
2008-03-21T12:31:03 <ThomasWaldmann> MoinMoin/script/moin.py xmlrpc write http://test17.moinmo.in/?action=xmlrpc2 user pass pagename
2008-03-21T12:31:03 <ThomasWaldmann> MoinMoin/script/moin.py xmlrpc write http://test17.moinmo.in/?action=xmlrpc2 user pass pagename
2008-03-21T12:31:12 <ThomasWaldmann> this is how it works now
2008-03-21T12:31:12 <ThomasWaldmann> this is how it works now
2008-03-21T12:31:38 <johill> oh the scripts
2008-03-21T12:31:38 <johill> oh the scripts
2008-03-21T12:31:42 <ThomasWaldmann> the ?action=xmlrpc2 could be auto appended
2008-03-21T12:31:42 <ThomasWaldmann> the ?action=xmlrpc2 could be auto appended
2008-03-21T12:31:58 <johill> no opinion
2008-03-21T12:31:58 <johill> no opinion
2008-03-21T12:33:04 <ThomasWaldmann> ok, I'll change it
2008-03-21T12:33:04 <ThomasWaldmann> ok, I'll change it
2008-03-21T12:33:58 <johill> ma
2008-03-21T12:33:59 <johill> ma
2008-03-21T12:34:08 <johill> maybe giving it on the command line would allow using the script with other wikis
2008-03-21T12:34:08 <johill> maybe giving it on the command line would allow using the script with other wikis
2008-03-21T12:34:17 <johill> but that's the only reason I could possibly see for it
2008-03-21T12:34:17 <johill> but that's the only reason I could possibly see for it
2008-03-21T12:35:51 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3666:e2b345030e42 1.7-storage-hwendel/MoinMoin/search/term.py: search term docstrings
2008-03-21T12:35:51 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3666:e2b345030e42 1.7-storage-hwendel/MoinMoin/search/term.py: search term docstrings
2008-03-21T12:41:03 <ThomasWaldmann> johill: i don't thing they have getAuthToken/applyAuthToken/deleteAuthToken
2008-03-21T12:41:03 <ThomasWaldmann> johill: i don't thing they have getAuthToken/applyAuthToken/deleteAuthToken
2008-03-21T12:41:15 <ThomasWaldmann> k
2008-03-21T12:41:15 <ThomasWaldmann> k
2008-03-21T12:42:33 <johill> good point
2008-03-21T12:42:34 <johill> good point
2008-03-21T12:50:54 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2589:ab0766295dab 1.6/ (MoinMoin/script/xmlrpc/write.py docs/CHANGES): moin xmlrpc write command: improve builtin docs, auto-append ?action=xmlrpc2
2008-03-21T12:50:54 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2589:ab0766295dab 1.6/ (MoinMoin/script/xmlrpc/write.py docs/CHANGES): moin xmlrpc write command: improve builtin docs, auto-append ?action=xmlrpc2
2008-03-21T12:52:05 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3667:79ac41368b57 1.7-storage-hwendel/MoinMoin/search/ (_tests/test_terms.py term.py): add Word, WordStart, WordEnd search terms + tests, small bugfix
2008-03-21T12:52:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3667:79ac41368b57 1.7-storage-hwendel/MoinMoin/search/ (_tests/test_terms.py term.py): add Word, WordStart, WordEnd search terms + tests, small bugfix
2008-03-21T12:52:37 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3389:df4f538f8e8f 1.7/MoinMoin/script/xmlrpc/write.py: moin xmlrpc write command: improve builtin docs, auto-append ?action=xmlrpc2
2008-03-21T12:52:37 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3389:df4f538f8e8f 1.7/MoinMoin/script/xmlrpc/write.py: moin xmlrpc write command: improve builtin docs, auto-append ?action=xmlrpc2
2008-03-21T12:58:52 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3668:7cc608ccca48 1.7-storage-hwendel/MoinMoin/search/ (_tests/test_terms.py term.py): add term.TRUE and term.FALSE
2008-03-21T12:58:52 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3668:7cc608ccca48 1.7-storage-hwendel/MoinMoin/search/ (_tests/test_terms.py term.py): add term.TRUE and term.FALSE
2008-03-21T13:00:46 * ThomasWaldmann works on http://moinmo.in/MoinMoinBugs/GuiEditorSyntaxError
2008-03-21T13:00:46 * ThomasWaldmann works on http://moinmo.in/MoinMoinBugs/GuiEditorSyntaxError
2008-03-21T13:16:32 <johill> uh huh
2008-03-21T13:16:32 <johill> uh huh
2008-03-21T13:16:41 <johill> now I created circular module dependencies
2008-03-21T13:16:41 <johill> now I created circular module dependencies
2008-03-21T13:20:09 <dennda> what happens?
2008-03-21T13:20:09 <dennda> what happens?
2008-03-21T13:20:48 <johill> p
2008-03-21T13:20:48 <johill> p
2008-03-21T13:20:58 <johill> python gives you a trackeback
2008-03-21T13:20:58 <johill> python gives you a trackeback
2008-03-21T13:22:06 <dennda> just wondering, because there is no problem if I construct the deadly diamond of death with two classes that extend each other
2008-03-21T13:22:07 <dennda> just wondering, because there is no problem if I construct the deadly diamond of death with two classes that extend each other
2008-03-21T13:22:39 <dennda> I can even instantiate an object
2008-03-21T13:22:40 <dennda> I can even instantiate an object
2008-03-21T13:24:03 <johill> try with modules :P
2008-03-21T13:24:03 <johill> try with modules :P
2008-03-21T13:26:00 <johill> wow it even works
2008-03-21T13:26:00 <johill> wow it even works
2008-03-21T13:31:16 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2590:48c17dda61ed 1.6/ (3 files in 3 dirs): GUI editor: fixed javascript error by using a simpler word_rule_js regex
2008-03-21T13:31:16 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2590:48c17dda61ed 1.6/ (3 files in 3 dirs): GUI editor: fixed javascript error by using a simpler word_rule_js regex
2008-03-21T13:35:56 <dennda> hg died again
2008-03-21T13:35:56 <dennda> hg died again
2008-03-21T13:36:14 <dennda> strange
2008-03-21T13:36:15 <dennda> strange
2008-03-21T13:37:08 <ThomasWaldmann> sorry, that was me restarting apache
2008-03-21T13:37:08 <ThomasWaldmann> sorry, that was me restarting apache
2008-03-21T13:38:10 <dennda> ah :)
2008-03-21T13:38:10 <dennda> ah :)
2008-03-21T13:45:09 <johill> cool
2008-03-21T13:45:09 <johill> cool
2008-03-21T13:45:17 <johill> now the backends use the search.term stuff
2008-03-21T13:45:18 <johill> now the backends use the search.term stuff
2008-03-21T13:45:21 <johill> that was easy
2008-03-21T13:45:21 <johill> that was easy
2008-03-21T13:53:45 <johill> now some weird tests fails that seems unrelated. odd
2008-03-21T13:53:45 <johill> now some weird tests fails that seems unrelated. odd
2008-03-21T13:59:47 <johill> or maybe about the deletd metadata
2008-03-21T13:59:47 <johill> or maybe about the deletd metadata
2008-03-21T14:01:38 <johill> ahrg
2008-03-21T14:01:38 <johill> ahrg
2008-03-21T14:01:38 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3669:d1dadcf4f1d6 1.7-storage-hwendel/MoinMoin/macro/_tests/test_EmbedObject.py: fix pep8 in MoinMoin/macro/_tests/test_EmbedObject.py
2008-03-21T14:01:38 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3669:d1dadcf4f1d6 1.7-storage-hwendel/MoinMoin/macro/_tests/test_EmbedObject.py: fix pep8 in MoinMoin/macro/_tests/test_EmbedObject.py
2008-03-21T14:01:41 * johill kicks dreimark
2008-03-21T14:01:41 * johill kicks dreimark
2008-03-21T14:04:27 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3390:f1ba94b7913d 1.7/MoinMoin/ (PageGraphicalEditor.py parser/text_moin_wiki.py): GUI editor: fixed javascript error by using a simpler word_rule_js regex
2008-03-21T14:04:27 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3390:f1ba94b7913d 1.7/MoinMoin/ (PageGraphicalEditor.py parser/text_moin_wiki.py): GUI editor: fixed javascript error by using a simpler word_rule_js regex
2008-03-21T14:10:32 <dreimark> johill: why
2008-03-21T14:10:32 <dreimark> johill: why
2008-03-21T14:12:17 <johill> pep8 failure again
2008-03-21T14:12:17 <johill> pep8 failure again
2008-03-21T14:14:15 * ThomasWaldmann runs tests
2008-03-21T14:14:15 * ThomasWaldmann runs tests
2008-03-21T14:14:26 <dreimark> johill: test
2008-03-21T14:14:26 <dreimark> johill: test
2008-03-21T14:14:46 <dreimark> test_source doesn't list that one, but sorry for the extra line
2008-03-21T14:14:46 <dreimark> test_source doesn't list that one, but sorry for the extra line
2008-03-21T14:15:45 <johill> test_source is weird
2008-03-21T14:15:45 <johill> test_source is weird
2008-03-21T14:15:50 <johill> and not really deterministic
2008-03-21T14:15:50 <johill> and not really deterministic
2008-03-21T14:15:57 <johill> sometimes it checks your files and sometimes it doesn't
2008-03-21T14:15:58 <johill> sometimes it checks your files and sometimes it doesn't
2008-03-21T14:16:36 <johill> ahrg
2008-03-21T14:16:36 <johill> ahrg
2008-03-21T14:16:41 <johill> -domain:underlay search is broken?
