| *** didymo has joined #schooltool | 00:20 | |
| *** povbot has joined #schooltool | 10:31 | |
| *** jdrake has quit IRC | 10:36 | |
| *** jdrake has joined #schooltool | 12:01 | |
| *** alga has joined #SchoolTool | 12:31 | |
| *** didymo has quit IRC | 12:44 | |
| *** ignas has joined #schooltool | 13:42 | |
| *** mgedmin has joined #schooltool | 14:26 | |
| *** mgallagh has quit IRC | 15:19 | |
| *** jinty has joined #schooltool | 15:30 | |
| ignas | jinty: hi | 15:31 |
|---|---|---|
| jinty | hey ignas | 15:32 |
| jinty | get my mail? | 15:32 |
| ignas | yes | 15:32 |
| ignas | got my emails? | 15:32 |
| jinty | hmm, moment | 15:35 |
| *** th1a_ is now known as th1a | 15:37 | |
| th1a | Catching up... | 15:37 |
| jinty | ignas: Makes us dependent on paste though? | 15:38 |
| th1a | Does paste make it easier for us to use wsgi? | 15:39 |
| ignas | yes | 15:40 |
| ignas | it makes it possible to use wsgi | 15:40 |
| ignas | and it's a very small dependency imho | 15:40 |
| th1a | I think it is worth adding it as a depencency then. | 15:40 |
| ignas | oh and we don't depend on zdaemon then ;) | 15:40 |
| ignas | i think schooltool egg depends on paste anyway | 15:41 |
| jinty | we never did.... | 15:41 |
| ignas | oh | 15:41 |
| ignas | hmm | 15:41 |
| ignas | i mean - schooltool has code that depends on paste | 15:41 |
| jinty | start-stop-daemon was what we used | 15:41 |
| ignas | though in the future we could split it as well | 15:41 |
| ignas | so we'd have "schooltool-server" "python-schooltool" "python-schooltool.paste" | 15:41 |
| ignas | and python-schooltool.paste would be the one that has the make-instance, start-instance scripts in it | 15:42 |
| jinty | er, so could you send me the default paste.ini | 15:42 |
| jinty | we should be using in production? | 15:42 |
| ignas | ok | 15:43 |
| ignas | lisppaste5: url | 15:43 |
| lisppaste5 | To use the lisppaste bot, visit http://paste.lisp.org/new/schooltool and enter your paste. | 15:43 |
| lisppaste5 | ignas pasted "server|paste|schooltool.ini" at http://paste.lisp.org/display/59463 | 15:44 |
| ignas | we depend on python-zope.server 3.5 implicitly if we use that | 15:44 |
| ignas | but as it is the version in PPA, we can get away without specifying it | 15:44 |
| jinty | hmm, that looks like something that shouldn't be in /etc | 15:45 |
| jinty | not sure | 15:46 |
| jinty | also, what about python-schooltool.stapp2007 | 15:46 |
| ignas | hmm, i don't have it as a separate egg at the moment | 15:46 |
| ignas | stapp2007 is in schooltool itself | 15:47 |
| ignas | stapp2008spring is a separate package though | 15:47 |
| jinty | what does this do: use = egg:schooltool#main | 15:47 |
| jinty | ? | 15:47 |
| ignas | it tells paste to run schooltool application | 15:47 |
| jinty | what is #main? | 15:48 |
| ignas | #main is the name of the entrypoint | 15:48 |
| ignas | of schooltool egg | 15:48 |
| ignas | the configuration is selected in the site.zcml | 15:48 |
| ignas | site.zcml is set in schooltool.conf | 15:48 |
| jinty | what will happen to "web 7080" in the schooltool.conf | 15:48 |
| ignas | it is ignored | 15:49 |
| jinty | seems to me that there is a lot of duplicate info in these | 15:49 |
| ignas | and the meaning of rest declaration changes | 15:49 |
| jinty | eeew | 15:49 |
| ignas | we'd get that with mod_wsgi too | 15:49 |
| jinty | hmm, can we perhaps get rid of schooltool.conf completely? | 15:49 |
| ignas | hmm, not at the moment | 15:50 |
| jinty | I mean there's not much it's doing anymore? | 15:50 |
| ignas | i could not do that because we have start-schooltool-server.py | 15:50 |
| ignas | and it depends on all that info in schooltool.conf | 15:50 |
| ignas | and i had to keep it working | 15:50 |
| ignas | because we use it in our releases | 15:50 |
| ignas | oh and things like devmode, CAS authentication | 15:50 |
| ignas | ZODB | 15:50 |
| ignas | are defined in schooltool.conf | 15:51 |
| ignas | LDAP authentication | 15:51 |
| ignas | there are like 2 configs | 15:51 |
| ignas | the ini one | 15:51 |
| ignas | is for setting up the server | 15:51 |
| ignas | that uses schooltool application | 15:51 |
| ignas | and maybe some middleware | 15:51 |
| ignas | and schooltool.conf that configures schooltool application | 15:51 |
| jinty | ok, I see... | 15:51 |
| jinty | then we'd better deprecate the directives in schooltool.conf that don't apply | 15:52 |
| ignas | yes | 15:52 |
| ignas | but i'll do that only when i'll be able to remove start-schooltool-server | 15:52 |
| jinty | and the rest directive shouldn't change meaning. | 15:52 |
| jinty | cant start-schooltool-server become a paste command? | 15:53 |
| ignas | well - rest goes from rest <PORT> to rest <on> | 15:53 |
| jinty | yes, but then it should be an "enable-rest" directive | 15:53 |
| ignas | ok | 15:53 |
| ignas | jinty: as for changing the command | 15:53 |
| jinty | pehaps the rest directive could take on the new meaning | 15:53 |
| ignas | we can do that | 15:53 |
| jinty | but still be deprecated | 15:54 |
| jinty | enable-rest true | 15:54 |
| jinty | enable-rest false | 15:54 |
| ignas | makes sense | 15:54 |
| jinty | so, should I make a release for stapp2007 | 15:54 |
| jinty | or only aim at stapp2008spring? | 15:55 |
| ignas | you can aim at stapp2008spring i think | 15:55 |
| ignas | unless you want to have 2 releases | 15:55 |
