IRC log of #schooltool for Tuesday, 2006-04-25

*** AnsiC has left #schooltool00:32
*** didymo has joined #schooltool00:41
*** jinty has joined #schooltool01:04
*** jinty has quit IRC03:30
povbot`/svn/commits: * hoffman committed revision 5967:09:52
povbot`/svn/commits: I don't seem to need this.09:52
*** mooky has quit IRC10:12
*** alga has joined #SchoolTool11:03
*** ignas has joined #schooltool11:37
*** jinty has joined #schooltool11:42
*** alga has quit IRC11:56
*** thisfred has joined #schooltool12:01
*** Aiste has quit IRC12:03
*** faassen has joined #schooltool12:24
*** alga has joined #SchoolTool12:25
*** ignas has quit IRC12:32
faassenI have a failed doctest example in the schooltool tests today12:43
faassenFile "/home/faassen/working/schooltool/schooltool/src/schooltool/setupdata/tests/", line 71, in schooltool.setupdata.tests.test_browser.doctest_SetupDataView_update12:43
faassenbasically it's this:12:44
faassen      [('work',12:44
faassen    -   < object at ...>,12:44
faassen    ?                                                       ^^^12:44
faassen    +   < object at 0xb14c63ac>,12:44
faassen    ?                                                       ^^^^^^^^^^12:44
*** Aiste has joined #schooltool12:52
*** mgedmin has joined #schooltool12:55
*** ignas has joined #schooltool13:09
*** srichter has joined #schooltool14:18
povbot`/svn/commits: * faassen committed revision 5968:14:59
povbot`/svn/commits: store version in a src/schooltool/version.txt text file, so that both the setup script as well as the schooltool UI can make use of it.14:59
*** mgedmin has quit IRC15:10
*** didymo has quit IRC15:16
*** alga has quit IRC15:54
th1asrichter:  Are you back in MA now?16:09
*** mgedmin has joined #schooltool16:18
faassenth1a: I replied to your post.16:20
faassenth1a: on the permissions table which I do not quite understand yet I'm afraid. :)16:20
* faassen waits for the virtual gravel thing.16:29
th1aI just sent a reply.16:31
th1aEssentially, you found a few bugs in the table.16:31
th1ahi ignas16:32
th1ais jinty here?16:32
faassenyes,I think the table effort is valuable, though I'm a bit worried we'll lose rules by trying to encapulate.16:32
faassenth1a: blanking things that make no sense makes sense.16:33
* mgedmin scoops a shovelful of gravel and tosses it towards faassen16:33
faassenth1a: I think we might see two types of permissions, some that apply to person-related data, and some that apply globally, and we may find we end up with two tables.16:33
th1aIs this some Lithuanian idiom?16:33
faassenth1a: anyway, I think we need regular updates.16:33
*** wrobel has quit IRC16:33
faassenno, I think that's just Marius ragging me. :)16:33
* faassen uses his force field to withstand the gravel attack.16:33
* faassen points his easter egg powered cannon in Marius's direction.16:34
th1afaassen:  OK we'll get back to the table in a bit.16:34
ignasth1a: one question16:34
ignasyellow bits signify global setting in schooltool configuration16:34
faassendid I get the word wrong and it is the thing you hammer the table with something else?16:34
ignasor will they have to be customizable per/user16:34
th1aOh, yes.  Gavel.16:35
faassengavel, right.16:35
faassensee, I got it all mixed up.16:35
faassenno wonder I got Marius's wrath.16:35
th1aignas:  Not per user.16:35
ignasth1a: cool :)16:35
ignasi think the yellow bits is what's important16:35
ignasas they will have to become a nice access control-pannel16:35
ignasor at least a set of zcml directives16:35
*** wrobel has joined #schooltool16:36
th1aOK, faassen: what have you been up to?16:36
th1aWhat's the plan for this week?16:36
faassenI've been up to..16:39
faassenthe plan I spend as much time as possible on schooltool.16:39
faassenwhat I've been up to is having a cold and stuff. :)16:39
faassenhowever, I made the eggs work on the branch.16:40
faassenand I hope to be able to merge that into the trunk.16:40
faassenif it doesn't meet with objections.16:40
th1aYes, I guess we have to determine that.16:40
faassenbasically all the branch does is pull in zc.table (and thus zc.resourcelibrary) as eggs.16:40
faassenwhen you do 'make'16:40
th1aThis is the kind of decision I'm not very useful for.16:41
faassenI want to land this branch.16:41
faassenas soon as possible.16:41
faassenas then I can proceed in trying to plug in my demographics stuff.16:41
th1ajinty:  Are you ok with this?16:41
faassenand then continue working on that.16:41
* jinty is comfortable merging as soon as he figures out what kind of tarball the releas scripts will generate16:41
th1aOr do you need to investigate?16:41