2008-03-21T14:16:41 <johill> -domain:underlay search is broken?
2008-03-21T14:17:27 <johill> oh, no, but it's stupid
2008-03-21T14:17:27 <johill> oh, no, but it's stupid
2008-03-21T14:17:28 <johill> ok
2008-03-21T14:17:28 <johill> ok
2008-03-21T14:18:31 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3670:57b972553801 1.7-storage-hwendel/MoinMoin/ (19 files in 8 dirs):
2008-03-21T14:18:31 <CIA-39> make backends/ItemCollection use the new search.term things for listing pages
2008-03-21T14:18:31 <CIA-39> TODO: add back optimisations in various places
2008-03-21T14:18:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3670:57b972553801 1.7-storage-hwendel/MoinMoin/ (19 files in 8 dirs):
2008-03-21T14:18:32 <CIA-39> make backends/ItemCollection use the new search.term things for listing pages
2008-03-21T14:18:32 <CIA-39> TODO: add back optimisations in various places
2008-03-21T14:18:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3671:91eb17f6aa08 1.7-storage-hwendel/MoinMoin/search/term.py: add __repr__ for some more search terms
2008-03-21T14:18:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3671:91eb17f6aa08 1.7-storage-hwendel/MoinMoin/search/term.py: add __repr__ for some more search terms
2008-03-21T14:18:34 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3672:4280aec4a547 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: optimise layer backend underlay search
2008-03-21T14:18:34 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3672:4280aec4a547 1.7-storage-hwendel/MoinMoin/storage/backends/meta.py: optimise layer backend underlay search
2008-03-21T14:19:32 <ThomasWaldmann> johill: you are sure you want to merge all that stuff with the xapian fixes that should be done in main branch?
2008-03-21T14:19:33 <ThomasWaldmann> johill: you are sure you want to merge all that stuff with the xapian fixes that should be done in main branch?
2008-03-21T14:20:02 <johill> not entirely
2008-03-21T14:20:03 <johill> not entirely
2008-03-21T14:20:07 <johill> so far I havne't touched xapian
2008-03-21T14:20:07 <johill> so far I havne't touched xapian
2008-03-21T14:20:12 <johill> but will anybody do xapian fixes?
2008-03-21T14:20:12 <johill> but will anybody do xapian fixes?
2008-03-21T14:20:19 <johill> I'm 60% expecting it to be released broken
2008-03-21T14:20:19 <johill> I'm 60% expecting it to be released broken
2008-03-21T14:20:36 <ThomasWaldmann> well, somebody will have to, some time...
2008-03-21T14:20:37 <ThomasWaldmann> well, somebody will have to, some time...
2008-03-21T14:21:26 <ThomasWaldmann> for 1.6.2 i guess there won't be fixes, but maybe afterwards
2008-03-21T14:21:26 <ThomasWaldmann> for 1.6.2 i guess there won't be fixes, but maybe afterwards
2008-03-21T14:21:27 <dreimark> does eventlog know it's size or length ?
2008-03-21T14:21:27 <dreimark> does eventlog know it's size or length ?
2008-03-21T14:22:06 <johill> ThomasWaldmann: yeah but it could just as well wait until the storage merge
2008-03-21T14:22:06 <johill> ThomasWaldmann: yeah but it could just as well wait until the storage merge
2008-03-21T14:22:29 <johill> dreimark: not in 1.6/1.7 but in storage I think it's written chunked rather than line-based
2008-03-21T14:22:29 <johill> dreimark: not in 1.6/1.7 but in storage I think it's written chunked rather than line-based
2008-03-21T14:23:10 <ThomasWaldmann> johill: that could be rather long
2008-03-21T14:23:10 <ThomasWaldmann> johill: that could be rather long
2008-03-21T14:23:55 <ThomasWaldmann> (until that reaches some released version)
2008-03-21T14:23:55 <ThomasWaldmann> (until that reaches some released version)
2008-03-21T14:24:26 <johill> yes
2008-03-21T14:24:26 <johill> yes
2008-03-21T14:24:57 <johill> but on the other hand, the current code is really crufty
2008-03-21T14:24:57 <johill> but on the other hand, the current code is really crufty
2008-03-21T14:25:04 <johill> and -storage seems to work fine except you can
2008-03-21T14:25:04 <johill> and -storage seems to work fine except you can
2008-03-21T14:25:14 <johill> except you can't actually use non-moin16 storage
2008-03-21T14:25:14 <johill> except you can't actually use non-moin16 storage
2008-03-21T14:26:41 * dreimark speeds up Hits
2008-03-21T14:26:41 * dreimark speeds up Hits
2008-03-21T14:26:46 <johill> how?
2008-03-21T14:26:46 <johill> how?
2008-03-21T14:27:51 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3677:ac09769962b7 1.7-storage-hwendel/MoinMoin/ (5 files in 5 dirs): merge with main 1.7
2008-03-21T14:27:51 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3677:ac09769962b7 1.7-storage-hwendel/MoinMoin/ (5 files in 5 dirs): merge with main 1.7
2008-03-21T14:28:04 <johill> ThomasWaldmann: maybe a pre-commit pep8 hook wouldn't be all bad
2008-03-21T14:28:04 <johill> ThomasWaldmann: maybe a pre-commit pep8 hook wouldn't be all bad
2008-03-21T14:28:16 <johill> then we could get rid of that test which takes a long time and is indeterministic
2008-03-21T14:28:16 <johill> then we could get rid of that test which takes a long time and is indeterministic
2008-03-21T14:28:22 <ThomasWaldmann> yeah, i also though about it
2008-03-21T14:28:22 <ThomasWaldmann> yeah, i also though about it
2008-03-21T14:28:54 <ThomasWaldmann> the pushes will then take longer
2008-03-21T14:28:54 <ThomasWaldmann> the pushes will then take longer
2008-03-21T14:28:58 <dreimark> for the case All: return u'%d' % event_log.lines() (it doesn't need to loop)
2008-03-21T14:28:58 <dreimark> for the case All: return u'%d' % event_log.lines() (it doesn't need to loop)
2008-03-21T14:29:13 <johill> dreimark: huh ok
2008-03-21T14:29:13 <johill> dreimark: huh ok
2008-03-21T14:29:23 <johill> ThomasWaldmann: pep8 test on single files shouldn't take too long
2008-03-21T14:29:24 <johill> ThomasWaldmann: pep8 test on single files shouldn't take too long
2008-03-21T14:29:55 <ThomasWaldmann> and the sourcecode test should work quite good except if your files' mtime is wrong or all too fresh
2008-03-21T14:29:56 <ThomasWaldmann> and the sourcecode test should work quite good except if your files' mtime is wrong or all too fresh
2008-03-21T14:30:16 <dennda> did you write these pep 8 tests yourself?
2008-03-21T14:30:16 <dennda> did you write these pep 8 tests yourself?
2008-03-21T14:31:12 <johill> ThomasWaldmann: not really, the RECENTLY thing is failry bogus
2008-03-21T14:31:12 <johill> ThomasWaldmann: not really, the RECENTLY thing is failry bogus
2008-03-21T14:31:13 <ThomasWaldmann> the py.test stuff has developed from a tab checker from the pypy guys to some more featured checker
2008-03-21T14:31:13 <ThomasWaldmann> the py.test stuff has developed from a tab checker from the pypy guys to some more featured checker
2008-03-21T14:31:29 <johill> ThomasWaldmann: it should have last-check-mtime instead but that's not really possible
2008-03-21T14:31:29 <johill> ThomasWaldmann: it should have last-check-mtime instead but that's not really possible
2008-03-21T14:31:32 <ThomasWaldmann> but i used a pep8 module i found on the inet
2008-03-21T14:31:32 <ThomasWaldmann> but i used a pep8 module i found on the inet
2008-03-21T14:32:12 <ThomasWaldmann> johill: it works, except if you create fresh clones often
2008-03-21T14:32:13 <ThomasWaldmann> johill: it works, except if you create fresh clones often
2008-03-21T14:32:47 <johill> ThomasWaldmann: no, if I edit a dozen files it'll always recheck them for quite a long time
2008-03-21T14:32:47 <johill> ThomasWaldmann: no, if I edit a dozen files it'll always recheck them for quite a long time
2008-03-21T14:33:15 <johill> a week or something
2008-03-21T14:33:15 <johill> a week or something
2008-03-21T14:33:31 <johill> that's what bothers me, if I do API updates I often tocuh many files
2008-03-21T14:33:31 <johill> that's what bothers me, if I do API updates I often tocuh many files
2008-03-21T14:33:53 <johill> alternatively I could make it write an extended attribute to the file (where possible) :)
2008-03-21T14:33:53 <johill> alternatively I could make it write an extended attribute to the file (where possible) :)
2008-03-21T14:34:19 <johill> with the last checked mtime
2008-03-21T14:34:19 <johill> with the last checked mtime
2008-03-21T14:35:04 <ThomasWaldmann> be happy that it does not check ALL files all the time >:)
2008-03-21T14:35:04 <ThomasWaldmann> be happy that it does not check ALL files all the time >:)
2008-03-21T14:35:22 <johill> doesn't take that much longer ;)
2008-03-21T14:35:23 <johill> doesn't take that much longer ;)
2008-03-21T14:39:45 <PierceTyler> how much time is left before the 1.7 code freeze?
