| *** bskahan has quit IRC | 00:06 | |
| *** FarcePest has quit IRC | 02:40 | |
| povbot | /svn/commits: * srichter committed revision 4617: | 04:02 |
|---|---|---|
| povbot | /svn/commits: I converted the implementation of ITimetabled from a mixin to an adapter. There some simple issues left to address, but it shows how things should be done. | 04:02 |
| povbot | /svn/commits: Also, note how many delcarations, classes and other support code I was able to remove immediately. | 04:02 |
| * tvon wonders if there is a quick way to make a canAccess call fail for testing | 04:02 | |
| srichter | create a security proxy that has an empty checker | 04:03 |
| srichter | then everything will be always denied | 04:03 |
| tvon | ah | 04:04 |
| tvon | erm, how might I go about creating a security proxy? | 04:05 |
| srichter | check the dopcs in zope.security | 04:05 |
| * tvon nods | 04:05 | |
| tvon | thanks | 04:05 |
| srichter | soemthing like: SecurityProxy(obj, checker=NamesChecker({}, {})) | 04:05 |
| povbot | /svn/commits: * srichter committed revision 4618: | 05:01 |
| povbot | /svn/commits: Made title a real property of ISchoolBellApplication. | 05:01 |
| povbot | /svn/commits: Simplified code a little bit. | 05:01 |
| povbot | /svn/commits: * srichter committed revision 4619: | 05:07 |
| povbot | /svn/commits: Simplify the SchoolToolApplication code a bit. | 05:07 |
| povbot | /svn/commits: * srichter committed revision 4620: | 05:48 |
| povbot | /svn/commits: First stab at a subscriber-based model of creating the schooltool application object. | 05:48 |
| *** bskahan has joined #schooltool | 05:51 | |
| bskahan | r4617 is a beautiful thing | 05:52 |
| bskahan | I've been bitten in the ass by schoolbell.app.app.Person vs. schooltool.app.Person several times | 05:53 |
| * tvon has high hopes for that branch | 05:55 | |
| *** jinty has quit IRC | 06:14 | |
| *** bskahan has quit IRC | 06:48 | |
| *** tvon has quit IRC | 07:37 | |
| *** tvon has joined #schooltool | 08:25 | |
| tvon | Where did the topic go? | 08:32 |
| *** tvon changes topic to "SchoolTool development discussion - http://www.schooltool.org/" | 08:43 | |
| *** tvon changes topic to "SchoolTool development | IRC logs at http://source.schooltool.org/irclogs/ | Buildbot: http://source.schooltool.org/buildbot/ | Dev meetings Mon, 13:30 UTC" | 08:45 | |
| * tvon sighs | 09:30 | |
| povbot | /svn/commits: * tvon committed revision 4621: | 09:30 |
| povbot | /svn/commits: view/update was missing from index.pt, please backport | 09:30 |
| *** erchache has joined #schooltool | 11:44 | |
| erchache | hi | 11:44 |
| *** erchache has quit IRC | 12:17 | |
| *** jinty has joined #schooltool | 13:14 | |
| *** mgedmin has joined #schooltool | 13:32 | |
| *** thisfred has joined #schooltool | 13:47 | |
| *** bskahan has joined #schooltool | 14:18 | |
| srichter | tvon: they switched servers the other day, so at some point there might have been noone here and the channel got deleted | 14:24 |
| srichter | mgedmin: good morning | 14:30 |
| srichter | mgedmin: so I played with the code yesterday and I am confident now that the merge is possible and a lot of duplicate code can be removed (also for views; w/o pagelets) | 14:31 |
| srichter | btw, what are overlays? | 14:33 |
| bskahan | srichter: external calendars layered over the users calendar view | 14:34 |
| bskahan | they're indicated be a subscrtiption relationship to the external calendar | 14:35 |
| srichter | but Person has an overlaid_calendars attribute | 14:36 |
| srichter | if it is view specific, why is it an attribute of the class? | 14:37 |
| srichter | bskahan: also, what is the difference between OverlaidCalendar[Property] and OverlaidCalendarAndTT[Property]? | 14:39 |
| bskahan | because it persists with the user I suspect | 14:39 |
| srichter | *shrug*, it should really be an annotation | 14:39 |
| bskahan | OC is a subscription to a regular ICalendar, OCaTT is a subscription to a Timetable | 14:40 |
| bskahan | so Johnny can look at his calendar, the lacrosse teams calendar, and his Biology class timetable all together | 14:41 |
| srichter | so Timetable subclasses Calendar? | 14:41 |
| mgedmin | morning | 14:42 |
| bskahan | I'm not sure about that part, I don't think it does | 14:42 |
| bskahan | hey mgedmin | 14:42 |
| srichter | the problem is that code between the SB/ST Person implementation cannot be shared because of this property | 14:43 |
| mgedmin | instead of the property you could use getRelatedObjects directly | 14:43 |
| mgedmin | in view classes | 14:43 |
| * mgedmin shrugs | 14:43 | |