faassenjinty: what do you mean about the tarball? I didn't think about generated tarballs at all. :)16:42
th1asrichter:  Do you want to take a look at this before we merge?16:42
jintyand a documented installation process16:42
faassenjinty: I'd like to merge today if possible and I don't know exactly what you want. :)16:42
srichterth1a: I don't understand eggs enough yet to give it a useful review; if it works, it's fine with me16:42
mgedmindoes it add any software requirements?16:43
faassenthe main thing it does now is pull in some eggs.16:43
faassenmgedmin: setuptools.16:43
jintyfaassen: well, er, we normally distribute tarballs to non-developers16:43
faassenit pulls in eggs beyond Zope 3 itself.16:43
mgedminis setuptools packaged for debian/ubuntu?16:43
jintymgedmin: yes16:43
faassenthe way it hooks in their ZCML is still hacked directly into site.zcml16:43
srichtermy main question is: Are we sure that all bugs concerning eggs and Zope 3 are worked out?16:44
faassenwhat bugs?16:44
srichterwell, whatever there might be16:44
th1aWe aren't using eggs for Zope 3 itself.16:44
faassenI'm just using eggs for Zope 3 extensions.16:44
srichterZope 3 has not committed to eggs in the upcoming release...16:44
faassenit is much less impact on the makefile and so on than actually extending the makefile.16:44
faassenthat's fine, I just need eggs for zc.table and zc.resourcelibrary, and those exist as I made them.16:44
faassenand any other package I want to pull in, I'll just make an egg. that's easy.16:45
srichterI would let Zope 3 make the step to eggification first and then use it in ST16:45
srichterwill those pulled in packages be writable checkouts?16:45
srichterthat is not good16:45
srichterI often fix bugs during development16:45
srichterthe current Makefile allows me to checkout writable versions if I wish16:46
faassenlook, I just want to use a few measly libraries to develop demographics stuff.16:46
faassenI don't want to eggify zope 3.16:46
faassenif you want to have writable checkout you can check them out yourself and install them in development mode.16:46
* th1a thinks srichter is overreacting a bit.16:46
jintysrichter: it is easy to checkout a writable checkout of an eggified package and have the schooltool build use that16:47
faassenI started out changing the makefile but it soon turned into a huge ball of make cruft.16:47
jintyit just needs documentation16:47
faassenunmaintainable makefile cruft.16:47
srichterI am just making an argument; at the end of the day I am far too tired right now to heavily argue for or against anything ;-)16:47
* mgedmin is inclined to try the egg solution16:47
th1aSo primarily this is jinty's domain.  If it is ok with him it is ok with me.16:48
faassenI shall add in a document on how to do the development mode checkout egg thing.16:48
th1amgedmin:  Do you want to personally inspect this before it is merged?16:48
*** alga has joined #SchoolTool16:48
mgedminwhat's the url?16:48
faassenit's the schooltool-eggs branch.16:48
algahello, sorry I'm late16:48
jintyI will make a test with the release scripts and come back to faassen with a response16:48
* mgedmin expected it to be schooltool-omlette16:49
jintythis afternoon, i think16:49
faassenok, that would be good.16:49
faassenit's one obstacle that makes it hard to start integrating my other stuff.16:49
th1aYes, it would be good to actually see some demographics code ;-)16:49
faassenso my plan is that after this lands..16:50
faassenI can start moving in some demographics stuff.16:50
faassenI'm still scared of the schooltool skin and viewlets and so on.16:50
faassenbut I'll just make a stupid UI first.16:50
faassenas I'm not the UI guy anyway. hopefully we'll get some UI work done here but I don't think we can count on it anytime soon.16:50
th1aA stupid UI will do for starters.16:51
th1aOK... POV?  ignas?16:52
th1aThanks faassen.16:52
ignaswe will have ~25 dd for may16:52
ignaswhat will be your orders ? :)16:53
faassenthe security work would be nice. :)16:54
th1aI guess that takes us to the next order of business.16:54
th1aYes, this might be the time for that, if we can figure out what we want.16:54
faassenI think the steps you and I have been taking were good.16:54
th1aWhat would we need to decide to do "the security work?"16:54
faassenthe longer I stared at the table the more problems I found but also the more sense it made. :)16:55
ignasas in - get rid of ACL views, replace everything with a centralized access control pannel16:55
ignasmake the panel usable16:55
faassenand a centralized policy for security, somehow.16:55
