IRC log of #schooltool for Thursday, 2005-08-25

povbot/svn/commits: * srichter committed revision 4859:00:20
povbot/svn/commits: 1. Added a couple more files to ignore on the root.00:20
povbot/svn/commits: 2. Import the new testbrowser using the same mechnism Zope 3 itself is00:20
povbot/svn/commits: downloaded. This is great, since I now do not have to constantly adjust our version.00:20
povbot/svn/commits: Fixed tests to match the new version.00:20
*** FarcePest has quit IRC01:45
*** srichter is now known as hter04:19
*** hter is now known as srichter04:20
povbot/svn/commits: * srichter committed revision 4860:04:56
povbot/svn/commits: Merge final trunk revisions into the branch:04:56
povbot/svn/commits: Merged: 4703, 4706, 4708, 4709 (partial), 4710, 4712 (partial), 4713, 4716,04:56
povbot/svn/commits: 4718, 4719, 4720, 4723, 4729, 4732, 4735, 4741, 4745, 4748, 4751, 4753, 4758, 4761, 4762, 4768, 4769, 4770, 4771, 4783, 4787, 4788, 4833, 483504:56
povbot/svn/commits: Ignored: 4722, 4731, 4746 (already fixed), 4760, 4775 (already fixed),04:56
povbot/svn/commits: 4898 (N/A), 4799 (N/A), 4837 (see 4835)04:56
*** tvon has quit IRC08:51
*** tvon has joined #schooltool09:16
*** SteveA has joined #schooltool09:21
*** azazel has joined #schooltool12:44
*** Ricey has quit IRC12:44
azazelhi all!12:45
azazelcan someone explain me the diff between schooltool and schoolbell?12:45
*** Ricey has joined #schooltool13:10
*** Aiste has joined #schooltool13:50
*** th1a has joined #schooltool14:10
*** mgedmin has joined #schooltool14:14
srichtermgedmin: I finished merging the trunk changes to the branch14:16
srichterI still have some issues with old location references though14:17
srichtereven though I convert all instances to the new class paths, for some reason the references to those objects do not get updated14:18
srichterit is very hard to track down all references systematically (due to circular refs)14:18
srichterI'll contact Tim and see what he has to say14:19
srichterone way would be to hack the ObejctReader to mark classes that have an old path automatically with _p_changed = True14:19
mgedminhm... a generation script that marks all objects as changed?14:20
srichteranother would be of course to leave the references around until we have a ZODB solution and remove the deprecation warnings14:20
srichterthat's all that is needed14:20
mgedminI once did some hacking so that I could traverse the graph of object references and set all Persistent objects reachable from a given source14:21
srichteruuuh, I would love to see that script14:21
mgedmin(I was debugging some Copy & Paste issues that left some shared subobjects between the original and the copy)14:21
mgedminit was a simple hack around locationCopy, IIRC14:21
srichterI also looked at fsrecover, but it seems too much work14:21
*** ignas has joined #schooltool14:24
mgedminannotatedLocationCopy(foo) returns a list of (object, id) for all objects (not necessarily Persistent) that are reachable from foo14:24
mgedminid shows whether that object would be cloned or aliased if you did copy & paste -- irrelevant for this use case14:24
srichterI gotta go14:25
srichterI will be back in about 3 hours14:25
srichterth1a: before I leave: I have not read the entire UI doc yet, but we also have nested menus now in Zope 3, so we might be able to utilize them15:07
th1aQuick question:  Does DTML actual work in Zope 3.1?15:08
srichterit should15:08
mgedmin(even if it does, we ought to pretend it doesn't)15:08
srichterhowever, not for views15:09
srichterso if someone asks whether they can use DTML to extend ST, we say no :-)15:09
th1aWell, we could use DTML to make non-XML templates (i.e., CSV).15:11
th1aBut we don't need to get into it now.15:11
*** bskahan has joined #schooltool15:11
th1abskahan:  Did you see the email I sent?15:12
srichterth1a: the UI doc is a good start15:13
th1aI'm not sure how it translates into Zope 3 jargon: views, skins, layers, adapters, etc.15:13
* srichter would really like to start working with UML for this stuff15:13
mgedminI'm not sure DTML would be a considerably better fit than ZPT for CSV files...15:16
mgedmin"<tal:x replace="structure var1"/>", "<tal:x replace="structure var2"/>"15:16
th1aI was under the impression that you can't use ZPT for non-XML, but perhaps that is just something I imagined.15:17