2008-03-21T14:39:45 <PierceTyler> how much time is left before the 1.7 code freeze?
2008-03-21T14:43:51 <dreimark> PierceTyler: major code freeze is already
2008-03-21T14:43:52 <dreimark> PierceTyler: major code freeze is already
2008-03-21T14:43:57 <dreimark> why do you ask ?
2008-03-21T14:43:57 <dreimark> why do you ask ?
2008-03-21T14:44:18 <dreimark> PierceTyler:
2008-03-21T14:44:18 <dreimark> PierceTyler:
2008-03-21T14:44:50 <PierceTyler> dreimark: we are interested in the DocBook features. very much want ImportDocBook utility merged from 1.6 branch if possible
2008-03-21T14:44:50 <PierceTyler> dreimark: we are interested in the DocBook features. very much want ImportDocBook utility merged from 1.6 branch if possible
2008-03-21T14:45:33 <dreimark> I do see only whitespaces
2008-03-21T14:45:33 <dreimark> I do see only whitespaces
2008-03-21T14:45:45 <PierceTyler> ?
2008-03-21T14:45:45 <PierceTyler> ?
2008-03-21T14:46:08 <dennda> it was a color-message
2008-03-21T14:46:09 <dennda> it was a color-message
2008-03-21T14:46:10 <johill> me too
2008-03-21T14:46:10 <johill> me too
2008-03-21T14:46:12 <johill> turn off colours
2008-03-21T14:46:12 <johill> turn off colours
2008-03-21T14:46:19 <PierceTyler> is that better?
2008-03-21T14:46:19 <PierceTyler> is that better?
2008-03-21T14:47:11 <dreimark> it is in the log readable http://moinmo.in/MoinMoinChat/Logs/moin-dev/2008-03-21
2008-03-21T14:47:12 <dreimark> it is in the log readable http://moinmo.in/MoinMoinChat/Logs/moin-dev/2008-03-21
2008-03-21T14:47:15 <johill> yes
2008-03-21T14:47:15 <johill> yes
2008-03-21T14:47:25 <PierceTyler> am i still emitting colors?
2008-03-21T14:47:26 <PierceTyler> am i still emitting colors?
2008-03-21T14:47:31 <dennda> no
2008-03-21T14:47:32 <dennda> no
2008-03-21T14:48:20 <dennda> why does the log contain every message twice?
2008-03-21T14:48:21 <dennda> why does the log contain every message twice?
2008-03-21T14:48:52 <dreimark> it is a bug one has to fix it ;)
2008-03-21T14:48:52 <dreimark> it is a bug one has to fix it ;)
2008-03-21T14:48:56 <PierceTyler> dreimark: again, the reason i ask is we are interested in the DocBook features. would like ImportDocBook utlility added to 1.7 if possible
2008-03-21T14:48:56 <PierceTyler> dreimark: again, the reason i ask is we are interested in the DocBook features. would like ImportDocBook utlility added to 1.7 if possible
2008-03-21T14:48:56 <dennda> is it because we have two moinbots in here? ;)
2008-03-21T14:48:57 <dennda> is it because we have two moinbots in here? ;)
2008-03-21T14:49:04 <johill> dennda: heh funky, probably
2008-03-21T14:49:04 <johill> dennda: heh funky, probably
2008-03-21T14:50:00 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3391:eea0edb31d9d 1.7/MoinMoin/_tests/test_sourcecode.py: pep8 test: use extended attributes (where available) to store last-tested-mtime
2008-03-21T14:50:00 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3391:eea0edb31d9d 1.7/MoinMoin/_tests/test_sourcecode.py: pep8 test: use extended attributes (where available) to store last-tested-mtime
2008-03-21T14:50:01 <johill> there. test imrpoved
2008-03-21T14:50:01 <johill> there. test imrpoved
2008-03-21T14:50:32 <dreimark> PierceTyler: mvirkkil is working on that branch and did already lots of clean commits into 1.7
2008-03-21T14:50:33 <dreimark> PierceTyler: mvirkkil is working on that branch and did already lots of clean commits into 1.7
2008-03-21T14:51:01 <johill> test now takes < 3 seconds even with RECENTLY = 0
2008-03-21T14:51:01 <johill> test now takes < 3 seconds even with RECENTLY = 0
2008-03-21T14:51:09 * johill <3's his filesystem
2008-03-21T14:51:09 * johill <3's his filesystem
2008-03-21T14:51:26 <PierceTyler> dreimark: yes, i know. i have been talking by email to mikko. but no word from him past few days. he said ImportDocBook was in "pretty good shape" and he was considering merging it, but it's not merged yet
2008-03-21T14:51:26 <PierceTyler> dreimark: yes, i know. i have been talking by email to mikko. but no word from him past few days. he said ImportDocBook was in "pretty good shape" and he was considering merging it, but it's not merged yet
2008-03-21T14:51:36 <johill> bbl
2008-03-21T14:51:36 <johill> bbl
2008-03-21T14:51:52 <PierceTyler> as far as i know
2008-03-21T14:51:52 <PierceTyler> as far as i know
2008-03-21T14:52:27 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3679:8ef1f5b392f9 1.7-storage-hwendel/MoinMoin/_tests/test_sourcecode.py: merge pep8 test improvement from main
2008-03-21T14:52:27 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3679:8ef1f5b392f9 1.7-storage-hwendel/MoinMoin/_tests/test_sourcecode.py: merge pep8 test improvement from main
2008-03-21T14:52:46 <xorAxAx> dennda: hehe, ThomasWaldmann is the admin :)
2008-03-21T14:52:46 <xorAxAx> dennda: hehe, ThomasWaldmann is the admin :)
2008-03-21T14:53:07 <dreimark> PierceTyler: http://hg.moinmo.in/moin/1.7/rev/7ff3c643631d
2008-03-21T14:53:07 <dreimark> PierceTyler: http://hg.moinmo.in/moin/1.7/rev/7ff3c643631d
2008-03-21T14:53:18 <dennda> xorAxAx: of the bots, you mean?
2008-03-21T14:53:18 <dennda> xorAxAx: of the bots, you mean?
2008-03-21T14:53:54 <ThomasWaldmann> PierceTyler: i think it is rather uncritical adding ImportDocBook, but there should be docs about it and it should have been tested that it works ok
2008-03-21T14:53:54 <ThomasWaldmann> PierceTyler: i think it is rather uncritical adding ImportDocBook, but there should be docs about it and it should have been tested that it works ok
2008-03-21T14:54:22 <ThomasWaldmann> see MoinMoin:DocBook for some hints
2008-03-21T14:54:22 <ThomasWaldmann> see MoinMoin:DocBook for some hints
2008-03-21T14:54:26 <xorAxAx> dennda: yes
2008-03-21T14:54:26 <xorAxAx> dennda: yes
2008-03-21T14:54:33 <xorAxAx> dennda: he should fix it
2008-03-21T14:54:34 <xorAxAx> dennda: he should fix it
2008-03-21T14:55:13 <PierceTyler> ThomasWaldmann: i have been reading docs closely for a couple weeks now. ImportDocBook seems critical to me from our perspective
2008-03-21T14:55:13 <PierceTyler> ThomasWaldmann: i have been reading docs closely for a couple weeks now. ImportDocBook seems critical to me from our perspective
2008-03-21T14:55:15 <dennda> well, it's easy to find out if the second bot causes the "bug" :)
2008-03-21T14:55:16 <dennda> well, it's easy to find out if the second bot causes the "bug" :)
2008-03-21T14:56:58 <PierceTyler> ThomasWaldmann: i mean critical because we are preparing to launch a wiki site and i need to load DocBook content into wiki markup
2008-03-21T14:56:58 <PierceTyler> ThomasWaldmann: i mean critical because we are preparing to launch a wiki site and i need to load DocBook content into wiki markup
2008-03-21T14:59:41 <PierceTyler> i can test ImportDocBook extensively.