faassenthis needs some implementation thought.16:55
th1afaassen:  I think it was particularly buggy because I kind of changed the design of the table as I went.16:55
faassenth1a: I hope we can get a bugfix release soon. :) I understand how that goes anyway.16:56
faassenanyway, we need to figure out what notions there are in the access control picture.16:56
faassenthere is a notion of 'my information'16:56
faassenthere's a notion of 'information I have access to through a kind of relationship'16:56
faassenand a notion of 'stuff I can do because I have a certain global group membership/role'16:57
faassenthe latter is simplest; those are permissions someone has globally.16:57
faassenanyway, the question I can't answer myself right now is whether the systm contains enough meta information to get the former two.16:58
faassenlike, there doesn't seem to be a notion of 'self' data yet, as far as I understand.16:58
faassenthere is a notion of relationship. This may be powerful enough, though the question remains whether we can make it all perform adequately, as far I understand.16:59
th1aThat pretty much just comes from context in the database.16:59
ignascould we please go back to the: "now that i own the PoV what am i supposed to do with this thing"16:59
faassenok, sorry.16:59
th1aignas:  What?16:59
ignasyou have got 25 dev days of PoV17:00
ignasand i still don't have a clear picture what do you want to do with that timew17:00
th1aWell, fixing access control would be a good use of that time.17:00
th1aThat's what launched us down that path.17:00
ignasthere are things like ACL, LDAP, SchoolBell extraction ...17:01
faassenhow important is schoolbell in all of this?17:01
th1aLDAP I think we should pass on.17:01
faassenschoolbell isn't on the critical path right?17:01
th1aIt is important that at some point in the next six months we have some kind of SchoolBell release.17:01
faassenah, ok.17:01
th1aWhich shouldn't be that hard.17:02
jintyas release manager, jinty, would like to see SchoolBell building sooner rather than later as he considers it a release blocker17:02
th1aUnless we miscalculated.17:02
th1ajinty:  That's what I'm thinking.17:02
faassenjinty talks about himself in the third person. :)17:02
jintyjust like Saddam Hussein;)17:02
faassenoh, he does?17:02
th1aSo the eggs related changes should be settled quickly enough.17:03
faassenjinty: right, please by the way if you want me to do more eggs stuff.17:03
th1aSo that's not a SchoolBell-packaging blocker (barring unforseen disaster).17:03
faassenjinty: that you need as a release manager, etc, let me know.17:03
faassenjinty: I already fixed the version number issue.17:03
faassenjinty: (though didn't do it for Schoolbell yet, that has its own version)17:03
jintyI think we are bound to release schoolbella and schooltool together17:04
faassenjinty: (don't know how schoolbell fits into all this)17:04
faassenwith the same version number?17:04
jintybecause they use the same base libraries17:04
faassenok, then it's a 1 line fix I think.17:04
th1aHm... we should probably sync the ST/SB version numbers.  Make them both 2006 x.x17:04
mgedminSchoolTool Vista17:05
mgedminthen we could postpone the release date a couple of years17:05
th1aOK so, for POV:17:06
th1a* I'll pick out some bugs to fix;17:06
th1a* extracting SchoolBell17:06
th1aIs there a reasonable first step in fixing permissions?17:06
th1aWithout trying to solve the problem all at once?17:07
ignasth1a: hiding the current ACL view behind an Access Control Pannel ?17:07
th1aCreating a "permissions control panel" for site-wide permissions configuration?17:07
faasseneven though it wouldn't do anything yet?17:08
th1aIt would do something.17:08
ignasfaassen: even though it would just set some "checkboxes"17:08
th1aThe limitation would be that it would only change group-related permissions.17:08
faassenright, if it only changed global group related permissions..17:09
th1aWe'll need more work under the hood to do relationship-based permissions.17:09
faassenand self-permissions.17:09
th1aYes, I think.17:09
faassenglobal group related permissions, i.e. parts of the permission role mapping. :)17:09
faassenwould work.17:09
th1aOK.  We'll shoot for that.17:10
th1aDoes that make sense to everyone?17:11
faassenfor the acl, let's continue the discussion.17:12
faassenrefining that table.17:12
faassenand thinking about implementation choices.17:12
th1aI think the "staff member" column should go.17:13
ignasth1a: i think i'll need more time to think of a viable first step, as doing the UI apparently has it's downsides17:14