mgedmin"&dtml.howdoidsablexmlquoting-var1;", "&dtml.howdoidisablexmlquoting-var2;"15:17
mgedminyou can use ZPT for anything (ok, not binary stuff), although that might not be the best option15:18
mgedminyou can disable XML quoting by using 'structure' where necessary, and you can get rid of all tags15:18
mgedminreplacing XML quoting with something else might be more difficult15:18
mgedmin<tal:x replace="structure python:csv_quote(var1)"/>15:19
mgedminor maybe <tal:x replace="structure var1/quote:csv"/>15:19
mgedmina custom templating mini-language for CSV migh work best15:19
th1aI was thinking that as well.  Regardless, I don't think we're going to deal with that question in the short term.15:20
bskahanth1a: just looking at it now15:25
bskahanI'm wondering what hardware paul saitta is using that has performance problems15:26
th1aDo we have some chunky .ics files for testing?15:26
bskahancan probably find them via google15:27
th1aWe need to start generating serious sample data.15:28
th1aThat's at the top of the list.15:28
bskahanth1a: principal_* views would be abstract views not tied directly to a single content object (like a 'dashboard' view), vs object_* views would be various views on a single object (like a person) that show different information to different classes of principals?15:38
azazeli'm trying to understand if schooltool can be a good base for developing a people and resource management app for a customerwith a core businness other than courses15:38
azazelany hint?15:39
bskahanazazel: probably schoolbell would.  its a good calendar application that you could add CRM features to15:39
th1aazazel:  Do you have much Python & Zope experience?15:40
azazelth1a: yes, expecially Zope 2.x and plone experience15:40
th1aOK. that's a good start.15:41
azazelmy customer has a vast (and good) database of resources in an existing postgres db, any idea if i can reuse it directly with an orm (like sqlobject?)15:43
th1aI don't really know.  I think there are adapters for postgres for Zope 3, but that's a bit out of my range.15:44
bskahanazazel: there isn't anything built into schoolbell to handle SQL backends, but the idea has been discussed and Marius seemed to think it would be possible for someone to develop15:44
th1abskahan:  object_* views don't change based on who is viewing them, except by permissions.15:45
th1ai.e., some people can't see them, or all of them.15:45
bskahanthe object_* views are essentially what we have now, correct?15:45
azazelbskahan: thanks,15:46
bskahanazazel: no problem15:46
th1aWell, we have the equivalent of both, really.15:46
th1aIf you think of a person's calendar view as a principal view.15:46
th1aBut in this case, all the principals get the same view.15:46
azazeljust another question is it possible to set the view permission an a cal to a person but not the write permission?15:46
th1aAlthough there is a bug in the new release that is causing the permissions to go wonky.15:47
th1aI'd like to get a bugfix release out today...15:47
th1aI don't know where gintas is.15:48
azazelth1a: from the standard schooltool interface or from the zmi?15:48
th1aI'm talking about the SchoolTool web interface.15:49
azazeli've found it, thanks... setup access15:49
th1abskahan:  Are you busy today?15:53
bskahanth1a: I'm sorry, I'm not quite following what the principal vs. object view is.  Why is the calendar a principal view15:53
bskahanfor the middle of the day15:53
bskahanI'll be available more in the evening, but that's the middle of the night for you15:53
th1aWell, YOUR calendar is a principal view, in that you can modify it to overlay other calendars & create events.15:54
th1aI guess other object calendars would become read only in this model.15:54
bskahanwhat do you need?15:54
th1aI have been reminded that gintas is in Denmark.15:54
th1aI *really* want to put an updated tarball on the site today.15:55
th1aBecause this permissions bug is a real problem.15:55
th1aThere are just two zcml files to switch.15:55
th1aAnd everyone here is a bit stressed and busy.15:55
bskahanI''m pretty swamped until 6, I can do it then15:56
th1aI should be able to do it before then.15:56
bskahansorry, if I had done a release previously I'd say I'd do it this morning, but I'm not entirely sure of everything that's involved15:58