| ignas | to start testing upgrade paths | 15:55 |
| jinty | what 's the paste for that? the same? | 15:55 |
| ignas | yes, the only difference is in site.zcml | 15:55 |
| jinty | we already have a stapp2007 package | 15:56 |
| ignas | the paste.ini is identical | 15:56 |
| jinty | or release | 15:56 |
| jinty | needs only to have a schooltool.stapp2007 egg | 15:56 |
| ignas | not sure I understand you, sorry | 15:57 |
| ignas | i can create an schooltool.stapp2007 egg if you want me to | 15:58 |
| jinty | i.e. the schooltool-2007 package corresponds to schooltool.stapp2007 and depends on python-schooltool.stapp2007 | 15:58 |
| jinty | which doesn't exist yet | 15:58 |
| ignas | could you just make it depend on "python-schooltool" or make python-schooltool provide schooltool.stapp2007 | 15:59 |
| th1a | I don't understand why we'd want a 2007 egg. | 15:59 |
| jinty | yeah, I can do that too | 15:59 |
| ignas | my idea was to see how it works with more than 1 configuration | 16:00 |
| ignas | so we'd resolve these issues earlier | 16:00 |
| ignas | than our real ubuntu release | 16:00 |
| th1a | OK. | 16:00 |
| ignas | because at least at the moment it seems that Brians and mine vision of how packages and servers should interact are slightly different | 16:00 |
| ignas | and it is better to have 6 months to resolve that than 1 week ;) | 16:01 |
| ignas | jinty: actually schooltool does really provide schooltool.stapp2007 | 16:01 |
| ignas | jinty: by the way - are you releasing the trunk or the 2007 branch at the moment? | 16:02 |
| jinty | ignas: wait a bit, I think it'll majke sense when you see the packages working | 16:02 |
| ignas | ok :) | 16:03 |
| jinty | can we use the default server as the paste server safely? | 16:04 |
| jinty | I'd rather not depend on zope.server... | 16:04 |
| jinty | but rather on a basic implementation | 16:04 |
| jinty | hmm, actually, it's not too bad | 16:05 |
| ignas | use = egg:Paste#http is the default i think | 16:05 |
| ignas | but i like zope.server better ;) | 16:06 |
| jinty | that also takes arguments host and port, right? | 16:07 |
| ignas | yes | 16:07 |
| ignas | identical | 16:07 |
| jinty | http://wiki.pylonshq.com/display/pylonscookbook/Scripts+for+paster+serve | 16:07 |
| jinty | how do I startup as a different user? | 16:07 |
| ignas | --user=USERNAME Set the user (usually only possible when run as root) | 16:08 |
| ignas | --group=GROUP Set the group (usually only possible when run as root) | 16:08 |
| ignas | paster serve --help | 16:08 |
| ignas | th1a: resolved all the "high" bugs, fixing the "medium" ones | 16:10 |
| th1a | Excellent. | 16:12 |
| jinty | ignas: what's the default for rest? | 16:21 |
| jinty | rest-enabled false? | 16:22 |
| ignas | yes | 16:22 |
| ignas | default is diabled | 16:22 |
| ignas | are you adding the section? | 16:22 |
| ignas | or just TODO documentation so i could add it | 16:22 |
| jinty | just documenting it a bit | 16:22 |
| jinty | pid-file is also deprecated, right? | 16:23 |
| ignas | yes | 16:23 |
| Lumiere | ignas: how does the start/stop script work then? | 16:26 |
| Lumiere | (hi) | 16:26 |
| ignas | Lumiere: in buildout you get bin/start_schooltool_server | 16:26 |
| ignas | and for people who are still not using buildout even though they should | 16:26 |
| ignas | i will probably change start-schooltool-server.py to do the right thing, which is use some paste.ini in schooltool-skel | 16:27 |
| ignas | and bin/start_schooltool_server sets the pid file to live in <instance>/var/schooltool.pid | 16:27 |
| ignas | i think | 16:27 |
| Lumiere | yes | 16:27 |
| Lumiere | ignas: the problem is, how does a start/stop script know which instance is which | 16:28 |
| Lumiere | I know because I rename the start/stop scripts | 16:28 |
| ignas | Lumiere: bin/start_schooltol_server is actually more like bin/control_schooltool_server, you pass it an instance directory | 16:29 |
| ignas | so and the script finds the right pid file | 16:29 |
| ignas | in it | 16:29 |
| ignas | so s_s_i instance | 16:29 |
| ignas | s_s_i /home/ignas/my_cando_test_instance | 16:30 |
| ignas | s_s_i /home/ignas/my_cando_test_instance --daemon | 16:30 |
| ignas | s_s_i /home/ignas/my_cando_test_instance --stop-daemon | 16:30 |
| Lumiere | yea | 16:30 |
| th1a | aelkner: Good morning. | 16:30 |
| aelkner | morning | 16:30 |
| ignas | makes it a lot lot easier to develop when you want to test something on an actual cando, lyceum and schooltool instance | 16:31 |
| ignas | without moving Data.fs around | 16:31 |
| th1a | ignas: So how are we looking for this release overall? | 16:32 |
| ignas | th1a: don't know really ;), we have Zope3 released and working, packaging schooltool egg is quite straightforward, so as soon as we will find out how we want our start/stop machinery looking on debian we will have an ubuntu release | 16:34 |
| ignas | *a ubuntu | 16:34 |
| th1a | OK. Based on the current trunk. | 16:34 |
| th1a | You don't think "an ubuntu" is correct? | 16:35 |
| ignas | hmm | 16:35 |
| ignas | in english pronounciation of ubuntu - it's a in african - it's an ;) | 16:36 |
| ignas | so i wrote it pronouncing ubuntu one way, and then re read it in english ;) | 16:36 |
| th1a | Anyway... | 16:36 |
| Lumiere | a(n) and every choice is covered | 16:36 |
| th1a | We just want to make sure this last batch of bugfixes is included. | 16:37 |