2008-03-21T14:59:41 <PierceTyler> i can test ImportDocBook extensively.
2008-03-21T15:00:14 <dennda> test
2008-03-21T15:00:16 <dennda> 123
2008-03-21T15:01:28 <dennda> welcome back moinBot
2008-03-21T15:01:45 <PierceTyler> BuildBook is not critical, and also needs more work, mikko told me
2008-03-21T15:03:10 <ThomasWaldmann> PierceTyler: those pages look a bit strange
2008-03-21T15:04:40 <ThomasWaldmann> but that test wiki likely runs on rather old code
2008-03-21T15:05:13 <PierceTyler> could be. it was possible to test ImportDocBook on it
2008-03-21T15:05:21 <ThomasWaldmann> PierceTyler: if you can test it extensively, just tell me when you are happy with it
2008-03-21T15:05:27 <ThomasWaldmann> (because i cant)
2008-03-21T15:05:48 <PierceTyler> i can definitely let you know. i have plenty of use cases ready to go into production
2008-03-21T15:08:28 <perryclar1> Yeah we do...
2008-03-21T15:08:31 <ThomasWaldmann> http://moinmo.in/DocBook that page maybe needs some updates
2008-03-21T15:08:41 <ThomasWaldmann> and some less "todo"
2008-03-21T15:09:30 <PierceTyler> once we are up and running, i can help with updating the docs at http://moinmo.in/DocBook
2008-03-21T15:14:34 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3680:030d644c574e 1.7-storage-hwendel/MoinMoin/storage/ (4 files in 3 dirs): clean up DELETED metadata handling (including a bugfix)
2008-03-21T15:18:53 <ThomasWaldmann> perryclar1: ImportDocbook.py has quite some XXX and TODO. It will also need some changes for stuff that changed after 1.6.
2008-03-21T15:19:31 <ThomasWaldmann> I can help you with the updating if there is someone else doing the docs / docbook related stuff.
2008-03-21T15:19:57 <dennda> what does XXX mean (in this context)?
2008-03-21T15:20:14 <ThomasWaldmann> something to look at :)
2008-03-21T15:21:01 <dennda> i see :)
2008-03-21T15:21:07 <ThomasWaldmann> also, i guess the transformation docbook > wiki will need some update for the new wiki link markup
2008-03-21T15:25:08 <perryclar1> So, are you saying that you will merge the ImportDocBook utility into 1.7 so that we can test it, OR are you saying that "These items need to be worked on, then I can merge it."?
2008-03-21T15:28:04 <johill> the query parser code is f'ed
2008-03-21T15:28:38 <ThomasWaldmann> i would prefer having it in a separate repo and merge it into main when working and documented
2008-03-21T15:30:23 <johill> lanius: ping
2008-03-21T15:30:36 <dreimark> 1.7-extensions/ ?
2008-03-21T15:35:18 <johill> lanius: never mind
2008-03-21T15:36:16 <dreimark> johill: can you please run py.test in a fresh testwiki
2008-03-21T15:36:24 <dreimark> I do get 19 failures
2008-03-21T15:37:35 * johill tries
2008-03-21T15:37:40 <johill> I get about 2 failures iirc
2008-03-21T15:37:49 <dreimark> ok I take new clone
2008-03-21T15:38:07 <johill> haha
2008-03-21T15:38:12 <johill> no I get 19 now too
2008-03-21T15:38:17 <johill> seems teh configuration is crapped
2008-03-21T15:39:03 <johill> ThomasWaldmann: you forgot to update the tests for the new underlay stuff
2008-03-21T15:39:32 <ThomasWaldmann> ehrm, the .tar?
2008-03-21T15:39:49 <johill> it tries to untar the .tar.bz2
2008-03-21T15:39:51 <johill> not the .tar
2008-03-21T15:39:57 <ThomasWaldmann> ok, will fix it
2008-03-21T15:40:44 <johill> never mind, I have a fix
2008-03-21T15:42:12 <johill> dreimark: that was pretty obvious from the test failures, no?
2008-03-21T15:42:13 <ThomasWaldmann> it's likely a single line change in maketestwiki
2008-03-21T15:42:16 <johill> yes
2008-03-21T15:42:48 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3392:48b95abf3994 1.7/MoinMoin/_tests/maketestwiki.py: fix tests: need to untar underlay.tar not underlay.tar.bz2
2008-03-21T15:43:26 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3682:9fcb7c9e5574 1.7-storage-hwendel/MoinMoin/_tests/maketestwiki.py: merge test fix from main 1.7
2008-03-21T15:43:34 <johill> tests finished: 909 passed, 2 failed, 65 skipped in 171.18 seconds
2008-03-21T15:43:47 <ThomasWaldmann> johill: you fix 1.6 also?
2008-03-21T15:43:56 <johill> oh, not yet, want me to?
2008-03-21T15:44:11 <ThomasWaldmann> i can also just commit :)
2008-03-21T15:44:31 <johill> go ahead
2008-03-21T15:44:41 <johill> my 1.6 isn't even up-to-date
2008-03-21T15:45:17 <ThomasWaldmann> mode='r' is default?
2008-03-21T15:46:12 <ThomasWaldmann> it is
2008-03-21T15:46:23 <johill> at least in our version, yes
2008-03-21T15:47:36 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2591:c157c20087d7 1.6/MoinMoin/_tests/maketestwiki.py: fix tests: need to untar underlay.tar not underlay.tar.bz2
2008-03-21T15:56:49 * dreimark dislikes copiing the whole data dir for tests
2008-03-21T15:57:57 <johill> I thought it just creates an empty one
2008-03-21T15:58:02 <dreimark> no
2008-03-21T15:58:19 <johill> oh
2008-03-21T15:58:25 <johill> that's why the tests are so weird then
2008-03-21T15:58:30 <johill> ;)
2008-03-21T15:59:01 <johill> lanius: ping again :)
2008-03-21T16:03:17 <lanius> johill: pong
2008-03-21T16:04:41 <johill> lanius: need more thoughts about underlay, especially scripted underlay update
2008-03-21T16:05:01 <johill> also, I have use cases for copy-
2008-03-21T16:05:08 <johill> copy-on-write semantics
2008-03-21T16:06:00 <johill> lanius: I have implemented the filter language, use it in backends, and removed the is_underlay stuff now
2008-03-21T16:06:57 <johill> and I have a question about the indexed backend
2008-03-21T16:07:07 <johill> oh no, never mind
2008-03-21T16:08:53 <lanius> yes
2008-03-21T16:12:54 <johill> any ideas how to do either?
2008-03-21T16:24:03 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3683:685902b9b0b1 1.7-storage-hwendel/MoinMoin/storage/backends/common.py: explain _get_item_metadata_cache
2008-03-21T16:24:03 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3684:db9daed814b1 1.7-storage-hwendel/MoinMoin/ (5 files in 3 dirs): rename Term.reset to Term.prepare
2008-03-21T16:33:22 <lanius> johill: scripted underlay update can be done by only using multiconfig.uderlay_backend
2008-03-21T16:33:32 <johill> yeah but I removed that ;)
2008-03-21T16:33:56 <johill> I guess I'll have to add it back but with a big fat note to kill it if one doesn't want underlay
2008-03-21T16:34:09 <lanius> nono you didn't remove that
2008-03-21T16:34:18 <lanius> you removed the item collectoin
2008-03-21T16:34:21 <johill> because the config code you had would have an underlay_backend even if not layered in data_backend
2008-03-21T16:34:52 <johill> no, I did remove that, it's a local variable only now
2008-03-21T16:34:57 <lanius> but the underlay backend is still there, so you can use it
2008-03-21T16:35:34 <johill> but you don't know about it directly
2008-03-21T16:35:44 <johill> only through the layered
2008-03-21T16:35:55 <lanius> but you can get it from the config
2008-03-21T16:36:04 <lanius> (there is still the addUnderlay method)
2008-03-21T16:36:28 <johill> hmm
2008-03-21T16:36:51 <johill> I think the config stuff needs to be changed a bit
2008-03-21T16:37:17 <johill> to postprocess the config the user gave
2008-03-21T16:37:27 <lanius> why
2008-03-21T16:37:34 <johill> rather than pre-process it
2008-03-21T16:37:48 <johill> because there's no way to know what the user intended
2008-03-21T16:38:00 <lanius> intended?
2008-03-21T16:38:13 <johill> say I configure data_backends to no layered one
2008-03-21T16:38:18 <lanius> yes
2008-03-21T16:38:21 <johill> what should the update script do if invoked?>
2008-03-21T16:38:38 <johill> (nothing, of course, but how?)