bskahanand I don't want to rush it15:58
th1aAs manager of the project, I have the most leeway to do it sloppily.15:59
th1aIf you know what I mean.15:59
*** ignas has quit IRC16:01
bskahannot "do it sloppily", you can "prioritize time constraints" ;)16:01
*** ignas has joined #schooltool16:26
mgedminth1a, I can do an updated tarball, if you give me the original tarball and the two new zcml files16:34
th1aI'm doing it right now...16:35
th1aThere are two files.16:47
th1aThe first one I sent twice.16:47
th1aLet me message you directly.16:47
mgedminI'll look into it after lunch16:49
srichtermgedmin: I don't understand how your code takes care of all the references17:44
srichterit seems very specific to your use case17:44
srichterupdate: Jim thinks that a straightforward change in ZODB can fix the problem18:05
*** eldafar has joined #schooltool18:08
eldafaranyone alive?18:09
*** eldafar has quit IRC18:10
*** eldafar has joined #schooltool18:11
*** eldafar has quit IRC18:11
*** eldafar has joined #schooltool18:12
th1aBack from lunch.18:44
eldafardo you know off the top of your head how to set cookies in zope3?18:49
th1aThe amount of specific technical details about Zope3 I keep on top of my head is still quite small.18:50
mgedminsrichter, I didn't know that references to a persistent object stored the class name18:57
mgedminI assumed references only stored the OID18:57
srichterdid you follow the #zope3-dev discussion18:58
srichterI have a solution to the problem now, since we have a file storage18:59
mgedminI just read it18:59
srichterI am simply going through all oids and _p_changed = True all objects18:59
srichterthat should take care of it18:59
srichtermgedmin: I would prefer if the translations are not retrieved for every "make run"19:04
srichterit makes startup much slower than it has to be19:04
mgedmin"make run" is a silly shortcut for developers19:07
mgedminrun "make" once, then run ./s* directly19:07
*** th1a has quit IRC19:36
*** gintas has joined #schooltool19:37
mgedminhi, gintas!19:38
*** alga has joined #SchoolTool19:38
gintashi mgedmin19:44
mgedminhow's internet down there?19:44
gintaspretty good19:45
gintasubuntu updates flew in at 700K/s19:45
gintasbut that's from a Danish mirror of course19:45
mgedminhow hard is to make a schooltool release?19:45
mgedmintom wants a bugfix release that contains those two zcml permission changes19:45
mgedminI looked at it, and I couldn't find a release process document or anything19:46
gintasit's no big deal, hopefully19:46
gintashas the fix been backported?19:46
mgedmineven the debian/ subdirs are gone19:46
mgedminAFAIU the fix is in both ST and SB release branches19:46
gintasthe debian subdirectories are in a different svn directory19:46
gintasbecause SB/ST is not a native Debian project19:46
ignasgintas, yes the fix was backported19:46
mgedminit'd be nice to tag the current release branches and release them as SB 1.2.1 and ST 0.11.119:46
mgedmintom just left19:47
gintasI can probably do that19:47
gintasactually packages are built automatically, except Windows ones19:47
mgedminI hesitate to ask you, but if you could do it, tom would be very happy19:47
mgedminI do not think he wants a new Windows package19:48
gintasI think I should be able to upload everything today19:48
gintasdo you know anything about i18n and rosetta?19:49
mgedminthere's a makefile rule that downloads and merges rosetta translations19:49
gintasI did not do any special things for the previous release19:49
mgedminjinty wrote that19:49
gintasperhaps I should have run that rule19:49
gintasthat would be nice to figure out before the release, so that we could add some possibly new translations in addition to the bugfix19:50
gintasis jinty around?19:50
mgedminsadly no19:51
mgedminhe's away this week19:51
mgedminin ivija we have a script that does a release19:52
gintasI have a similar script, it's in the Debian packaging directory19:52
mgedminis it completely automated?19:53
mgedmindoes it do everything?19:53
gintasmost things, yes19:54
gintasit exports SB & ST, and builds the tarballs and Debian packages19:54
mgedminbut it doesn't create tags in svn or change version numbers in __init__ files?19:54
gintasthat's fairly easy to do; and I wanted the script not to affect the repository19:55
gintasOTOH that should be easy to code19:55