| mgedmin | the data is stored in annotations, like all relationship data | 14:44 |
| bskahan | mgedmin: we could store it in the preferences annotation | 14:44 |
| mgedmin | the property is just syntactic sugar to make the code more readable | 14:44 |
| bskahan | the actual property | 14:44 |
| bskahan | since it is a perference of sorts | 14:44 |
| srichter | in that case I would suggest a helper function | 14:45 |
| srichter | in the view code | 14:45 |
| srichter | this property has really nothing to do there | 14:45 |
| srichter | seeing the word "color" in schooltool.app is an indication that something is fishy | 14:46 |
| mgedmin | the view already has a helper function | 14:48 |
| mgedmin | I think | 14:48 |
| mgedmin | ah, but there's more than one view | 14:49 |
| srichter | so, could we not move this code there then? | 14:49 |
| mgedmin | maybe an adapter would be cleaner then? | 14:49 |
| srichter | so the helper function should be a function, not a method on the class | 14:49 |
| *** ignas has joined #schooltool | 14:49 | |
| mgedmin | there's more than just a single helper function, I think | 14:50 |
| mgedmin | views need enumeration and modification also | 14:50 |
| srichter | right | 14:50 |
| srichter | I just saw that | 14:50 |
| srichter | mgedmin: btw, I think the merge should not be as bad as we have feared | 14:53 |
| srichter | I cannot see any show-stopping obstacles anymore | 14:54 |
| srichter | bskahan: mgedmin: one way of keeping overlaid_calendars would be to put it in an adapter | 14:55 |
| bskahan | I'm guessing that's what has to happen, iirc, "colors" is in sb.app because we want the color associated with a particular overlaid calendar to remain consistent | 14:57 |
| srichter | ok | 14:58 |
| *** jinty has quit IRC | 15:04 | |
| srichter | I also think that your entire preferences system needs to change; currently, whenever some feature requires a new preference you just stick it on | 15:05 |
| srichter | this means that any feature that requires preferences has to be hard-wired into the application for good | 15:05 |
| srichter | I think the new zope.app.preference code can be reutilized (rewritten) to address the issue better | 15:06 |
| srichter | since you have preference namespaces | 15:06 |
| bskahan | I'll look at that, I noticed zope.app.preference about 3 days after doing our preferences | 15:07 |
| bskahan | I think I was wrong about making date/time formatting a preference as well, it would have been better to use the locale | 15:08 |
| srichter | zope.app.preference is for principals only so I hope it will be flexible enough for all sorts of objects | 15:08 |
| srichter | yep | 15:09 |
| srichter | but that can be quickly redeemed | 15:09 |
| srichter | bskahan: do we have preferences for anything else other than people? | 15:09 |
| srichter | I guess an application can have preferences as well | 15:09 |
| bskahan | it does | 15:10 |
| srichter | ok, so I hope it would be as simple as registering a different adapter to get the preferences work | 15:10 |
| bskahan | there is a site preference and a person preference | 15:10 |
| bskahan | I have to run to the store, I'll be back in about 30 minutes and I'll take a look then | 15:11 |
| srichter | ok | 15:11 |
| srichter | bskahan: I would prefer this to be done on the refactor branch, so that we keep those changes together | 15:11 |
| bskahan | srichter: ok | 15:14 |
| *** SteveA has joined #schooltool | 15:46 | |
| povbot | /svn/commits: * srichter committed revision 4622: | 16:02 |
| povbot | /svn/commits: Moved application initialization event notification to SchoolBell. | 16:02 |
| povbot | /svn/commits: Cleaned up the interfaces and app module a bit. | 16:02 |
| *** gintas has joined #schooltool | 16:23 | |
| povbot | /svn/commits: * srichter committed revision 4623: | 16:36 |
| povbot | /svn/commits: Moved canModify from SchoolTool to SchoolBell. | 16:36 |
| povbot | /svn/commits: Made the SchoolBellApplicationTraverser somewhat sane, so that it can be simply reused in SchoolTool. (But I really think an adapter is not the right thing to do here. The calendar, the reason the custom traverser exists, should be available via a special namespace ++calendar++) | 16:36 |
| mgedmin | "The calendar, the reason the custom traverser exists, should be available via a special namespace ++calendar++" | 16:37 |
| mgedmin | -1 for ugly URLs | 16:37 |
| mgedmin | Zope 3 should allow me to specify traversal in a more fine-grained manner | 16:37 |
| *** jinty has joined #schooltool | 16:38 | |