2008-03-21T16:38:39 <lanius> two possibilities
2008-03-21T16:39:03 <lanius> a.) write everthing to the only data_backend
2008-03-21T16:39:06 <lanius> b.) do nothing
2008-03-21T16:39:33 <lanius> c.) manually create a backend for the underlay dir
2008-03-21T16:39:59 <johill> imho the only correct thing is (b)
2008-03-21T16:40:09 <johill> and data_underlay_dir shouldn't ever be used anywhere but in the config
2008-03-21T16:40:16 <lanius> yes
2008-03-21T16:40:19 <johill> that's why I say postprocess
2008-03-21T16:40:27 <johill> if the user doesn't set data_backend
2008-03-21T16:40:37 <johill> then we create one for data_dir and one for data_underlay_dir
2008-03-21T16:40:52 <johill> otherwise, we don't do anything and leave it up to the user
2008-03-21T16:41:06 <johill> and we can have an underlay_backend config too
2008-03-21T16:41:14 <johill> which would be used by the update code
2008-03-21T16:41:29 <johill> which we *only* assign when the user didn't any backends at all
2008-03-21T16:41:58 <johill> the code you had before *always* assigned underlay_backend even if the user configured own data_backend
2008-03-21T16:42:20 <lanius> not setting data_backend is a configuration error
2008-03-21T16:42:39 <lanius> that's why multiconfig has defaults
2008-03-21T16:43:00 <johill> eh
2008-03-21T16:43:05 <johill> not setting it isn't an error if there are defaults
2008-03-21T16:43:17 <johill> the point is
2008-03-21T16:43:17 <lanius> define what you mean by not set
2008-03-21T16:43:25 <johill> if you set data_backend but not underlay_backend
2008-03-21T16:43:30 <lanius> where
2008-03-21T16:43:31 <johill> then the two may have no relation
2008-03-21T16:43:33 <johill> in your config
2008-03-21T16:43:36 <lanius> which config
2008-03-21T16:43:40 <johill> any wikiconfig.py
2008-03-21T16:44:15 <lanius> so we have the current defaults in multiconfig which are just like you describ
2008-03-21T16:44:16 <lanius> e
2008-03-21T16:44:36 <johill> yeah
2008-03-21T16:44:39 <johill> I know how to fix it now :)
2008-03-21T16:51:24 <johill> uh
2008-03-21T16:51:28 <johill> where is cfg.user_dir set?
2008-03-21T16:51:44 <lanius> somewhere with some magi
2008-03-21T16:51:44 <lanius> c
2008-03-21T16:51:50 <johill> oh
2008-03-21T16:52:30 <ThomasWaldmann> mvirkkil: are you there? maybe join #moin also.
2008-03-21T17:10:11 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3393:93aa3b53ddc9 1.7/MoinMoin/config/_tests/test_configs.py:
2008-03-21T17:10:11 <CIA-39> add configuration tests
2008-03-21T17:10:11 <CIA-39> currently only one checking that data_underlay_dir = None won't screw up right away
2008-03-21T17:11:55 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3686:c9c9bbb274bf 1.7-storage-hwendel/MoinMoin/config/_tests/test_configs.py: pull config test from main
2008-03-21T17:11:56 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3687:e66af1d8e789 1.7-storage-hwendel/MoinMoin/config/multiconfig.py: make backend configuration defaults smarter (fixes test)
2008-03-21T17:12:00 <johill> lanius: there :)
2008-03-21T17:12:26 * ThomasWaldmann curses screen/mc/ssh/whatever
2008-03-21T17:13:40 <johill> what for?
2008-03-21T17:15:23 <johill> ok
2008-03-21T17:15:28 <johill> that fixes the underlay issue completely
2008-03-21T17:15:28 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3688:0be76939ebab 1.7-storage-hwendel/MoinMoin/packages.py: packages can use underlay_backend now
2008-03-21T17:15:30 <johill> afaict
2008-03-21T17:15:52 <johill> lanius: what about copy-on-write?
2008-03-21T17:25:52 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3689:43042627cb4e 1.7-storage-hwendel/MoinMoin/ (config/multiconfig.py storage/backends/meta.py): give LayerBackend a copy_on_write configuration
2008-03-21T17:44:10 <johill> copy-on-write is.. interesting
2008-03-21T17:44:30 <johill> what if the underlayed data adds a revision? do all revision numbers change?
2008-03-21T17:47:47 <johill> or can we implemented non-linear revisoining :)
2008-03-21T18:05:54 <dreimark> I do want to change maketestwiki by http://moinmo.in/ReimarBauer?action=AttachFile&do=view&target=maketestwiki.patch
2008-03-21T18:07:04 <johill> why don't you filter for the unwanted ones before rather than deleting them later?
2008-03-21T18:07:25 <johill> wait, you do, why bother removing them then?
2008-03-21T18:09:08 <dreimark> cp -r of some dirs and touching some files
2008-03-21T18:09:19 <dreimark> and creating of empty dirs
2008-03-21T18:11:40 <johill> oh
2008-03-21T18:12:14 <dreimark> We may want to think about using a data.tar too
2008-03-21T18:12:28 <dreimark> this fixes now only the tests and not an installation
2008-03-21T18:13:34 <dreimark> if you try out in a DektopEdition and then do setup install
2008-03-21T18:14:16 <dreimark> you can easily get all your examples on all new wiki instances
2008-03-21T18:14:19 <johill> I think we should have a code snippet that creates a valid data dir instead
2008-03-21T18:14:34 <johill> that way we can also upgrade easily to create missing dirs etc
2008-03-21T18:16:19 <dreimark> hmm, we don't have only a dir structure to create. e.g. the plugin dir has files
2008-03-21T18:16:49 <johill> indeed, I think I changed the code though to not require plugin modules
2008-03-21T18:16:59 <johill> so just having an empty plugin dir would be sufficient
2008-03-21T18:18:53 <dreimark> but we should be able to test if plugins could be used, e.g. all this plugin import methods
2008-03-21T18:20:20 <johill> lanius: wtf, why do you index on 'name'?
2008-03-21T18:22:41 <johill> oh that's for users?
2008-03-21T18:25:34 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3394:396d3434ab92 1.7/MoinMoin/Page.py: move some code under openid_server_enabled where it belongs
2008-03-21T18:26:33 <dreimark> bbl
2008-03-21T18:38:24 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3690:caea6833d0da 1.7-storage-hwendel/docs/CHANGES.storage: update CHANGES
2008-03-21T18:38:25 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3691:edcc83a384cf 1.7-storage-hwendel/MoinMoin/search/term.py: correct some __repr__ errors
2008-03-21T18:40:34 <johill> lanius: why is indexed backend implicit? I think it would better be explicit
2008-03-21T18:47:17 <lanius> johill: no reason
2008-03-21T18:48:32 <johill> think integrated search and storage
2008-03-21T18:48:39 <johill> if you have xapian, you want it between index and filesystem
2008-03-21T18:48:50 <johill> hm otoh, how does xapian work here?
2008-03-21T18:48:56 <johill> I guess you'd want it *instead* of index
2008-03-21T18:49:27 <lanius> yes i guess that too
2008-03-21T18:50:48 <johill> I also wonder if the indexed backend should save a list of all item names
2008-03-21T18:51:00 <johill> would probably be faster than readdir()
2008-03-21T18:51:10 <lanius> that is the question
2008-03-21T18:51:19 <lanius> we would need a test to assure that
2008-03-21T18:52:13 <johill> keep in mind that most filesystems don't cache readdir()
2008-03-21T18:52:31 <johill> while the index db would be cached
2008-03-21T18:52:42 <lanius> would it?
2008-03-21T18:52:43 <xorAxAx> johill: how do you want to invalidate the cache?
2008-03-21T18:52:52 <johill> xorAxAx: hm?
2008-03-21T18:52:58 <xorAxAx> johill: whats unclear?
2008-03-21T18:53:00 <johill> which cache?
2008-03-21T18:53:01 <lanius> you mean cached by the filesystem directly?
2008-03-21T18:53:05 <johill> yes
2008-03-21T18:53:05 <xorAxAx> 18:50:48 < johill> I also wonder if the indexed backend should save a list of all item names
2008-03-21T18:53:11 <xorAxAx> thats a cache
2008-03-21T18:53:15 <johill> lanius: yes, in the kernel
2008-03-21T18:53:18 <johill> xorAxAx: yes
2008-03-21T18:53:21 <xorAxAx> johill: so?
2008-03-21T18:53:26 <johill> xorAxAx: the indexed backend layers above it so it knows when a new page is added/removed
2008-03-21T18:53:30 <johill> where's the problem?
2008-03-21T18:53:52 <xorAxAx> johill: how does it know? note that any other process could mutate the data
2008-03-21T18:54:15 <johill> then you can throw away the indexed backend completely
2008-03-21T18:54:15 <lanius> xorAxAx: but it writes to a general file
2008-03-21T18:54:17 * johill deletes code
2008-03-21T18:54:28 <xorAxAx> johill: ?
2008-03-21T18:54:38 <xorAxAx> lanius: well, does it "cache" a list of items?
2008-03-21T18:54:40 <johill> other code could also change metadata keys
2008-03-21T18:54:41 <xorAxAx> in memory
2008-03-21T18:54:49 <johill> xorAxAx: ehh, it uses bdb
2008-03-21T18:54:53 <lanius> xorAxAx: not in memory
2008-03-21T18:54:55 <xorAxAx> johill: so where is the problem?