mgedminthe only problem is that when you and jinty are both away, the rest of us do not know how to make a release19:56
mgedminand be sure not to miss any steps19:56
gintashmm, I thought we had some document somewhere19:56
gintasI'll see to this19:57
*** azazel_ has joined #schooltool19:57
azazel_hi all19:57
azazel_groups in schoolbell are used to group :-) both people and resources19:59
gintasthe latest SchoolBell version does not support adding resources to groups, AFAIK20:00
gintasouch, the 'get-rosetta-translations' makefile rule does not work any more, it shows a heap of 404 errors20:06
gintasthe Rosetta website is confusing for me20:10
gintasin fact, I can't find any SchoolBell translations on Rosetta at all20:17
*** lunarosity has joined #schooltool20:18
*** hwo4 has quit IRC20:36
srichtermgedmin: based on the outcome at #zope3-dev, what do we want to do?21:13
mgedminI would like to understand what the problem is21:14
srichterso I did P1.N1.P2 (P -> persistent; N -> non-persistent)21:15
srichterif I move the references, then P2 gets not updated when doing P1._p_changed = True21:16
srichterbut the iterator should get P2 too, huh?21:16
mgedmindid you add a debug print and notice that both P1 and P2 are returned by the iterator?21:16
srichterI am trying to do that next21:17
srichterI have not updated using the iterator yet21:17
*** eldafar has quit IRC21:19
*** azazel_ has quit IRC21:39
srichtermgedmin: note that you used p_changed instead of _p_changed21:41
srichterand then it fails21:41
mgedminit works if I do obj._p_activate(); obj._p_changed = True21:43
mgedminbut not if I try to do iteration over records21:44
mgedminworks now21:45
mgedminmy iteration loop is this:21:45
mgedminrecord = ('\x00\x00\x00\x00\x00\x00\x00\x00',)21:45
mgedminwhile True:21:45
mgedmin    obj = conn.get(record[0])21:45
mgedmin    obj._p_activate()21:45
mgedmin    obj._p_changed = True21:45
mgedmin    if record[-1] is None:21:45
mgedmin        break21:45
mgedmin    record = conn._storage.record_iternext(record[-1])21:45
mgedminI'd be surprised if it were correct, though21:46
srichterthat should work21:46
srichterjust print the record21:46
mgedminif I comment out the _p_activate() line, the code breaks21:46
srichterif you get three entries you are good to go21:46
mgedminI get21:47
mgedmin$ ./run21:47
mgedmin<class 'new.A'>21:47
mgedmin<class 'new.B'>21:47
mgedmin<class 'new.C'>21:47
mgedminwhen I use the loop I pasted here in zodb-test2/migrate.py21:47
mgedmininstead of mucking with conn.root()21:47
srichterthe first entry should be conn.root()21:47
mgedminwanna another tarball?21:47
* mgedmin nods21:48
srichterjust paste all of test2/migrate21:48
srichterit works!21:49
srichterI just tried _p_activate() in the evolution script21:49
srichtermgedmin: ok, everything is great now :-)22:20
*** Aiste has quit IRC22:20
*** alga has quit IRC22:21
*** mgedmin has quit IRC22:22
*** ignas has quit IRC22:28
povbot/svn/commits: * srichter committed revision 4861:22:34
povbot/svn/commits: Finally got the evolution script fixed. Both, Marius and I, learned a lot about the ZODB. But all is fixed now and look how simple the code is!!!22:34
povbot/svn/commits: * gintas committed revision 4862:22:42
povbot/svn/commits: Tagging SchoolBell 1.2.122:42
povbot/svn/commits: * gintas committed revision 4863:22:42
povbot/svn/commits: Tagging SchoolTool 0.11.122:42
povbot/svn/commits: * gintas committed revision 4864:22:51
povbot/svn/commits: Set SchoolBell version number in release22:51
povbot/svn/commits: * gintas committed revision 4865:22:52
povbot/svn/commits: Set SchoolTool version number in release22:52
povbot/svn/commits: * gintas committed revision 4866:22:55
povbot/svn/commits: Set SchoolBell version number in release22:55
povbot/svn/commits: * gintas committed revision 4867:22:55
povbot/svn/commits: Set SchoolTool version number in release22:55
povbot/svn/commits: * gintas committed revision 4868:22:57
povbot/svn/commits: Bound SchoolTool 0.11.1 to SchoolBell 1.2.122:57
*** bskahan has quit IRC23:25
*** admp has joined #schooltool23:26
povbot/svn/commits: * gintas committed revision 4869:23:58
povbot/svn/commits: Revision 4841 directly backported revision 4616 from trunk, which brought in a dependency of WFMC, which we do not have in the tarball.23:58

Generated by 2.15.1 by Marius Gedminas - find it at!