th1aSuch as?17:15
mgedminhaving to make the UI work with a backend that is not suitable for it?17:15
* mgedmin hasn't thought very deeply about security so his comments may be entirely mistaken17:16
th1aIf we're just doing group-based access control, I think our back end is sufficient.17:17
algabut then we'll have to do this stuff over again when the backend changes17:17
th1aI guess it depends on how much the backend has to change.17:18
faassenwell, there are three categories of ..17:18
th1aI'm not against doing the backend first, though.17:18
ignasth1a: well - with the notion of local permissions gone - probably a lot17:18
faassensome things you are allowed to do globally, or not.17:18
faassenlike, add new uers.17:18
faassenif you're a school administrator, you can just add new users.17:18
faassenI think th1a was suggesting we create a UI for the configurable parts of that first.17:18
faassen(not all of those settings are configurable; some things a school admin will always be allowed to do, but in some cases there exist choices in what a student may see, for instance)17:19
faassensome other permissions depend on your relationship with the 'owner' of the object.17:19
faassenlike, I may see a student's demographics if I'm his teacher.17:19
faassenth1a wasn't proposing we do a UI for that yet.17:19
faassenI'm not sure how much sense a UI would make right now though.17:20
faassenas after all it's fairly simple check boxes.17:20
faassenall of those settings are global and they need good descriptions on what they do if you check or uncheck them.17:20
faassen(or whatever the UI will look like)17:20
th1afaassen:  Yes.17:20
th1aIt could be quite simple.17:20
faassenso perhaps it makes more sense to just work on the security story as a design first.17:20
faassenas the UI work wouldn't help (or harm) us that much in the scope of things.17:21
faassenand the shape of the configuration UI *might* be affected by further thinking along the road.17:21
th1aWell, one question is whether or not any changes actually have to be made to implement group based permission.17:21
faassenI think group based permission is the one thing that more or less works now.17:22
faassenexcept perhaps the permission distribution needs to be more fine grained.17:22
faassenor altered in some way.17:22
faasseni.e. modifying lots of ZCML.17:22
faassenanyway, whatever we do, we need to continue the actual design.17:22
th1aI should note that we can get through SchoolTool 2006 with just group and self based permission if we need to.17:23
th1aself-based meaning the system knows who "owns" their own data.17:23
faassenwell, dunno.17:23
th1aWhich we can do now reasonably well.17:23
faassenwhat about a gradebook?17:23
faassenis that owned by the teachers of the section?17:24
faassenthen that looks like it could work.17:24
jintymgedmin: apparently the egg setup in schooltool-eggs needs a later version of setuptools than that currently in Debian17:25
mgedminjinty: is 0.6a9-1ubuntu1 recent enough?17:25
jintymgedmin: nope needs 0.6a11 AFAIK17:26
ignasth1a: so it's ACL design + SchoolBell extraction and some bugs?17:26
th1aOK.  I'm thinking that we won't actually have relationship-based permissions in ST 2006.17:26
th1aignas:  That sounds reasonable.17:26
mgedminthis is why I'm not too keen on trying eggs myself -- everything needs bleeding edge, released just yesterday, version of setuptools17:26
ignasfaassen: when will your demographics changes be ready, as they might conflict with the extraction in some places ?17:27
faassenignas: what extraction?17:27
ignasfaassen: not demographics, but skin if to be exact17:27
ignasfaassen: schoolbell from schooltool17:27
faassenignas: oh, the skin changes are currently on hold, I mean, a lot more work could be done but I'm not going to do it soon.17:28
faassenignas: what's in is in there.17:28
ignasfaassen: you will be the one merging then :D17:28
faassenignas: sure.17:28
faassenmgedmin: that dependency thing sucks. :(17:28
faassenmgedmin: then again it's reasonably easy to bootstraps the whole thing.17:29
th1aIf we don't have the version of setuptools we need in Dapper... is that a showstopper?17:29
mgedminmaybe not, but it is a pain in the donkey17:30
faassenmgedmin: dapper doesn't have it either?17:30
mgedmindapper has 0.6a917:30
* mgedmin runs dapper17:30
th1aOK.  I'll keep working on the permissions table.17:31
th1aHave a good week, folks.17:31
* th1a bangs the virtual bag of gravel.17:31
faassengravel. :)17:32
ignaswith a shovel :)17:32
faassenjinty: could you let me/the mailing list know what's needed before you deem the egg branch landable?17:32