| mgedmin | <browser:traverse for="...ICalendarOwner" name="calendar" component=".traverseIntoCalendar" /> | 16:38 |
| mgedmin | def traverseIntoCalendar(context): return context.calendar | 16:38 |
| srichter | that would be one | 16:39 |
| srichter | however, what if ever someone calls an item in STApp 'calendar' | 16:39 |
| srichter | then it will be a heck of a debugging session | 16:40 |
| srichter | the current traverser is the worst case, since it is not extensible at all | 16:40 |
| mgedmin | yeah... | 16:40 |
| mgedmin | by the way, what exactly forbids me to call an item '@@contents.html'? | 16:40 |
| *** admp has joined #schooltool | 16:40 | |
| srichter | I will never forget the namespaces/traversal discussion a couple of years ago | 16:41 |
| mgedmin | what discussion? | 16:41 |
| * mgedmin has apparently forgotten it :-) | 16:41 | |
| *** admp has quit IRC | 16:41 | |
| srichter | about why we want to have traversal namespaces | 16:41 |
| srichter | I think you might not have been around back then ;-) | 16:41 |
| mgedmin | no, I used to read zope3-dev a couple of years ago | 16:42 |
| mgedmin | I can't find the time to do read it now :( | 16:42 |
| srichter | @@ is a shortcut for ++view++, so what happens, if someone starts with "++"? | 16:43 |
| mgedmin | anyway, I guess it is a question of point of view | 16:43 |
| mgedmin | if you're building a CMS, namespaces are important | 16:43 |
| srichter | it turns out that we decided to document these cases well and note it as the only real restriction of the name | 16:43 |
| mgedmin | if you're building an app, you can declare "I know that no one will ever have an item called 'calendar' within the top-level app container" | 16:43 |
| bskahan | I don't think users like to see ++foo++ and @@foo | 16:43 |
| mgedmin | users probably don't care | 16:44 |
| mgedmin | nitpicky developers like me do | 16:44 |
| srichter | that is true as long as you control the development | 16:44 |
| mgedmin | as a web debeloper I want to be able to use the URL scheme that I want to use | 16:44 |
| mgedmin | not have one enforced by the framework | 16:44 |
| srichter | but once you want a development platform, you should loose as many assumptions as possible | 16:44 |
| srichter | I think Zope is the most flexible framework of all ;-) look at other application URLs | 16:45 |
| mgedmin | twisted wasn't too bad | 16:45 |
| srichter | but twisted really does not offere a Web framework | 16:45 |
| mgedmin | although it was very centralized | 16:45 |
| srichter | its just the server | 16:45 |
| mgedmin | (I meant twisted.web) | 16:46 |
| srichter | isn't twsited.web just the server? | 16:46 |
| ignas | srichter, well zope is not that flexible, yet flexible/capable ratio is very high | 16:46 |
| mgedmin | no, twisted.internet.http was the server | 16:46 |
| ignas | assembler is the most flexible :) | 16:46 |
| mgedmin | twisted.web was a publication framework | 16:46 |
| srichter | oh, ok | 16:46 |
| mgedmin | it has been superseded by nevow | 16:46 |
| srichter | I see | 16:46 |
| mgedmin | or maybe it was twisted.woven, not twisted.web? | 16:47 |
| mgedmin | I forgot | 16:47 |
| srichter | yeah, twisted.woven, probably | 16:47 |
| srichter | because the new Web server is called twisted.web2 | 16:47 |
| mgedmin | I found it difficult to write custom traversers in zope 3 | 16:48 |
| mgedmin | you have to implement IBrowserPublisher, figure out what to do with browserDefault when in fact you do not want to do anything | 16:48 |
| mgedmin | and it is hard to chain traversers | 16:48 |
| srichter | well, you could write an IChainTraverser | 16:49 |
| srichter | that uses subscribers | 16:49 |
| srichter | I think this might be the way to go in Zope 3 for the future | 16:49 |
| mgedmin | if I could then make it part of Zope 3 core, used by default, then it would be worthwile | 16:49 |
| srichter | I guess I talk to Jim about it in F12g | 16:49 |
| srichter | mgedmin: I think you would have good chances to get it in the core | 16:50 |
| srichter | I would certainly support it | 16:50 |
| srichter | Jim probably too | 16:50 |
| mgedmin | the hard part is designing it correctly, I think | 16:50 |
| mgedmin | how do you combine subscribers with container items, e.g. | 16:50 |
| srichter | <still writing?> | 16:53 |
| mgedmin | no | 16:54 |
| srichter | ok | 16:54 |
| mgedmin | perhaps my sentence was badly phrased | 16:54 |
| srichter | what do you mean by combining problems | 16:54 |
| srichter | so you have an IChainableTraverser registered for all interfaces | 16:54 |
| srichter | then you register a mult adapter for (IChainableTraverser, IContainer) that looks up items | 16:55 |