2008-03-21T18:55:06 <xorAxAx> well, it sounded like you want to do it in ram :)
2008-03-21T18:55:11 <johill> you're making it one
2008-03-21T18:55:32 <xorAxAx> i misunderstood you
2008-03-21T18:55:41 <johill> I never wanted to do it in ram :)
2008-03-21T18:55:43 <johill> ok
2008-03-21T18:56:04 <johill> now with the index optimisations I get rather strange test failures
2008-03-21T18:56:04 <johill> wtf
2008-03-21T18:56:46 <johill> oh
2008-03-21T18:56:52 <johill> py.test doesn't like being run concurrently?
2008-03-21T19:01:30 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3692:7de9938e8b03 1.7-storage-hwendel/MoinMoin/search/term.py: add term.BOOL
2008-03-21T19:01:31 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3693:296a0481ac38 1.7-storage-hwendel/MoinMoin/storage/_tests/test_backends_meta.py: add copy_on_write parameter to LayerBackend test
2008-03-21T19:01:32 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3694:41dde8f3daa0 1.7-storage-hwendel/MoinMoin/Page.py:
2008-03-21T19:01:32 <CIA-39> fix getPageCount bug
2008-03-21T19:01:34 <CIA-39> why didn't this show up in earlier testing? must be a bug elsewhere
2008-03-21T19:04:45 <johill> why do the search tests take forever?
2008-03-21T19:05:56 <xorAxAx> johill: py.test doesnt care
2008-03-21T19:06:08 <johill> xorAxAx: then something else does
2008-03-21T19:06:20 <johill> xorAxAx: I was getting output from failed test X and Y while running only test Z
2008-03-21T19:06:38 <xorAxAx> how did ensure to run only Z?
2008-03-21T19:06:50 <johill> py.test path/to/Z
2008-03-21T19:07:05 <xorAxAx> well, that will run all test from that module, maybe somebody imported others?
2008-03-21T19:07:10 <johill> no
2008-03-21T19:08:10 <johill> maybe I read the messages wrongly
2008-03-21T19:10:43 <johill> hrm
2008-03-21T19:10:54 <johill> the indexed backend doesn't seem to invalidate the cache properly
2008-03-21T19:13:34 <johill> ah, I see
2008-03-21T19:17:15 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3695:aad6b7112158 1.7-storage-hwendel/MoinMoin/storage/backends/filesystem.py: actually make use of indexes in indexing backend again
2008-03-21T19:17:16 <johill> ok, done
2008-03-21T19:17:27 <johill> that query stuff is pretty nice
2008-03-21T19:17:38 <johill> if anybody tells me how to use xapian I'll port that
2008-03-21T19:20:01 <johill> down to the regular two test failures
2008-03-21T19:20:12 <johill> dennda: what is your proposal about then?
2008-03-21T19:20:45 <xorAxAx> johill: did you read the helppage?
2008-03-21T19:21:14 <johill> you mean help on xapian?
2008-03-21T19:24:21 <xorAxAx> yes
2008-03-21T19:25:29 <johill> that's not very helpful
2008-03-21T19:26:55 <johill> hm. I wonder if a page's category should be treated as metadata
2008-03-21T19:27:20 <johill> categories
2008-03-21T19:30:25 <xorAxAx> johill: it was enough for me to get it running
2008-03-21T19:30:36 <johill> ah
2008-03-21T19:30:39 <johill> you misunderstood me again
2008-03-21T19:31:00 <johill> I want to use it from code, not just as an end user :)
2008-03-21T19:33:50 <xorAxAx> well, what exactly?
2008-03-21T19:33:58 <xorAxAx> xapwrap? the official bindings?
2008-03-21T19:34:31 <johill> good question
2008-03-21T19:34:34 <johill> how does it currently work?
2008-03-21T19:35:00 <johill> or not work ;)
2008-03-21T19:35:31 <xorAxAx> moin is using xapwrap
2008-03-21T19:35:43 <xorAxAx> but why isnt that obvious from the source?
2008-03-21T19:38:13 <johill> that part is
2008-03-21T19:38:23 <johill> but there's no high-level overview over how it works
2008-03-21T19:39:46 <xorAxAx> is there any such overview for parts of moin? :)
2008-03-21T19:40:02 <johill> there is for most of storage :)
2008-03-21T19:40:25 <johill> you make it sound like there shouldn't be
2008-03-21T19:40:32 <johill> I'm just asking. feel free to admit ignorance :)
2008-03-21T19:42:39 <xorAxAx> no, i am not implying that
2008-03-21T19:43:15 <xorAxAx> if you have some concrete questions, you can ask ThomasWaldmann who wrote the first parts of the xapian integration IIRC
2008-03-21T19:43:43 <johill> no, the point is that I don't know what to ask yet :)
2008-03-21T19:53:08 <johill> ahem
2008-03-21T19:53:09 <johill> Your search query "category:CategoryFeatureImplemented" didn't return any results. Please change some terms and refer to HelpOnSearching for more information.
2008-03-21T19:53:50 <johill> it should be category:FeatureImplemented
2008-03-21T19:54:17 <xorAxAx> ?
2008-03-21T19:54:35 <xorAxAx> category:FeatureImplemented returns something? then its broken
2008-03-21T19:54:38 <johill> yes
2008-03-21T19:54:54 <johill> the docs say it must be category:namew/oprefix
2008-03-21T19:54:57 <xorAxAx> because moin cannot determine which string to strip and it seems to strip the same token
2008-03-21T19:55:07 <johill> yah I was just going to say that
2008-03-21T19:59:23 <dennda> johill: hum?
2008-03-21T20:00:07 <johill> dennda: you earlier said you were writing an application, just curious, my latency is high today ;_
2008-03-21T20:00:23 <dennda> um
2008-03-21T20:00:26 <dennda> did I say that?
2008-03-21T20:00:45 <dennda> aj
2008-03-21T20:00:50 <dennda> that kind of "application"
2008-03-21T20:01:10 <dennda> thought you were talking about a program
2008-03-21T20:01:29 <dennda> I think I will file two applications for moin when google starts accepting student applications
2008-03-21T20:01:37 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3395:d5ad5578b2a5 1.7/MoinMoin/search/queryparser.py: fix category search
2008-03-21T20:01:50 <johill> ThomasWaldmann: want to port that to 1.6? :)
2008-03-21T20:01:58 <dennda> one for the doodle calendar thing and one for the extension of the moinmoin storage engine
2008-03-21T20:03:00 <johill> :)
2008-03-21T20:11:35 * ThomasWaldmann looks
2008-03-21T20:12:20 <dennda> but actually I just filled out your developer application thing. it's still missing information
2008-03-21T20:12:35 <johill> wait that might not be right either
2008-03-21T20:13:17 <johill> it's still missing some places
2008-03-21T20:13:54 <johill> I don't understand this
2008-03-21T20:14:01 <johill> # oh, let's better call xapian if we encounter this nasty regexp ;)
2008-03-21T20:14:01 <johill> if not category:
2008-03-21T20:14:01 <johill> cat_re = re.compile(r'----\(-\*\)\(\\r\)\?\\n\)\(\.\*\)Category(.*)\\b', re.U)
2008-03-21T20:14:04 <johill> [..]
2008-03-21T20:14:14 <johill> 'smart' comments like that make me feel dumb
2008-03-21T20:15:00 <johill> oh
2008-03-21T20:15:02 <johill> I see now
2008-03-21T20:15:03 <johill> wtf
2008-03-21T20:15:15 <johill> is that supposed to be somehow backward compatible for existing category search linkis?
2008-03-21T20:15:18 <johill> -i
2008-03-21T20:15:21 <johill> let's just kill that code
2008-03-21T20:16:53 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3396:b30370fa2f41 1.7/MoinMoin/search/queryparser.py: queryparser: kill nasty attempt to detect category searches
2008-03-21T20:17:30 <ThomasWaldmann> johill: in your cs, isn't there some \b|^ missing?
2008-03-21T20:18:25 <johill> I suppose, it wasn't there before though
2008-03-21T20:20:40 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3397:2377e18e4a10 1.7/MoinMoin/search/queryparser.py: make category search match the category indicator properly
2008-03-21T20:22:32 <johill> oh. I see why that stupid category thing was there
2008-03-21T20:22:34 <johill> how dumb!
2008-03-21T20:25:07 <johill> and now the query parser is fucking it up anyway
2008-03-21T20:25:12 <johill> oh god
2008-03-21T20:25:24 * johill just wants to rip out all that code
2008-03-21T20:25:33 <johill> <<FullSearchCached(regex:(----(-*)(\r)?\n)(.*)CategoryHomepage\b)>>
2008-03-21T20:25:38 <johill> the query parser will see that as
2008-03-21T20:26:03 <johill> regex:(----(-*)(\r)?\n) (.*) CategoryHomepage\n
2008-03-21T20:26:09 <johill> and search for an OR between those three for some reason
2008-03-21T20:27:35 <johill> now
2008-03-21T20:27:37 <johill> WHY??