faassenmgedmin: and the same to you?17:32
mgedminfaassen: I'm not opposed to landing that branch17:33
mgedminlet's just see what happens17:33
jintyfaassen: sure, am busy with it now,17:33
faassenok. :)17:33
faassenthe main advantage is that it doesn't add a whole lot of new infrastructure to the makefile.17:34
faassenjust a few minor tweaks.17:34
*** srichter has quit IRC17:50
*** wrobel has quit IRC18:08
*** wrobel has joined #schooltool18:10
*** alga has quit IRC18:19
*** faassen has quit IRC18:59
ignasth1a: ping19:11
th1aignas: pong.19:23
ignasmaybe you do imagine the requirements for booker/ booked resources in calendars19:23
ignasas at the moment i am planning to simplify that bit to make it manageable19:24
ignasthe problem is:19:24
ignaswe had some weird logic that was sorting out information seen on events19:24
ignasdepending on what calendar and who was looking at19:24
ignasif you are looking at the calendar of ResourceA you will see only "Booked by: john" not both "Booked by: john" + "Booked resources: ResourceA"19:25
ignasif you are looking at the calendar of john you will not see "Booked by john"19:25
th1aIs that more of a pain than it is worth?19:26
ignasand everything mangled up for overlaying19:26
ignaswell with sections and timetable events19:26
ignasit is geting out of hand19:26
ignasas i can't reliably decide what should be seen where19:26
ignasand usecases are pretty difficult to express19:26
ignasin some manageable way19:26
th1aOn the other hand, all we're trying to avoid is a little redundancy, right?19:27
ignasat the moment i am leaning to: show all the booking information19:27
ignasth1a: yes19:27
th1aOK.  Screw it.19:27
th1aIt won't kill you to see "Booked resources: Resource A" on Resource A's calendar.19:28
Aisteth1a: can I quote you on that? :DDD19:28
Aistethe 'Screw it' bit?19:28
th1aWell, whether or not ignas can screw Resource A depends on what Resource A is.19:28
th1aPresumably A does not stand for Aiste.19:29
ignashu-mans are not resources :)19:30
th1aGood point.19:30
Aistethey are19:30
Aistehaven't you heard of HR?19:30
th1aNot in SchoolTool.19:30
Aistewell, I'm not in school19:30
Aistehaven't been for a long time19:30
th1aYou can't book a person.19:30
th1aIn SchoolTool.19:30
th1aAnyhow, if that's the only cost, I think it is a reasonable move.19:31
ignasif users will want that data hidden they will know what they don't what to see better than we do anyway :)19:33
povbot`/svn/commits: * jinty committed revision 5969:19:56
povbot`/svn/commits: Make get_version a bit more OS independent and port it to i18nextract.py19:56
povbot`/svn/commits: * jinty committed revision 5970:20:12
povbot`/svn/commits: Change the schooltool release script configuration to test with the eggs branch. Also change the version number test to the new version.txt file.20:12
*** thisfred has left #schooltool20:15
*** Aiste has quit IRC20:24
ignasth1a: what is the semantic meaning of group timetables ?20:25
ignasbecause apparently one can add timetables to groups through restive interface20:26
ignaswhile not through the normal web interface20:27
povbot`/svn/commits: * jinty committed revision 5971:20:43
povbot`/svn/commits: Include the version.txt file in the zpkg metadata.20:43
mgedminignas: aren't group timetables automatically composited into the timetables of their members?20:45
ignasyes they are20:50
ignasbut you can't create them through normal user accessible means20:50
ignasand you can't book resources for groups20:50
ignas2 choices20:50
ignas1. Dump group timetables in favour of Recurring Events20:51
ignas2. Add resource booking and scheduling to Groups20:51
ignasbecause now i will have to add code that works around the fact that groups can't book resources while they can theoretically have timetables20:52
*** Aiste has joined #schooltool21:14
*** jinty has quit IRC21:16
*** mgedmin has quit IRC21:20
*** ikonia has left #schooltool21:33
th1aignas:  Still there?21:46
th1aI don't think there should be group timetables.21:47
th1aI don't know what that would be either.21:47
ignasi see21:49
th1aThere is, quite frankly, probably still some conceptual weirdness around timetables we've never completely straightened out.21:50
th1aLike this case.21:50
ignaswell - clients should help us see what is really required and which excess flexibility should be slowly removed21:52
th1aYes, some fresh eyes should be able to sort it out at some point.21:53
*** ignas has quit IRC22:24
*** Joe_Violanti has joined #SchoolTool23:25

Generated by 2.15.1 by Marius Gedminas - find it at!