| mgedmin | I am confused | 16:56 |
| mgedmin | currently URL traversal relies on IBrowserPublisher | 16:56 |
| srichter | you leave the current system in place | 16:56 |
| mgedmin | so you would have an IBrowserPublisher multi-adapter registered for all interfaces | 16:56 |
| srichter | yes | 16:56 |
| mgedmin | that does zapi.queryAdapter(context, ISingleNameTraverser, name) | 16:57 |
| mgedmin | while <browser:traverse for="..." name="..." /> registers these ISingleNameTraverser | 16:57 |
| mgedmin | s | 16:57 |
| srichter | for example, but I think this is not correct | 16:57 |
| mgedmin | (better name anyone)? | 16:57 |
| mgedmin | yes, this is not entirely correct | 16:57 |
| srichter | you should make use of subscribers | 16:57 |
| mgedmin | perhaps | 16:58 |
| srichter | go through a list of subscribers, until one returned a valid object | 16:58 |
| mgedmin | however with subscribers don't you get ordering issues? | 16:58 |
| srichter | the assumption would be what comes first is found first | 16:58 |
| srichter | this assumption is true right now as well | 16:59 |
| mgedmin | so, zcml declaration order suddenly influences things | 16:59 |
| mgedmin | ok, not "suddenly" | 16:59 |
| mgedmin | it's just that I don't like it much | 16:59 |
| mgedmin | zcml ordering is hard to control | 16:59 |
| srichter | yes | 16:59 |
| mgedmin | when you have many zcml files scattered through multiple packages | 17:00 |
| srichter | at some point we probably want to add a weight to a subscriber | 17:00 |
| mgedmin | I once tried to register a subscriber for DatabaseOpenedEvent | 17:00 |
| mgedmin | that would get called before the one in zope.app | 17:00 |
| povbot | /svn/commits: * jinty committed revision 4624: | 17:00 |
| povbot | /svn/commits: add debian revision to changelog entry | 17:00 |
| ignas | srichter, can't we have the ordering done on "which traverser is the most specific one" ? | 17:00 |
| srichter | ahh, this is a bad assumption | 17:00 |
| srichter | subscribers should *never* rely on other subscribers or their result on the same level | 17:01 |
| mgedmin | (I gave up, because ZCML metadirectives that let me register subscribers were registered from the same zcml file that pulled in that bootstrap subscriber) | 17:01 |
| * mgedmin thinks that bootstrap subscribers currently *do* rely on other subscribers | 17:01 | |
| srichter | then that's bad | 17:02 |
| * mgedmin remembers that one subscriber creates the root folder, while another subscriber populates the site with some default utilities | 17:02 | |
| mgedmin | uhh, assuming I remember it correctly | 17:02 |
| * jinty idly wonders if it will help to have schoolbell under trunk/schoolbell and an svn external under trunk/schoolbell | 17:02 | |
| srichter | jinty: huh? | 17:03 |
| srichter | jinty: typo? | 17:03 |
| srichter | mgedmin: yeah, if this is the case, then it is used incorrectly | 17:03 |
| jinty | srichter: i mean the same repo layout as now, without zpkgtools, but with schoolbell stored in schooltool and the trunk/schoolbell/src/schoolbell an external | 17:05 |
| srichter | jinty: yes, that would be a step in the right direction, so I am working on refactoring all this now to make schoolbell depend on schooltool and not the other way around | 17:06 |
| srichter | why do we have a ST/SB container.pt | 17:07 |
| jinty | gintas: thanks for adding the translations, but one is missing the japanese schooltool one | 17:07 |
| srichter | they pretty much seem to be the same; ST has the canModify code, but that should be ported to SB anyways | 17:08 |
| gintas | jinty, thanks for noticing | 17:08 |
| jinty | strichter: I copied it, to get a release out the door | 17:08 |
| srichter | ok, so I can merge the two again? | 17:08 |
| gintas | for some reason I thought I had added it ;-/ | 17:08 |
| srichter | basically moving ST's into SB's? | 17:08 |
| srichter | SB seems to support a SEARCH attr in request | 17:09 |
| jinty | srichter: when you told me how to do it correctly, the diff was too big for me to resolve (as I did with the others) | 17:09 |
| * jinty knows very little about zpt's | 17:09 | |
| srichter | I use kompare | 17:09 |
| bskahan | http://linuxreviews.org/kde/kde-user-persp/kompare.html | 17:10 |
| bskahan | that? | 17:10 |
| srichter | yes | 17:10 |
| srichter | it is good for large diffs | 17:11 |
| jinty | gintas: no-worries. Hopefully I will move the debian stuff to alioth.debian.org and then will be able to commit my own translations... | 17:11 |
| gintas | kompare is OK | 17:11 |