| jinty | ignas: | 16:37 |
| jinty | failed! | 16:37 |
| jinty | ? | 16:37 |
| th1a | So I guess we should tag this at some point. | 16:37 |
| jinty | usr/bin/paster serve: error: no such option: --monitor-restart | 16:37 |
| ignas | paster serve some.ini --monitor-restart? | 16:37 |
| jinty | paster serve --monitor-restart some.ini | 16:38 |
| jinty | is what i'm doing | 16:38 |
| ignas | --daemon ? | 16:38 |
| jinty | $PASTER serve $PASTER_ARGS $CONFIGFILE start | 16:39 |
| jinty | where | 16:39 |
| ignas | paster serve schooltool.ini --daemon --monitor-restart goes into infinite loop if i do not set PYTHONPATH right, so it seems that it's working | 16:39 |
| jinty | PASTER_ARGS="--monitor-restart --user=schooltool --group=schooltool --daemon --pid-file=$PIDFILE --log-file=/var/log/schooltool/paste.log" | 16:39 |
| ignas | put --monitor-restart after --daemon | 16:39 |
| ignas | though ok - no difference | 16:39 |
| ignas | PasteScript 1.3.5 from /usr/lib/python2.5/site-packages (python 2.5.1 (r251:54863, Mar 7 2008, 04:10:12)) | 16:40 |
| ignas | paster --version | 16:40 |
| jinty | jinty@etch:~/src/schooltoolrepo/schooltool-2007 # paster --version | 16:40 |
| jinty | PasteScript 1.0 from /usr/lib/python2.4/site-packages (python 2.4.4 (#2, Apr 5 2007, 20:11:18)) | 16:40 |
| jinty | still using debian etch | 16:40 |
| ignas | ubntu has a more up to date version :/ | 16:41 |
| ignas | so i guess we'll have to do without monitoring | 16:41 |
| ignas | on etch | 16:41 |
| jinty | not sure if it's a good idea anyway | 16:41 |
| ignas | th1a: as for releases it seems that the release infrastructure by jinty is so good, that there should not be any problems with bugfix releases at all | 16:43 |
| th1a | Yes, I'm very happy about that. | 16:43 |
| Lumiere | is there any chance we can replicate it? :) | 16:43 |
| Lumiere | I'd love to be able to push tweaked schooltool/cando for internal releases | 16:44 |
| ignas | Lumiere: as soon as cando will be compatible with schooltool trunk | 16:45 |
| ignas | Lumiere: we'll be able to make a cando package that will be installable with schooltool without tweaking i think | 16:45 |
| Lumiere | ignas: i was more thinking about if acps added stuff that wasn't in the release | 16:46 |
| Lumiere | I agree with the cando package installing with schooltool without tweaking | 16:46 |
| Lumiere | (although it may need to make its own instance) | 16:46 |
| ignas | Lumiere: if acps would be in an egg, releasing it into schooltool ppa or into cando ppa should not be too difficult | 16:47 |
| ignas | and then you could just apt-get install it | 16:48 |
| Lumiere | yea | 16:50 |
| th1a | Make sense? | 16:50 |
| Lumiere | let me stop derailing the conversation | 16:50 |
| Lumiere | yep | 16:50 |
| th1a | Well, this is the unstructured working mode meeting anyhow. | 16:50 |
| th1a | aelkner: Want to squeeze in your report here? | 16:50 |
| aelkner | sure | 16:51 |
| aelkner | last week i had a great and really revealing meeting with Chris Lehmann | 16:51 |
| aelkner | in the meeting he revealed what he really wanted from us, and it wasn't anything so specific as CSAP | 16:51 |
| aelkner | it was more a generic intervention system that could be used around the country | 16:52 |
| aelkner | not just in philly where they have that horrible CSAP system | 16:52 |
| aelkner | so i was able to refactor the sla sub-packages | 16:52 |
| aelkner | out into one genereic package called 'intervention' | 16:52 |
| aelkner | this added work, but was offset by the work i will no longer need to do | 16:53 |
| aelkner | or that i thought i was going to need to do with interfacing with the CSAP database | 16:53 |
| aelkner | Karina, the guidance counselor there will still have to manually enter the CSAP info | 16:54 |
| aelkner | into the district's web site | 16:54 |
| aelkner | though Chris and I were happy with the meeting, Karina was frowning :) | 16:54 |
| aelkner | however, she will be glad to have the intervention system when she sees it in action | 16:55 |
| aelkner | because it will replace the need for teacher to constantly fill out reports | 16:55 |
| aelkner | and for her to have to look for them all over the place | 16:55 |
| th1a | I think we should be able to save her some re-entry eventually. | 16:55 |
| aelkner | that will be a discussion that you and Chris will need to have | 16:56 |
| aelkner | it's a question of priorities that he was discussing in the meeting | 16:56 |
| Lumiere | I am personally looking forward to that | 16:56 |
| aelkner | namely, data integration | 16:56 |
| aelkner | i.e., the killer app | 16:56 |
| Lumiere | I am wondering if ACC could use something like it | 16:56 |
| aelkner | he was hoping to have that for Sept 1 | 16:56 |
| aelkner | i don't know | 16:56 |
| aelkner | we'll have to show it to them | 16:57 |
| aelkner | and see | 16:57 |
| aelkner | i'll be working with them this summer | 16:57 |
| aelkner | that's if you authorize it | 16:57 |
| th1a | Intervention tracking in general is a common need. | 16:57 |
| aelkner | that's another discussion | 16:57 |
| aelkner | that's the idea | 16:57 |
| th1a | Usually involving forms in triplicate and folders. | 16:57 |
| aelkner | i didn't know chriss had that much vision | 16:57 |
| aelkner | i won't underestimate him again | 16:57 |
| th1a | Oh, Chris has vision. | 16:58 |