2008-03-21T20:27:46 <johill> there's supposed to be an implicit AND anyway
2008-03-21T20:28:45 <ThomasWaldmann> btw, while you fix stuff, could you also add some tests?
2008-03-21T20:29:26 <johill> I think that's beyond fixing
2008-03-21T20:30:55 <xorAxAx> johill: hmm, once upon a time, that worked :)
2008-03-21T20:31:11 <xorAxAx> in the old days of 1.3 when i wrote that regex :)
2008-03-21T20:31:12 <johill> xorAxAx: it should be category:CategoryHomepage anywya
2008-03-21T20:31:22 <johill> (and then it'd actually still work)
2008-03-21T20:31:29 <xorAxAx> well, still the parser is broken
2008-03-21T20:31:36 <johill> yeah
2008-03-21T20:31:47 <johill> too bad the code is spaghetti
2008-03-21T20:32:47 <johill> I'm seriously considering rewriting it
2008-03-21T20:32:56 <johill> I'd have that done (including tests) before understanding the current mess
2008-03-21T20:33:21 <johill> that would, of course, intially break xapian
2008-03-21T20:33:27 <lanius> i guess you even would have it rewritten in the time you waste considering
2008-03-21T20:33:41 <johill> heh
2008-03-21T20:35:39 <johill> in fact
2008-03-21T20:35:44 <johill> I think I already have a lot of it
2008-03-21T20:35:53 <johill> in the wikiutil.parse
2008-03-21T20:35:57 <johill> _quoted_separated
2008-03-21T20:35:58 <johill> bbl, phone
2008-03-21T20:36:44 <johill> oh, no real call. heh
2008-03-21T20:38:17 <johill> hm, no, that function can't be used
2008-03-21T20:44:58 <johill> not unmodified
2008-03-21T20:52:47 <kikka> Heya.
2008-03-21T20:55:30 <johill> hi
2008-03-21T20:56:21 <kikka> Hello johill.
2008-03-21T20:59:13 <dennda> I need a bit of advice
2008-03-21T21:00:10 <dennda> My checkout of 1.7 hwendell storage contains 5270 files & folders. Is there any wiki page that explains the internals of moinmoin? Or any special file that I should start with?
2008-03-21T21:00:57 <johill> dennda: unfortunately not much, what do you want to know
2008-03-21T21:01:20 <dennda> oh, easy
2008-03-21T21:01:22 <dennda> just how it works
2008-03-21T21:01:23 <dennda> :)
2008-03-21T21:01:58 <johill> well MoinMoin.request contains where the stuff starts
2008-03-21T21:02:06 <johill> it's pretty horrible code ;)
2008-03-21T21:02:24 <johill> that calls out to actions (MoinMoin.action)
2008-03-21T21:02:30 <johill> possibly do
2008-03-21T21:02:37 <johill> normally do_show
2008-03-21T21:26:26 <dennda> i see a blank except there :)
2008-03-21T22:01:04 <ThomasWaldmann> johill: + return u'(\b%s\b)' % self._pattern \b is a backspace if you don't use raw strings
2008-03-21T22:01:35 <johill> ahrg
2008-03-21T22:01:45 <johill> (I always use raw strings for regexes in my code)
2008-03-21T22:06:06 <CIA-39> Johannes Berg <johannes AT sipsolutions DOT net> default * 3398:f212fe993c14 1.7/MoinMoin/search/queryparser.py: fix category highlight re
2008-03-21T22:07:00 <johill> xorAxAx: wouldn't it be better to make categories proper metadata?
2008-03-21T22:07:21 <xorAxAx> johill: how unwikilike :)
2008-03-21T22:07:37 <xorAxAx> well, thats first of all an UI question IMHO :)
2008-03-21T22:07:43 <johill> yeah
2008-03-21T22:07:59 <xorAxAx> (not that the current solution exhibits good usability - its just the canonical one)
2008-03-21T22:12:34 <johill> the current solution is odd
2008-03-21T22:13:00 <johill> especially searching for ----stuff rather than linkto:CategoryPage
2008-03-21T22:13:05 <CIA-39> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2592:05b7a9d2167c 1.6/ (MoinMoin/search/queryparser.py docs/CHANGES): Fix category search (use category:CategoryFoo), kill nasty attempt to detect category search (backport from 1.7)
2008-03-21T22:16:57 <johill> I'll fix the searches on moinmaster
2008-03-21T22:17:37 <ThomasWaldmann> wait a se
2008-03-21T22:17:39 <ThomasWaldmann> c
2008-03-21T22:18:07 <johill> ok, what for?
2008-03-21T22:20:43 <ThomasWaldmann> ok, feel free. just wanted to have a test first. :)
2008-03-21T22:20:56 * dreimark will add some german docu soon to MM
2008-03-21T22:24:14 <johill> done
2008-03-21T22:24:23 <johill> this is odd
2008-03-21T22:24:33 <johill> some translators have translated CategoryHomepage and others haven't
2008-03-21T22:26:00 <johill> if it was real metadata it could be translated and still be found, i.e. always stored as 'homepage' adn displayed as 'Domačastran'
2008-03-21T22:41:37 <johill> WAAHH
2008-03-21T22:41:39 * johill cries
2008-03-21T22:41:46 <johill> def getCategories(self, request):
2008-03-21T22:41:46 <johill> """ Get categories this page belongs to.
2008-03-21T22:41:46 <johill> @param request: the request object
2008-03-21T22:41:46 <johill> @rtype: list
2008-03-21T22:41:46 <johill> @return: categories this page belongs to
2008-03-21T22:41:48 <johill> """
2008-03-21T22:41:50 <johill> return wikiutil.filterCategoryPages(request, self.getPageLinks(request))
2008-03-21T22:42:05 <johill> so WTF is with that special ----() regex foo?
2008-03-21T22:42:23 <xorAxAx> is it in the code?
2008-03-21T22:42:30 <johill> both, yes
2008-03-21T22:43:06 <ThomasWaldmann> grown over the years...
2008-03-21T22:43:54 <xorAxAx> historically grown chaos :)
2008-03-21T22:44:10 <johill> so which way do we want?
2008-03-21T22:44:44 * ThomasWaldmann looks at getPageLinks...
2008-03-21T22:45:06 <xorAxAx> johill: the one you just pasted of course
2008-03-21T22:45:20 <johill> xorAxAx: but that one isn't translatable
2008-03-21T22:45:27 <xorAxAx> hmm?
2008-03-21T22:45:34 <johill> or rather, you cannot use multiple language categories in one wiki
2008-03-21T22:45:40 <xorAxAx> why?
2008-03-21T22:45:53 <johill> well you can by making a rather complex page_category_regex
2008-03-21T22:45:55 <xorAxAx> wikiutil.filterCategoryPages will be fine with any lang you configure
2008-03-21T22:45:58 <xorAxAx> yes
2008-03-21T22:46:16 <johill> not very useful in a multi-language wiki
2008-03-21T22:46:22 <johill> like moinmaster
2008-03-21T22:46:51 <xorAxAx> well, thats a different problem
2008-03-21T22:47:03 <xorAxAx> you seem to want to change how categories work in general
2008-03-21T22:47:14 <xorAxAx> and as i said, thats primarily a UI problem
2008-03-21T22:47:15 <johill> something sane and consistent would be a nice start
2008-03-21T22:47:25 <xorAxAx> currently its consistent :)
2008-03-21T22:47:28 <johill> no
2008-03-21T22:47:30 <xorAxAx> besides the search code
2008-03-21T22:47:36 <johill> right
2008-03-21T22:47:47 <johill> which you caused, btw, if you introduced that weird regex
2008-03-21T22:48:05 <johill> indirectly
2008-03-21T22:48:23 <johill> because whoever wrote the search code just pasted things over without thinkin
2008-03-21T22:48:24 <xorAxAx> wellll
2008-03-21T22:48:38 <xorAxAx> i am not responsible for people who dont think :)
2008-03-21T22:48:43 <xorAxAx> the regex makes complete sense
2008-03-21T22:48:55 <johill> and I suppose previously that was the only way since one couldn't search for linkto:
2008-03-21T22:48:59 <xorAxAx> in case of having a search system without link or category search
2008-03-21T22:50:19 <johill> except category search makes no sense
2008-03-21T22:50:21 <ThomasWaldmann> i guess for 1.6.2 we keep the regex approach and do more tests in 1.7
2008-03-21T22:50:25 <johill> since you can just do linkto:category
2008-03-21T22:51:14 <johill> unless, of course, you see it as a migration towards category metadata ;)
2008-03-21T22:52:32 <xorAxAx> johill: as i said, thats not a data issue in the first place
2008-03-21T22:52:56 <xorAxAx> even though that seems to be a sane prerequisite
2008-03-21T22:53:34 <johill> ?