| bskahan | there's something similar for gnome (meld, iirc) but it didn't seem substantially better than gvimdiff | 17:12 |
| jinty | srichter: the diff was small, the problem is that I didn't know how to resolve it | 17:12 |
| srichter | I see | 17:12 |
| gintas | it's just that I had such a strange feeling that I had committed that translation, I don't know why | 17:12 |
| srichter | ok, I am just copying ST's version into SB | 17:12 |
| jinty | gintas: perhaps svn ci? | 17:13 |
| mgedmin | I think it was either gintas and ignas | 17:13 |
| mgedmin | the change was committed while schoolbell was frozen for release | 17:13 |
| mgedmin | and I even saw an XXX saying "merge this back to schoolbell once release is done" | 17:13 |
| srichter | yes | 17:13 |
| srichter | I did that part already | 17:13 |
| mgedmin | thanks! | 17:14 |
| srichter | but I did not realize that this was true for the template as well | 17:14 |
| srichter | (note that it is all on the branch) | 17:14 |
| srichter | if we decide not to go for the refactorings, it will probably get lost, since it was not the first thing I did | 17:14 |
| jinty | while all are here I want to remove the zope external on schoolbell | 17:14 |
| gintas | I don't mind | 17:15 |
| gintas | that external is a pain | 17:15 |
| srichter | I don't mind, because I am on the branch :-) and I can still test schoolbell separately :-) | 17:15 |
| jinty | ok, in progress | 17:16 |
| mgedmin | how about a new Makefile rule then | 17:16 |
| mgedmin | Zope3: | 17:16 |
| mgedmin | <tab>svn co svn://svn.zope.org/t | 17:16 |
| mgedmin | ... | 17:16 |
| mgedmin | my goal is that the following should work | 17:16 |
| mgedmin | svn co http://source.schooltool.org/.../schooltool; cd schooltool; make test ftest | 17:16 |
| mgedmin | (assuming you have all the external dependencies -- python, libxml2, reportlab, whatever) | 17:17 |
| gintas | cd: Bad command or file name | 17:17 |
| gintas | make: Bad command or file name | 17:17 |
| gintas | ;) | 17:17 |
| mgedmin | if you insist using a silly OS, you can translate the commands yourseld | 17:18 |
| gintas | no, sorry, cd should work | 17:18 |
| mgedmin | besides, even MS-DOS has 'cd' as a shell builtin | 17:18 |
| * jinty thinks he understands what mgedmin wants and agrees | 17:18 | |
| mgedmin | if you have a Zope3 symlink, or check it out yourself, make will skip that rule, right? | 17:18 |
| mgedmin | otherwise we make Zope3 a dependency for the 'build' target | 17:19 |
| mgedmin | and 'make test' will then automatically do a checkout | 17:19 |
| gintas | jinty, I'm a bit lost with those translations | 17:19 |
| jinty | but I want to test against the debian zope packages | 17:19 |
| mgedmin | (once we can depend on a released version of Zope 3, we'll drop that rule, and end users will just have to install zope3 as another dependency) | 17:20 |
| gintas | do schooltool debconf translations work on SchoolBell? | 17:20 |
| gintas | the other way around? | 17:20 |
| gintas | the Japanese translation is for the 'schoolbell' package, not schooltool | 17:20 |
| gintas | and I indeed did add it earlier, but to schoolbell, as the .po file says in the header | 17:20 |
| jinty | gintas: I will investigate... the two should be almost exactly the same | 17:21 |
| jinty | perhaps exactly | 17:21 |
| gintas | jinty, if you don't want Z3 built, you can do `touch Zope3` | 17:21 |
| gintas | s/built/checked out/g | 17:21 |
| jinty | gintas: true, but in the future I will be sure to drop that rule (once Z3.1 is out) | 17:23 |
| mgedmin | what I said | 17:23 |
| jinty | yep, I cheekely copied your words (they are GPL?) | 17:24 |
| jinty | but tests should really be done against the released tarball rather than an svn checkout | 17:24 |
| mgedmin | or both | 17:27 |
| jinty | indeed | 17:27 |
| povbot | /svn/commits: * jinty committed revision 4625: | 17:28 |
| povbot | /svn/commits: small sync with schooltool setup.py | 17:28 |
| mgedmin | yay, if we merge st & sb into one source tree, there will be only one copy of test.py again! | 17:28 |
| povbot | /svn/commits: * jinty committed revision 4626: | 17:30 |
| povbot | /svn/commits: Remove the Zope3 external and provide a makefile rule to checkout the Zope 3.1 branch. | 17:30 |
| *** bskahan has quit IRC | 17:32 | |
| jinty | gintas: I mean this translation: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=319589 | 17:35 |
| *** mgedmin has quit IRC | 17:40 | |
| gintas | jinty, oh, right | 17:45 |
| povbot | /svn/commits: * srichter committed revision 4627: | 17:46 |