| aelkner | yes, i'm looking forward to our meeting when you come down | 16:58 |
| th1a | Me too. | 16:58 |
| aelkner | so i have a clear vision of what i need to have ready for May 1 | 16:59 |
| th1a | Good. | 16:59 |
| aelkner | that's my report | 16:59 |
| ignas | th1a: i can't reproduce https://bugs.launchpad.net/schooltool/+bug/80105 no matter how i try, it seems impossible with current security policy, but your comment that it's still there was added when we were already using the new security policy :/ | 16:59 |
| th1a | Checking... | 17:00 |
| th1a | I'll try it again; you can move on until I confirm it again. | 17:00 |
| th1a | So right now I can add our PPA, do a big apt-get of the zope stuff, and then use a buildout? | 17:02 |
| th1a | To get a fresh SchoolTool? | 17:02 |
| ignas | no | 17:02 |
| ignas | buildout is not using things on your system | 17:02 |
| jinty | ignas: I just added uploaded new schooltool-common and schooltool-2007 packages | 17:02 |
| th1a | OK. | 17:02 |
| jinty | ignas: but I can't test them - don't have gutsy | 17:02 |
| ignas | schooltool-2007 is the branch or trunk? | 17:03 |
| jinty | schooltool-2007 is just a package | 17:03 |
| jinty | it's not even in subversion | 17:03 |
| ignas | oh | 17:03 |
| ignas | ahh ok, wrong question | 17:03 |
| ignas | schooltool-common is what? | 17:03 |
| jinty | its a set of reusable scripts and filesystem locations for schooltool packages | 17:04 |
| ignas | ahh | 17:04 |
| jinty | schooltool-2007 is an implementation | 17:04 |
| jinty | based on the postgresql-common and postgresql-8.x packaging | 17:05 |
| ignas | i see | 17:05 |
| ignas | and the python-schooltool package does not exist at the moment | 17:05 |
| ignas | oh, it does | 17:05 |
| jinty | yes, that's still from the old packaging | 17:06 |
| jinty | we should get rid of that | 17:06 |
| jinty | and make a new one from the egg | 17:06 |
| ignas | hmm, how do I package the trunk? | 17:06 |
| ignas | and what version number do we want? | 17:06 |
| jinty | I think the version number of the egg is not important | 17:07 |
| jinty | I would just make a schooltool egg | 17:07 |
| jinty | and use the release infrastructure | 17:07 |
| ignas | as in - upload a tarball ? | 17:08 |
| ignas | what I was thinking about was - setting up schooltool trunk in schooltool.buildconfig, setting some initial version and making a release using that | 17:09 |
| ignas | then checking out the tag | 17:09 |
| jinty | hmm, this is going to get complicated | 17:09 |
| ignas | packaging the egg | 17:10 |
| ignas | uploading the egg | 17:10 |
| ignas | signing the package and uploading to PPA | 17:10 |
| jinty | er, the builder should make the package for you | 17:10 |
| ignas | oh, it does? | 17:10 |
| ignas | where does it put it? | 17:10 |
| jinty | yeah, like it does for all other eggs | 17:10 |
| ignas | i haven't seen schooltool.lyceum.journal debian package anywhere | 17:10 |
| jinty | not sure anymore, defined in schooltool.buildconfig | 17:11 |
| ignas | i see, so we have most if not all the dependencies of schooltool covered already and only need the python-schooltool egg | 17:11 |
| jinty | more or less | 17:11 |
| jinty | the python-schooltool package just puts the module on the filesystem | 17:12 |
| ignas | yes, the egg | 17:12 |
| ignas | so paster for example could see the egg and it's entry points | 17:12 |
| jinty | the schooltool-common and schooltool-2007 packages are a higher layer | 17:12 |
| jinty | and add an apt-gettable server to that | 17:12 |
| ignas | and you could import schooltool in pyhon prompt | 17:12 |
| jinty | exactly | 17:12 |
| ignas | yep, so if you know what you are doing - you can run schooltool yourself using mod_wsgi for example | 17:13 |
| jinty | bingo | 17:13 |
| ignas | just after installing python-schooltool | 17:13 |
| jinty | yes | 17:13 |
| Lumiere | wsgi is? | 17:13 |
| ignas | Lumiere: JFGI | 17:13 |
| ignas | ;) | 17:13 |
| jinty | if you don't, you can do apt-get install schooltool-2007 | 17:13 |
| Lumiere | hai hai | 17:13 |
| ignas | jinty: so can I just say that schooltool is in 0.9.0 | 17:14 |
| ignas | jinty: release it using your tool | 17:14 |
| ignas | and remove the schooltool package that is 2007.0.alpha5 in ppa | 17:15 |
| ignas | or is there a better way | 17:15 |
| jinty | I think that, for historical reasons, you will want a version greater than 2007 | 17:15 |
| ignas | like having the new version 2008.0 | 17:15 |
| jinty | 2008.0.0 | 17:15 |
| ignas | 2008.0.0 | 17:15 |
| th1a | Or use a date format. | 17:15 |
| jinty | because the old python-schooltool package is going to get in the way | 17:16 |
| Lumiere | I suggest 2007.0.pre2008rc1 or some such | 17:16 |
| th1a | 2008.4.21 | 17:16 |
| ignas | 2008.04.0 | 17:16 |
| Lumiere | at least until you finish the first real release | 17:16 |
| ignas | th1a: i'd rather keep the last number non-day one so i could add more than 30 bugfix releases ;) | 17:16 |
| th1a | OK. | 17:16 |
| Lumiere | ignas: lets hope that doesn't happen | 17:16 |
| ignas | Lumiere: we won't get more real than that ;) | 17:16 |
| ignas | Lumiere: i hope it will happen, because we can release it like after every bugfix | 17:17 |
| ignas | Lumiere: and fixing all the bugs would be way cool ;) | 17:17 |
| ignas | 2008.4.0, and then +1 for every bug fixed ;) | 17:17 |
| th1a | In the long run, cranking out bugfixes into our PPA is very important. | 17:17 |