2008-03-21T22:53:39 <johill> what is a prereq?
2008-03-21T22:53:46 <xorAxAx> requirement
2008-03-21T22:54:15 <xorAxAx> well, the possibility to store metadata
2008-03-21T22:54:25 <xorAxAx> in a controlled manner
2008-03-21T22:54:28 <johill> right
2008-03-21T22:56:25 <johill> but I agree, it's not so much about storing the metadata than representing it
2008-03-21T23:01:11 <dreimark> Do we have that textmarker function for a word too?
2008-03-21T23:01:35 <johill> textmarker function?
2008-03-21T23:02:04 <dreimark> e.g. {{#!wiki blue HilfeZurKonfiguration}}}
2008-03-21T23:02:09 <dreimark> {{{
2008-03-21T23:02:25 <johill> ah, no
2008-03-21T23:03:50 <dreimark> hmm for comments it would work
2008-03-21T23:04:10 <dreimark> ok I do use comment syntax
2008-03-21T23:09:52 * dreimark has lots of new ideas
2008-03-21T23:11:21 <johill> :)
2008-03-21T23:13:07 <dennda> What is a "backend" in the storage-context?
2008-03-21T23:13:22 <johill> dennda: the code responsible for actually storing stuff
2008-03-21T23:13:47 <dennda> johill: sure
2008-03-21T23:13:55 <dennda> johill: but could you name a few backends?
2008-03-21T23:14:33 <dennda> (I know what a backend basically is. I wrote one myself once :))
2008-03-21T23:14:49 <johill> right now there's only one functional backend
2008-03-21T23:14:51 <johill> moin16.py
2008-03-21T23:15:08 <johill> which sort of consists of two, one for page and one for user storage
2008-03-21T23:15:24 <dennda> I assume user storage is content the users upload
2008-03-21T23:15:40 <dennda> What is page content?
2008-03-21T23:16:23 <johill> no
2008-03-21T23:16:27 <johill> user storage is user profile storage
2008-03-21T23:16:36 <johill> what settings they save etc.
2008-03-21T23:16:46 <johill> page contents is just the regular moin data
2008-03-21T23:17:04 <johill> and, hopefully soon, also other stuff like images, movies etc
2008-03-21T23:17:22 <dennda> yes, that's what I want to take care of :)
2008-03-21T23:17:27 <dennda> (If you let me ;))
2008-03-21T23:17:39 <johill> right :)
2008-03-21T23:18:14 <dennda> hm
2008-03-21T23:18:26 <johill> for that, MoinMoin/action/AttachFile.py is most relevant
2008-03-21T23:18:28 <dennda> I quite often stumbled over something like this: _ = lambda x: x
2008-03-21T23:18:35 <dennda> at the end of a file
2008-03-21T23:18:36 <johill> dennda: that's just for marking
2008-03-21T23:18:51 <johill> dennda: _("asdf") will mean that the string will later be translated somewhere
2008-03-21T23:18:57 <johill> and should be marked as 'to translate'
2008-03-21T23:19:07 <johill> I think many of the uses of this in the storage code are wrong though
2008-03-21T23:19:27 <xorAxAx> johill: why?
2008-03-21T23:19:40 <dennda> hm, what kind of translation?
2008-03-21T23:19:51 <dennda> (sorry for asking something that sounds quite dumb)
2008-03-21T23:20:19 <johill> actually, they seem mostly ok
2008-03-21T23:20:32 <xorAxAx> dennda: _() is the canonical function name for gettext translations, i.e. those that end up translated at the user's screen
2008-03-21T23:20:33 <johill> except it's not documented that the exceptions will be translated later
2008-03-21T23:20:44 <johill> dennda: just translation to say Chinese
2008-03-21T23:21:25 <johill> if you edit a page with a browser that says it wants Chinese or you set your profile to Chinese, you'll get all the messages in Chinese rather than English
2008-03-21T23:21:28 <johill> hopefully
2008-03-21T23:21:48 <johill> and to know what the translators have to translate, you scan the source code for _() invocations
2008-03-21T23:22:01 <johill> [now, here's actually a nice SoC task, use gettext2]
2008-03-21T23:24:33 <johill> dennda: http://localhost/doc/python/html/lib/node337.html
2008-03-21T23:25:03 <xorAxAx> johill: its full of naked women :-)
2008-03-21T23:25:36 <dennda> :)
2008-03-21T23:26:06 <johill> xorAxAx: cool. my localhost isn't :((
2008-03-21T23:26:11 <dennda> johill: What's the topic of that page? I didn't set it up this way (prefer using devhelp)
2008-03-21T23:26:15 <johill> actually I mean lngettext
2008-03-21T23:26:26 <johill> dennda: 6.28.3.4 Deferred translations
2008-03-21T23:27:36 <dennda> ok, let me take a look
2008-03-21T23:29:13 <johill> "Make bullet-proof round-tripping possible, building on Mikko Virkkilä's work for Google Soc 2006"
2008-03-21T23:29:18 <johill> how can round-tripping possibly work?
2008-03-21T23:30:21 <johill> actually.. it doesn't say lossless :)
2008-03-21T23:30:47 <xorAxAx> johill: by transforming additional information into an orthogonal dimension
2008-03-21T23:31:13 <johill> xorAxAx: would you agree to make category: searches identical to linkto: searches with the added restriction that the category must match the category regex?
2008-03-21T23:31:16 <xorAxAx> like ... comments in the wiki text or docbook markup :)
2008-03-21T23:31:24 <xorAxAx> johill: yes, makes sense
2008-03-21T23:31:25 <johill> yeah that might work
2008-03-21T23:31:38 <johill> I'm just not entirely sure what to do when people say category:Abcde
2008-03-21T23:31:41 <johill> and it's not valid
2008-03-21T23:31:56 <johill> match nothing or match everything?
2008-03-21T23:32:17 <xorAxAx> hmm, match nothing
2008-03-21T23:32:39 <johill> my preference too
2008-03-21T23:37:15 <johill> ahrg
2008-03-21T23:37:16 <johill> wtf
2008-03-21T23:37:28 <johill> why aren't these things passed a request?
2008-03-21T23:39:22 <johill> I don't understand the __unicode__ methods
2008-03-21T23:39:27 <johill> what are they supposed to do?
2008-03-21T23:39:36 <johill> they always seem to return [-]!"%s"
2008-03-21T23:40:22 * dennda likes pythons documentation
2008-03-21T23:42:13 <johill> xorAxAx: how useful is category:regex:something?
2008-03-21T23:42:30 <xorAxAx> hmm
2008-03-21T23:42:35 <xorAxAx> generally useful
2008-03-21T23:42:36 <johill> that's like impossible to implement sanely
2008-03-21T23:42:47 <xorAxAx> why?
2008-03-21T23:43:02 <johill> because you can't check if the regex will match page_category_regex
2008-03-21T23:43:06 <xorAxAx> a page has a list of categories and the regex is matched against any of them
2008-03-21T23:43:08 <johill> eh, you know what I mean
2008-03-21T23:43:15 <xorAxAx> no, i dont :)
2008-03-21T23:43:23 <johill> oh, ok, yeah we can implement it backwards too
2008-03-21T23:45:10 * johill hopes the xapian code is right
2008-03-21T23:45:14 <johill> doubtful, but hey
2008-03-21T23:46:56 <johill> hrm
2008-03-21T23:47:00 <johill> no, I think I can't make that change
2008-03-21T23:47:09 <johill> the queryparser code is too obfuscated for me
2008-03-21T23:50:49 <dennda> Just wondering, the moin16.py module is marked as being translatable, too, but there are hardly any strings a user would ever see. Is it your intention to also translate docstrings?
2008-03-21T23:52:34 <johill> no
2008-03-21T23:52:44 <johill> there are exception strings it it I guess
2008-03-21T23:52:52 <johill> and it's not the module marked translatable any way
2008-03-21T23:52:55 <johill> each single string must be marked
2008-03-21T23:56:20 <dennda> so _ = lambda x: x says "there are translatable strings in this file"? (By the way I couldn't find any other strings in there then docstrings or parameters)
2008-03-21T23:58:03 <johill> no
2008-03-21T23:58:06 <johill> it means nothing on its own
2008-03-21T23:58:14 <johill> _('string') means "string" can be translated
2008-03-21T23:58:21 <johill> eh, will be translated
2008-03-21T23:58:24 <dennda> yes, got that
2008-03-21T23:58:52 <johill> _ = lambda x: x
2008-03-21T23:58:58 <johill> just means that we can't do the translation right now
2008-03-21T23:59:02 <johill> and something else will do them later
2008-03-21T23:59:17 <dennda> ah ok
2008-03-21T23:59:47 <johill> that's the deferred translations stuff I pointed you to MoinMoin: MoinMoinChat/Logs/moin-dev/2008-03-21 (last edited 2008-03-20 23:15:02 by IrcLogImporter)