| povbot | /svn/commits: Merged container.pt from ST back here. This way, we only need one. | 17:46 |
| povbot | /svn/commits: * gintas committed revision 4628: | 17:47 |
| povbot | /svn/commits: Added Japanese debconf translation for SchoolTool. | 17:47 |
| *** ignas has quit IRC | 17:57 | |
| * jinty is quite happy with the Zope3 makefile rule (thanks mgedmin) and will do the same for schooltool soon | 18:10 | |
| povbot | /svn/commits: * jinty committed revision 4629: | 18:11 |
| povbot | /svn/commits: Remove Zope3 from the schoolbell tarball. | 18:11 |
| povbot | /svn/commits: * jinty committed revision 4630: | 18:13 |
| povbot | /svn/commits: Update the schoolbell readme to reflect the removal of Zope3 and debian directories from the tarball. | 18:13 |
| povbot | /svn/commits: * jinty committed revision 4631: | 18:18 |
| povbot | /svn/commits: Make a semi-branch for the post Zope3 packaging, this will replace the current debian-packaging trunk after the current release and is the place for new packaging development (hi gintas!) | 18:18 |
| *** FarcePest has joined #schooltool | 18:50 | |
| povbot | /svn/commits: * gintas committed revision 4632: | 19:25 |
| povbot | /svn/commits: Fixed double quotes / missing quotes. | 19:25 |
| gintas | jinty, to be honest, I don't like the new packaging branch very much while the current one is being worked on | 19:25 |
| gintas | on the other hand, we should be able to merge the changes of the 'stable' branch to the new one before replacing the old one | 19:25 |
| gintas | still this looks a bit fishy | 19:25 |
| gintas | tvon, I really want you to fix issue328 | 19:27 |
| gintas | I intend to do an rc3 soon, probably tomorrow | 19:27 |
| gintas | issue329 would be nice to close too | 19:28 |
| tvon | gintas: I just need to finish up some tests for 328 and I'll move to 329 | 19:32 |
| gintas | great | 19:34 |
| jinty | gintas: I'll take care of the merge if that is what is worrying you. | 19:35 |
| gintas | jinty, OK then | 19:35 |
| gintas | I still intend to do some improvements to debian packaging, and we'll probably want to land those in your branch as well | 19:36 |
| jinty | FYI: I am going to land some big changes there soon | 19:37 |
| gintas | hmm, wasn't the site calendar in SchoolBell supposed to be public? | 19:37 |
| povbot | /svn/commits: * gintas committed revision 4633: | 19:41 |
| povbot | /svn/commits: Fixed a stupid bug in functional test. Added a commented out section which fails because of issue330. | 19:41 |
| * jinty want to scream at the person who broke pbuilder and debootstrap in etch.... aaargh | 19:43 | |
| povbot | /svn/commits: * gintas committed revision 4634: | 19:59 |
| povbot | /svn/commits: Workaround for issue330. Turned SBApplication title into a property and added security declaration. | 19:59 |
| tvon | gintas: did you see my commit from last night? | 20:00 |
| tvon | index.pt in schoolbell was missing 'view/update' | 20:00 |
| povbot | /svn/commits: * jinty committed revision 4635: | 20:02 |
| povbot | /svn/commits: Extend build-debs.py to optionally use pdebuild. | 20:02 |
| *** thisfred has quit IRC | 20:06 | |
| *** mgedmin has joined #schooltool | 20:07 | |
| gintas | tvon, yes, I did | 20:07 |
| gintas | didn't get to it yet | 20:07 |
| tvon | What did you mean about the site calendar being public (in the scrollback)? | 20:08 |
| gintas | well, I was updating one functional test and found that in the ftest an ordinary user could not see the site calendar | 20:09 |
| tvon | There was some discussion around the idea that SB should have an auth-only site-calendar, but in the end I think it is supposed to be public | 20:10 |
| * tvon looks into adding an ftest for anon site-calendar | 20:13 | |
| jinty | if a test fails, does test.py have a non-zero exit status? | 20:15 |
| gintas | yes | 20:15 |
| jinty | yay | 20:16 |
| gintas | oh, right, bskahan was working on this | 20:16 |
| gintas | he had a fixed but was lacking a generation script | 20:16 |
| gintas | s/fixed/fix/ | 20:16 |
| tvon | ah right | 20:21 |
| gintas | I didn't see him around today | 20:22 |
| *** bskahan has joined #schooltool | 20:24 | |
| bskahan | tvon: there are ftests in trunk for anonymous site calendars, iirc | 20:27 |
| povbot | /svn/commits: * jinty committed revision 4636: | 20:28 |
| povbot | /svn/commits: SchoolBell Packaging Improvements: | 20:28 |
| povbot | /svn/commits: * Zope3 removed from upstream tarball: | 20:28 |
| povbot | /svn/commits: * convert libschoolbell to an arch independent python2.4-schoolbell * depend on zope3-lib * remove the install-arch rule | 20:28 |