| ignas | jinty: so should I set up schooltool.buildconfig or will you do it? | 17:19 |
| jinty | ignas: Though I am of the opinion that having a schooltool egg as well a s a schooltool.gradebook egg is not such a nice thing, and that now is probably a good time to go exploding the schooltool egg. | 17:19 |
| ignas | jinty: schooltool.gradebook is the first one to split off | 17:19 |
| jinty | ignas: please do, I've spent too much time on this already... | 17:20 |
| ignas | jinty: later i will work on other packages | 17:20 |
| ignas | jinty: slowly and methodically | 17:20 |
| jinty | I would try installing both at the same time then | 17:20 |
| ignas | schooltool depends on schooltool.gradebook | 17:20 |
| jinty | to see any packaging conflicts | 17:20 |
| jinty | that'll be fun then | 17:20 |
| ignas | hmm we don't have an empty "python-schooltool" package though | 17:21 |
| ignas | and python-schooltool should probably be python-schooltool.the_rest_of_the_stuff | 17:21 |
| ignas | i'll try finding out | 17:22 |
| ignas | what and where and how | 17:22 |
| jinty | ok, I think I'll be off then | 17:22 |
| ignas | bye | 17:22 |
| jinty | let me know of any bugs by mail | 17:23 |
| th1a | jinty: Thanks! | 17:23 |
| ignas | Thanks a lot | 17:23 |
| jinty | ciao | 17:23 |
| *** jinty has quit IRC | 17:23 | |
| th1a | OK, any last words? | 17:23 |
| ignas | when's the deadline? ;) | 17:24 |
| th1a | When's hardy come out -- Thursday? | 17:24 |
| ignas | yep | 17:25 |
| th1a | That's the deadline then. | 17:25 |
| ignas | wish me luck ;) if everything is good - we should have it working today ;) | 17:25 |
| th1a | Excellent. | 17:25 |
| Lumiere | gl | 17:25 |
| th1a | We'll let you work then. | 17:25 |
| * th1a drops the bag of gravel. | 17:26 | |
| Lumiere | I am working on some backup scripts for CanDo | 17:26 |
| Lumiere | Once i get them finished | 17:26 |
| aelkner | i saw you asking about start/stop | 17:26 |
| Lumiere | and ignas finishes up his work | 17:26 |
| aelkner | did you get the answer you needed? | 17:26 |
| Lumiere | not yet | 17:26 |
| aelkner | lke how to stop, right? | 17:26 |
| Lumiere | mgall and I were ships passing in the night | 17:26 |
| Lumiere | over the weekend | 17:26 |
| Lumiere | I owe him stuff | 17:26 |
| Lumiere | and he owes me stuff | 17:27 |
| Lumiere | and we can't quite find each other at times where we can get the stuff | 17:27 |
| aelkner | he's with my brother every day in class | 17:27 |
| aelkner | and there's also a open class where jelkner doesn't even teach that matt attends | 17:27 |
| Lumiere | yes at times where I am in the shower most days | 17:27 |
| Lumiere | yes... we'll work it out | 17:28 |
| aelkner | cool | 17:28 |
| *** wbrady has joined #schooltool | 19:33 | |
| *** pcardune has joined #schooltool | 19:49 | |
| pcardune | hey ignas | 19:49 |
| ignas | hi | 19:49 |
| pcardune | ignas: you know that ascii art table help function you had? | 19:50 |
| ignas | yep | 19:50 |
| ignas | mgedmin is the author | 19:50 |
| pcardune | any plans on sticking that into some helpful package on the zope repo? | 19:50 |
| ignas | ask him ;) | 19:50 |
| pcardune | mgedmin: ? | 19:50 |
| ignas | the drawing was done by him | 19:50 |
| mgedmin | where would you suggest? | 19:51 |
| ignas | the parsing of the table is very very schooltool.lyceum.journal specific | 19:51 |
| pcardune | mgedmin: maybe somewhere like zope.testbrowser, or z3c.etestbrowser | 19:51 |
| pcardune | probably the latter | 19:52 |
| ignas | and was done by me | 19:52 |
| pcardune | ignas: hmm, maybe I will refactor that to work for html tables in general... | 19:52 |
| ignas | you can try doing that | 19:52 |
| * mgedmin dreams about >>> print browser.render_to_ascii() | 19:52 | |
| ignas | you can get a table element quite easily using xpath and just pass that to some "draw table" code then i guess | 19:53 |
| mgedmin | well, ascii is silly, render_to_text() perhaps | 19:53 |
| ignas | mgedmin: if that comes with ablity to select subsets of the page using xpath ;) | 19:53 |
| pcardune | that would be very cool | 19:54 |
| mgedmin | xpath may be a problem: a dependency on lxml (or whatever) | 19:54 |
| pcardune | can't we somehow hook into lynx rendering? | 19:54 |
| mgedmin | lynx, links, or w3m? | 19:54 |
| ignas | hmm, what lynx does with non-complete xtml | 19:54 |
| pcardune | mgedmin: z3c.etestbrowser already has an lxml dependency (that's why it was made in fact) | 19:54 |
| ignas | pcardune: lynx foo.html -dump works even if there are no body tags around the table | 19:59 |
| pcardune | ignas: oooo, what does it look like? | 19:59 |
| ignas | cat foo.html | lynx -stdin -dump | 20:00 |
| ignas | Lala blas asdfds | 20:00 |
| ignas | asfdfdsa fdsafdsa dsagfg | 20:00 |
| ignas | 12332 567 55 | 20:00 |
| ignas | aaa [_] vvv | 20:00 |
| ignas | so you can actually use popen or something like it | 20:00 |
| ignas | and just push the input to lynx | 20:00 |
| ignas | will work on linux | 20:00 |
| ignas | ;) | 20:00 |
| pcardune | is there a python integration package?! | 20:01 |
| ignas | don't think so | 20:01 |
| mgedmin | subprocess.communicate rather than popen | 20:01 |
| pcardune | cool | 20:03 |
| ignas | pcardune: anyway did you know that we have Zope3.4 KGS in launchpad PPA | 20:06 |
| ignas | ? | 20:06 |
| mgedmin | ignas: does anyone know? | 20:12 |
| ignas | mgedmin: nope ;) | 20:12 |
| mgedmin | announce it in #zope3-dev/mailing list | 20:12 |
| ignas | i'd rather start slow ;) | 20:12 |
| *** mgedmin has quit IRC | 20:14 | |
| pcardune | ignas: whoa... | 20:19 |
| pcardune | ignas: that sounds pretty crazy | 20:19 |
| ignas | eh ;) | 20:20 |
| ignas | schooltool tests are passing without any dependencies being downloaded | 20:20 |
| *** wbrady has quit IRC | 20:20 | |
| ignas | so it works at least for some value of the word "works" ;) | 20:20 |
| *** jinty has joined #schooltool | 20:50 | |
| ignas | jinty: woo-hoo apt-get install python-schooltool gives make-schooltool-instance and start-schooltool-instance work as designed, "python2.4 /usr/bin/paster" with proper parameters work as well | 20:53 |
| jinty | cool! | 20:54 |
| jinty | now just try apt-get install schooltool-2007 | 20:54 |
| jinty | .... | 20:54 |
| ignas | your init.d scripts are trying to run paster using python2.5 | 20:54 |
| ignas | which does not work | 20:54 |
| jinty | ah, my init scripts run the binary /usr/bin/paster | 20:55 |
| jinty | which seems like an intelligent thing to do | 20:55 |
| jinty | what should I be doing instead? | 20:55 |
| ignas | zodb is only available on python2.4 | 20:55 |
| ignas | /usr/bin/python2.4 /usr/bin/paster ... | 20:55 |
| jinty | does 3.8 work with 2.5? | 20:55 |
| ignas | yes it does, but zope does not | 20:56 |
| ignas | we have zope packages that think they do work on python2.5 and python2.4 though | 20:56 |
| jinty | why not | 20:56 |
| jinty | I thought it was only RestrictedPython | 20:56 |
| ignas | they are not done with porting it yet ... | 20:56 |
| jinty | which was the problem | 20:56 |
| jinty | and we don't depend on that, do we? | 20:56 |
| ignas | zope3 still depends on it i think | 20:56 |
| ignas | packages we depend on depend on restrictedpython | 20:57 |
| jinty | but does schooltool? | 20:57 |
| jinty | that sucks, badly | 20:57 |
| jinty | hmm, how bad is that dependency? | 20:57 |
| ignas | no idea, really | 20:58 |
| jinty | could you do apt-cache rdepends python-restrictedpython | 20:58 |
| ignas | hmm | 20:58 |
| ignas | nothing | 20:58 |
| ignas | so it might be working with python2.5 then | 20:58 |
| ignas | but we have a broken ZODB package then | 20:58 |
| jinty | er, yep | 20:58 |
| jinty | yes, let me change the metadata a bit | 20:59 |
| jinty | (after I check the ZODB changelog to be sure) | 20:59 |
| jinty | our python-restrictedpython package is probably aslo broken | 20:59 |
| ignas | yes, it says it works on python2.5 | 21:00 |
| jinty | because it's claiming to work on 2.5, but doesn't | 21:00 |
| ignas | i have releases schooltool_2008.0.0 for now by the way, because I was pretty sure it won't work the first time | 21:00 |
| ignas | as we have schoolyool_2008.0.4 now - it seems that i was right ;) | 21:01 |
| ignas | jinty: are you coming to Europython this year? | 21:02 |
| ignas | because I owe you a lot of lunch ;) | 21:02 |
| jinty | ignas: unfortunately I don't think I'll make it | 21:03 |
| jinty | really, really busy | 21:03 |
| ignas | i see | 21:03 |
| * jinty tries to find the log entry proving that the ZODB works with 2.5 | 21:04 | |
| ignas | jinty: i have asked in zope3-dev and srichter seems to be using zodb with python2.5 ... | 21:08 |
| jinty | ignas: http://mail.zope.org/pipermail/zodb-dev/2007-June/011094.html | 21:09 |
| ignas | jinty: cool | 21:09 |
| jinty | ignas: uploaded new zodb packages | 21:15 |
| jinty | will restrict restricted python now | 21:15 |
| ignas | jinty: can I assume that if i did apt-get source, have changed setup.py, did dch -i, and reuploaded the new package that vanguardistas.builder will not lose that small fix as long as the patch applies cleanly to new tarballs? | 21:18 |
| ignas | or will i have to do all that setup.py fixing for every new release | 21:18 |
| jinty | yep, you are assuming correctly | 21:18 |
| jinty | but it's much better to just fix it in svn.zope.org | 21:19 |
| jinty | make a new release | 21:19 |
| jinty | patching the original source code is just storing up problems | 21:19 |
| jinty | because at some point the patch won't apply cleanly | 21:20 |
| jinty | and someone will have to clean it up | 21:20 |
| ignas | i am talking about schooltool releases | 21:20 |
| ignas | PILwoTk and Reportlab eggs do not exist | 21:20 |
| ignas | in ubuntu | 21:21 |
| ignas | and when schooltool is loaded through the egg machinery you get missing dependencies in python | 21:21 |
| jinty | I thought that the pydebdep name translation took care of this? | 21:21 |
| ignas | it takes care on debian level | 21:21 |
| ignas | not on python level | 21:21 |
| ignas | so if some egg states that it depends on PILwoTk in setup.py | 21:21 |
| ignas | setuptools will try and find that egg | 21:22 |
| ignas | if you do pkg_resources.require("schooltool") | 21:22 |
| ignas | instead of import schooltool | 21:22 |
| jinty | ah, well, I'd report a lot of bugs somewhere... | 21:22 |
| ignas | and Reportlab is not an egg on ubuntu | 21:22 |
| ignas | well - PILwoTk is a hacky egg for buildouts, because else you'd need to have tk libraries to build your zope application that uses PIL | 21:23 |
| ignas | so either I put that dependency in every and each buildout.cfg | 21:23 |
| ignas | or in schooltool/setup.py | 21:23 |