| povbot | /svn/commits: * implement build time test running | 20:28 |
| povbot | /svn/commits: * build depend on dependencies so we can run the upsream tests at build time * fail to build if tests fail * add check rule to run tests | 20:28 |
| povbot | /svn/commits: * don't pass the python version to dh_python, it is not necessary anymore as | 20:28 |
| povbot | /svn/commits: the package names follow the python policy | 20:29 |
| povbot | /svn/commits: * srichter committed revision 4637: | 20:29 |
| povbot | /svn/commits: Add the only SchoolTool PersonPreference here for now till we develop a better system. | 20:29 |
| gintas | bskahan, what about that missing generation script? | 20:30 |
| gintas | can you finish it up today? | 20:30 |
| bskahan | gintas: yes, sorry about that I meant to do it yesterday and lost track | 20:30 |
| bskahan | doing it now | 20:30 |
| *** ignas has joined #schooltool | 20:33 | |
| bskahan | gintas: I may not finish this afternoon, but I'll get it done by tonight at the latest for an RC3 | 20:33 |
| povbot | /svn/commits: * jinty committed revision 4638: | 20:34 |
| povbot | /svn/commits: Make sure that *.mo files dont make it into the tarball. Also add *.pyd files for good measure. | 20:34 |
| bskahan | gintas: the release branch has evolve1 and evolve2, trunk has an additional evolve3 | 20:37 |
| bskahan | can I create evolve4 in trunk and skip evolve3 in the branch? | 20:38 |
| gintas | oh | 20:42 |
| gintas | this is sticky | 20:42 |
| povbot | /svn/commits: * gintas committed revision 4639: | 20:42 |
| povbot | /svn/commits: The build-debs script now does not expect Zope 3 externals and will build distribution tarballs itself. | 20:42 |
| povbot | /svn/commits: * jinty committed revision 4640: | 20:44 |
| povbot | /svn/commits: Sync changes from schoolbell MANIFEST.in. | 20:45 |
| gintas | I think you should move the current evolve3 to evolve4 and add the new evolution script as evolve3 | 20:45 |
| bskahan | gintas: ok, thanks | 20:45 |
| gintas | this is not nice, but things would get much nastier during the next release if we did not do that | 20:45 |
| * bskahan nods | 20:45 | |
| *** jinty has quit IRC | 20:59 | |
| povbot | /svn/commits: * jinty committed revision 4641: | 21:06 |
| povbot | /svn/commits: delete experimental nozope3 branch | 21:06 |
| bskahan | gintas: I'm not sure we should have a generation for this change | 21:11 |
| bskahan | its a simple generation script, but it make in place, unalerted security policy changes | 21:11 |
| tvon | true | 21:11 |
| bskahan | I would be very unhappy if this happened to my production site | 21:12 |
| povbot | /svn/commits: * jinty committed revision 4642: | 21:13 |
| povbot | /svn/commits: forgot to close the bug, also fixed formatting. | 21:13 |
| bskahan | its a tricky qestion, and generally a thorny issue though | 21:14 |
| bskahan | if I'm an admin and I go to the ACL form at the root of my site and grant 'view', and 'viewCalendar' to unauthenticated users I'm opening up the whole site to unauth | 21:15 |
| gintas | do you imply that an evolution script is not needed here? | 21:21 |
| bskahan | I'm saying the evolution script will surprise site admins running 0.10 | 21:23 |
| bskahan | because of inherited security policy | 21:23 |
| bskahan | I'm not sure what the correct answer is, I have a working generations script that will move a 0.10 site to the 0.11 security policy | 21:24 |
| tvon | there was no site-wide calendar | 21:24 |
| tvon | so we are not opening up anything new :) | 21:25 |
| bskahan | tvon: evolve2.py creates the calendar | 21:25 |
| bskahan | we are though | 21:25 |
| bskahan | or closing things they may have opened | 21:25 |
| bskahan | because we set view access on the root, anything under that with unset security policy (which used to be restricted) is open. | 21:25 |
| tvon | evolve2? | 21:26 |
| tvon | the site-wide calendar? | 21:26 |
| tvon | ah | 21:26 |
| bskahan | tvon: yeah, there's allready a generation for that | 21:26 |
| bskahan | this generation is just supposed to get the ACL set for the calendar | 21:27 |
| tvon | but that was not in .10 | 21:27 |
| bskahan | yeah | 21:27 |
| tvon | so this is only going to hit people running from svn | 21:27 |
| tvon | or RC's | 21:27 |
| bskahan | no, because the security policy is inherited | 21:28 |
| tvon | what? | 21:28 |
| tvon | I dont follow | 21:28 |
| bskahan | someone with 0.10 used to have one set of default security policy, plus any changes they made since install | 21:28 |
| tvon | so we might close-off containers that were previously opened? | 21:29 |