| ignas | either way it's not too good, but at least just patching setup.py is a single point fix, that I personaly can maintain | 21:24 |
| jinty | ok, sounds like your situation is bad enough that this is your only way out... | 21:25 |
| jinty | that sucks that reportlab is not an egg in ubuntu | 21:25 |
| jinty | guess you will have to add the dependencies by hand | 21:26 |
| jinty | but I seem to remember that we had other eggs with this problem... | 21:26 |
| ignas | what i have found interesting is that PIL dependency got handled properly | 21:28 |
| ignas | and schooltool dependency on reportlab is a weak one | 21:28 |
| ignas | as in reportlab enhances schooltool | 21:28 |
| ignas | ok, i'll run home now, and write you about my results from there | 21:30 |
| ignas | hopefully zodb will get rebuilt by that time | 21:30 |
| ignas | thanks for all the help and support | 21:31 |
| ignas | byew | 21:31 |
| *** ignas has quit IRC | 21:31 | |
| Lumiere | schooltool doesn't depend on reportlab | 21:38 |
| Lumiere | it suggests reportlab | 21:38 |
| Lumiere | I believe would be the right way | 21:39 |
| *** pcardune has left #schooltool | 21:39 | |
| *** alga has quit IRC | 22:11 | |
| *** ignas has joined #schooltool | 22:20 | |
| ignas | jinty: found who needs restrictedpython | 22:23 |
| jinty | ? | 22:24 |
| ignas | zope.security has | 22:24 |
| ignas | [untrustedpython] | 22:24 |
| ignas | RestrictedPython | 22:24 |
| ignas | in it's requires.txt | 22:24 |
| ignas | so if you try loading it as an egg | 22:24 |
| ignas | it tries to load restrictedpython | 22:24 |
| ignas | same was with python-zope.component i think, it has zope.hookable dependency under | 22:25 |
| ignas | [hook] | 22:25 |
| ignas | zope.hookable | 22:25 |
| ignas | so your dependency grokker skips them | 22:25 |
| ignas | but because of some reason pkg_resources tries to load them | 22:26 |
| jinty | extras_require = dict( | 22:26 |
| jinty | untrustedpython=["RestrictedPython"] | 22:26 |
| jinty | ), | 22:26 |
| jinty | in the setup.py of zope.security | 22:27 |
| jinty | er, so how can we just do a NORMAL import | 22:27 |
| ignas | yeah, that | 22:27 |
| jinty | of schooltool | 22:27 |
| jinty | rather than trying some kind of: let's import but not really | 22:27 |
| jinty | and do crazy stuff? | 22:28 |
| ignas | we can do a normal import and i think it will work, but wsgi stuff at least paste stuff is based on egg entry points, actually schooltool scripts are based on that too | 22:28 |
| ignas | i'll have to read on setuptools i guess | 22:28 |
| ignas | because if it is an extra | 22:28 |
| ignas | there must be a way to not load it | 22:28 |
| jinty | it shouldn't be a problem if it is not there | 22:28 |
| ignas | i mean - what kind of extra is it otherwise | 22:28 |
| jinty | what exactly is your error? | 22:28 |
| ignas | <class 'pkg_resources.DistributionNotFound'>: RestrictedPython | 22:29 |
| jinty | ah, so RestrictedPython then is not on the python path? | 22:29 |
| jinty | can you give me the full traceback? | 22:29 |
| ignas | give me a sec | 22:29 |
| ignas | reading documentation | 22:29 |
| ignas | These requirements will not be automatically installed unless another package depends on them (directly or indirectly) by including the desired "extras" in square brackets after the associated project name. | 22:30 |
| * jinty hates automatically | 22:30 | |
| ignas | zope.pagetemplate module | 22:31 |
| ignas | depends on zope.security [untrustedpython] | 22:31 |
| ignas | and zope.component [hook] | 22:32 |
| ignas | so i have fixed zope.component to depend on it's "hook" parts, but zope.security dependency on restricted python was left outside ... | 22:32 |
| ignas | but apparently zope.pagetemplate depends on zope.security + RestrictedPython | 22:32 |
| ignas | zpoe.app.pagetemplate | 22:33 |
| jinty | zope.app.pagetemplate... | 22:36 |
| jinty | how strong is that dependency... | 22:36 |
| jinty | I would guess it's not even necessary, because no-one uses page templates stored in the ZODB | 22:36 |
| jinty | or at least definitely not schooltool | 22:36 |
| jinty | but sigh, so because of this schooltool doesn't work on 2.5... | 22:37 |
| jinty | jinty@etch:~/src/zope.security/zope.app.pagetemplate # grep -r RestrictedPy . | 22:38 |
| jinty | ? | 22:38 |
| jinty | ? | 22:38 |
| jinty | ? | 22:38 |
| ignas | it seems quite strong :/ zope.app.pagetemplate:engine.py | 22:38 |
| ignas | no no | 22:38 |
| ignas | zope.app.pagetemplate | 22:39 |
| ignas | depends on parts of zope.security | 22:39 |
| ignas | that depend on RestrictedPython | 22:39 |
| jinty | eeew | 22:39 |
| ignas | zope.secirity.untrustedpython | 22:39 |
| ignas | and yes it seems the only reason for schooltool not working on python2.5 :/ there are no other modules except zope.app.pagetemplate that would need that part of zope.security | 22:40 |
| ignas | and only zope.security depends on RestrictedPython | 22:41 |
| jinty | src/zope/app/pagetemplate/viewpagetemplatefile.py | 22:41 |
| jinty | appears only to use the trusted engine | 22:41 |
| jinty | so, restricted python is a dependency of schooltool, but not actually used anywhere... | 22:43 |
| ignas | we use this - ViewPageTemplateFile a lot ... | 22:43 |
| ignas | and it is defined in there | 22:43 |
| jinty | yes, but that uses the trusted engine | 22:43 |
| ignas | oh | 22:43 |
| ignas | hmm, you can try hacking around it | 22:44 |
| th1a | jinty: If you find the time to come to Europython, I'll pay for it. | 22:44 |
| ignas | just set restrictedpython to think that it works on python2.4 | 22:44 |
| ignas | though that is just ugly :/ | 22:44 |
| ignas | but if what you say is true - schooltool tests should probably pass | 22:45 |
| ignas | not sure though :/ | 22:45 |
| jinty | th1a: it's not really a question of money, but I'll speak to the other partners | 22:45 |
| jinty | ignas: src/zope/app/pagetemplate/engine.py look at the parallel construction of | 22:46 |
| jinty | AppPT | 22:46 |
| ignas | yeah, i can see it | 22:46 |
| jinty | and TrustedAppPT | 22:46 |
| ignas | AppPT is not used in any of the eggs schooltool depends on | 22:46 |
| jinty | it's nasty to hack around, I think | 22:46 |
| ignas | i know | 22:47 |
| ignas | so either this or fixing schooltool init.d scripts to use python2.4 to run /usr/bin/paster | 22:47 |
| jinty | unfortunately I think <browser:page also get's it's template implementation from there | 22:47 |
| ignas | that module is kind of vital | 22:47 |
| ignas | the Untrusted part is not though | 22:48 |
| jinty | I'd vote for fixing the init scripts | 22:48 |
| jinty | ... | 22:48 |
| jinty | I'm surprised no-one has cared about this yet | 22:48 |
| jinty | that pagetemplate code is a mess anyway... | 22:49 |
| jinty | really difficult to understand | 22:49 |
| jinty | ignas: PASTER=/usr/bin/python2.4 /usr/bin/paster | 22:50 |
| jinty | try add that to: | 22:50 |
| ignas | /usr/share/schooltool-common/init.d-script | 22:51 |
| jinty | yes | 22:51 |
| * jinty was trying to execute the /usr IRC command | 22:51 | |
| jinty | ... | 22:51 |
| ignas | * Starting SchoolTool 2007... /usr/share/schooltool-common/init.d-script: 52: serve: not found | 22:52 |
| jinty | I've made the change here, just want to see if any other fixes are required | 22:52 |
| jinty | ah, I see | 22:53 |
| jinty | PASTER="/usr/bin/python2.4 /usr/bin/paster" | 22:53 |
| ignas | test: 17: /usr/bin/paster: unexpected operator | 22:54 |
| * jinty bows to the gods of the shell | 22:55 | |
| jinty | and does: | 22:55 |
| jinty | PYTHON=/usr/bin/python2.4 | 22:55 |
| jinty | PASTER=/usr/bin/paster | 22:55 |
| ignas | schooltool: error opening file /etc/schooltool/schooltool-2007/schooltool.conf: [Errno 2] No such file or directory: '/etc/schooltool/schooltool-2007/schooltool.conf' | 22:55 |
| ignas | it's main.conf in there | 22:55 |
| jinty | ok, I fixed the template | 22:56 |
| jinty | [app:main] | 22:56 |
| jinty | use = egg:schooltool#main | 22:56 |
| jinty | config_file=main.conf | 22:56 |
| jinty | anything else? | 22:56 |
| ignas | it works it seems | 22:57 |
| jinty | wow | 22:57 |
| ignas | now i will have to fix the schooltool.stapp2008spring package | 22:58 |
| ignas | and we can have schooltool-2008spring | 22:58 |
| ignas | or schooltool-2008.04 | 22:58 |
| ignas | don't know how to call it | 22:58 |
| jinty | it should be pretty easy to make that | 22:58 |
| * ignas did that | 22:58 | |
| jinty | cool | 22:58 |
| ignas | and botched the namespace_packages | 22:58 |
| ignas | again ;) | 22:58 |
| jinty | Ok, I'll upload a new schooltool-common | 22:58 |
| jinty | gonna need to trigger re-builds for schooltool-2007 and schooltool-stapp2008spring | 22:59 |
| ignas | i'll re release schooltool.stapp2008spring and schooltool.lyceum.journal | 23:00 |
| jinty | ok, uploaded new schooltool-common | 23:04 |
| jinty | er, so, we're done? | 23:04 |
| ignas | hope so | 23:05 |
| ignas | :) | 23:05 |
| ignas | we'll now in 40 minutes, which is tomorrow ;) | 23:05 |
| th1a | Are these hardy packages? | 23:06 |
| jinty | th1a: no, but there is a small script you can write to re-upload them for hardy | 23:07 |
| th1a | OK. | 23:07 |
| Lumiere | that was a fun read | 23:07 |
| jinty | I'm pretty sure ignas knows how to write that script by now | 23:08 |
| ignas | you mean - download, patch the changes file, reupload? | 23:09 |
| ignas | should the changes file have gutsy or "gutsy, hardy"? | 23:10 |
| jinty | some combination of the arguments to dch should do the right patching | 23:11 |
| jinty | dch -i -r hardy "Re-build for hardy" | 23:12 |
| jinty | or something | 23:12 |
| ignas | i see | 23:12 |
| ignas | i'll look through the list and skip the packages that we are not releasing i guess | 23:12 |
| jinty | but, yes, apt-get source package, cd package, dch, dpkg-buildpacakge and then re-upload | 23:12 |
| jinty | you can even get the list of possible targets from the releases file | 23:13 |
| jinty | http://ppa.launchpad.net/schooltool-owners/ubuntu/dists/gutsy/main/binary-i386/Packages | 23:13 |
| jinty | Packages file, I mean | 23:14 |
| ignas | what i would like to find out is which packages are old style ones and don't have auto dependencies set properly. Even though we can release it the way it is, i'd still like to have it clean. | 23:15 |
| ignas | i know i should rebuild zc.i18n | 23:16 |
| ignas | jinty: error: (4, 'Interrupted system call')<br /> :/ | 23:24 |
| ignas | bzr does not play with builder that well | 23:24 |
| ignas | but the second try works out most of the time | 23:25 |
| ignas | jinty: the server started up properly so schooltool-common is fixed. But schooltool-2007 is still using schooltool.conf, not main.conf in paste.ini | 23:47 |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!