| tvon | by the admin? | 21:29 |
| bskahan | now we want to grantAccess for 'view' and 'viewCalendar' to the root, that means anything under the root that has an unset security policy may be opened | 21:29 |
| bskahan | right | 21:29 |
| bskahan | because after opening root we close individual containers | 21:30 |
| bskahan | with a denyAccess, rather than an unsetAccess | 21:30 |
| gintas | so close the individual containers in the evolution script then | 21:30 |
| * bskahan shrugs | 21:30 | |
| tvon | you'd have to check the settings in the evolve script before you change anything | 21:30 |
| bskahan | ugh | 21:32 |
| bskahan | currently I have a working script that open root, closes persons, groups, resource, sections, and courses. which will make a 0.10 function like a new 0.11 site. | 21:33 |
| bskahan | tvon: I'm afraid you might be right about checking the current settings though | 21:33 |
| *** SteveA has quit IRC | 21:37 | |
| *** gintas has quit IRC | 21:37 | |
| *** munkee has quit IRC | 21:37 | |
| *** bskahan has quit IRC | 21:37 | |
| *** ignas has quit IRC | 21:37 | |
| *** auxesis has quit IRC | 21:37 | |
| *** mgedmin has quit IRC | 21:37 | |
| *** FarcePest has quit IRC | 21:37 | |
| *** tvon has quit IRC | 21:37 | |
| *** Aiste has quit IRC | 21:37 | |
| *** Ricey has quit IRC | 21:37 | |
| *** srichter has quit IRC | 21:37 | |
| *** povbuildbot has quit IRC | 21:37 | |
| *** povbuildbot has joined #schooltool | 21:38 | |
| *** ignas has joined #schooltool | 21:38 | |
| *** gintas has joined #schooltool | 21:38 | |
| *** SteveA has joined #schooltool | 21:38 | |
| *** auxesis has joined #schooltool | 21:38 | |
| *** munkee has joined #schooltool | 21:38 | |
| *** mgedmin has joined #schooltool | 21:38 | |
| *** FarcePest has joined #schooltool | 21:38 | |
| *** tvon has joined #schooltool | 21:38 | |
| *** Aiste has joined #schooltool | 21:38 | |
| *** Ricey has joined #schooltool | 21:38 | |
| *** srichter has joined #schooltool | 21:38 | |
| *** bskahan has joined #schooltool | 21:40 | |
| bskahan | netsplit | 21:40 |
| tvon | silly opn | 21:42 |
| *** bskahan has quit IRC | 21:43 | |
| tvon | er, freenode | 21:43 |
| *** Aiste has quit IRC | 21:48 | |
| povbot | /svn/commits: * jinty committed revision 4643: | 21:52 |
| povbot | /svn/commits: Tweak to allow building the new non-zope3 schoolbell packages with the 1.2 tarballs. It builds, but I havnt tested it at all. | 21:52 |
| *** ignas has quit IRC | 22:02 | |
| povbot | /svn/commits: * srichter committed revision 4644: | 22:06 |
| povbot | /svn/commits: Removed a lot of duplicated code. Especially, I automated the setup of the SchoolToolApplication. | 22:06 |
| povbot | /svn/commits: * jinty committed revision 4645: | 22:06 |
| povbot | /svn/commits: Sync the schoolbell packaging changes to schooltool. Completely untested. | 22:06 |
| povbot | /svn/commits: * gintas committed revision 4646: | 22:08 |
| povbot | /svn/commits: Ported revision 4621 (missing call to view/update). | 22:08 |
| *** mgedmin has quit IRC | 22:10 | |
| srichter | tvon: are any of the calendar macros in view_macros.pt used? | 22:48 |
| tvon | erm, I'm not sure | 22:55 |
| tvon | schoolbell uses them afaik, schooltool has a calendar_macros file | 22:57 |
| srichter | no, it's the other way around :-) | 23:00 |
| srichter | calendar_macros is in SB | 23:01 |
| srichter | which makes all those macros in ST/browser/templates/view_macros.pt unused | 23:01 |
| tvon | ahh | 23:01 |
| srichter | I am developing right now a new Macros class | 23:02 |
| srichter | that allows you to say: view/@@schooltool_macros/calendar/page instead of view/@@standard_macros/calendar-page | 23:03 |
| tvon | that would be nice | 23:04 |
| srichter | the code is written and works | 23:04 |
| srichter | I just have to change all the templates now | 23:04 |
| povbot | /svn/commits: * srichter committed revision 4647: | 23:33 |
| povbot | /svn/commits: This is a new way of doing macros. It gives us another level for the macro namespace, which I think makes everything much nicer. | 23:33 |
| povbot | /svn/commits: Converted all calendar macros to use it. | 23:33 |
| tvon | nice | 23:34 |
| srichter | ok, done | 23:51 |
| srichter | I also demonstrated how to use macros effectively to write common view code | 23:52 |
| povbot | /svn/commits: * srichter committed revision 4648: | 23:52 |
| povbot | /svn/commits: Made main template more flexible. It should now be possible to remove the SchoolTool view_macros.pt file. | 23:52 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!