IRC log of #schooltool for Tuesday, 2005-09-20

jintytvon: also waiting for the meeting?00:08
tvonjinty: meeting? today?00:19
jintyaargh, forget about wrong timezone, think wrong day00:19
tvonheh :)00:20
jintyI think it's time for bed00:20
*** jinty has quit IRC00:20
tvontalk to you tomorrow ;)00:22
povbot/svn/commits: * srichter committed revision 5020:00:38
povbot/svn/commits: Add viewlet dependency.00:38
th1asrichter: ayt?01:08
srichterth1a: yeah01:12
srichterth1a: btw, you realize that the testbrowser recorder from Brian is in a Zope 3 sandbox these days, right? :-)01:13
th1aI'm trying an example from Philipp's book, so perhaps I should be bugging him abou this ;-)01:13
th1aam unaware of the testbrowser recorder.01:13
th1aAnyhow, did IPluggableAuthenticationService cease to exist?01:14
srichterit is replaced with
srichterthe old code is still in pluggableauth, but should not be used anymore01:15
th1aHrm.  Well, it doesn't appear to be in the rc of 3.1 I'm using.01:16
srichterthat would be bad!01:17
srichterplease report it on the mailing list01:17
th1aWell, I should probably get the latest RC and make sure I'm not smoking dope.01:17
*** th1a is now known as th1a|dinner01:21
*** bskahan has joined #schooltool02:13
*** tvon has quit IRC02:23
*** tvon has joined #schooltool02:24
*** Ricey has joined #schooltool02:35
*** Ricey_ has quit IRC02:35
*** bskahan has quit IRC02:56
*** bskahan has joined #schooltool03:14 is now up03:16
bskahanth1a|dinner: I'd put the schoolbell stuff under the schooltool directory in ftp03:16
bskahanschooltool at that level being the "project"03:17
bskahanschoolbell tarballs would just be in
*** bskahan has quit IRC03:37
*** th1a|dinner is now known as th1a03:48
*** jinty has joined #schooltool04:10
*** bskahan has joined #schooltool05:20
srichterman, there are many links in the calendar code that are not reachable from the UI06:00
bskahansrichter: what links?06:02
*** jinty has quit IRC06:03
srichterlike calendar.vfb06:44
srichteralso daily.html, weekly.html, etc are not directly available06:44
bskahando we need daily/weekly/etc?06:46
bskahanwe have the day/week/etc view links at the top06:46
bskahanthat use the date in the url06:46
*** bskahan has quit IRC06:56
*** mgedmin has joined #schooltool10:14
*** alga has joined #SchoolTool10:36
*** ignas has quit IRC11:01
* mgedmin wonders what's happening with demo.schooltool.org11:05
*** ignas has joined #schooltool11:28
*** Aiste has quit IRC11:59
*** Aiste has joined #schooltool12:00
*** Aiste has quit IRC12:03
*** Aiste has joined #schooltool12:17
*** th1a has quit IRC12:17
povbot/svn/commits: * mg committed revision 5021:12:25
povbot/svn/commits: Merged optimisations from the alga-sb-optimization branch.12:25
povbot/svn/commits: - Significantly improved the speed of ACL and calendar views.12:25
povbot/svn/commits: - Added some ad-hoc benchmarks for measuring rendering time.12:25
povbot/svn/commits: - Hopefully fixed issue 377 (intermittent ConnectionStateError).12:25
povbot/svn/commits: - Partially fixed issue 379 (PDF calendars always in UTC), although12:25
povbot/svn/commits: end-users will probably not notice (the more visible part of the bug remains unfixed).12:25
povbot/svn/commits: Commands used for merging:12:25
povbot/svn/commits: svn merge -r 4921:4922 svn+ssh:// .  svn merge -r 4923:HEAD svn+ssh:// .12:25
povbot/svn/commits: (HEAD was revision 5020 at the time)12:25
povbot/svn/commits: svn revert src/schoolbell/app/browser/configure.zcml12:25
povbot/svn/commits: (had conflicting changes that added some apidoc directives; not needed in this branch)12:25
povbot/svn/commits: * mg committed revision 5022:12:31
povbot/svn/commits: Fixed SchoolTool unit tests after the changes applied by the optimization branch merge broke them.12:31
mgedminmerge complete12:31
mgedminjinty will be happy12:31
mgedminignas, the url is svn+ssh://
*** maharaja has joined #schooltool12:46
*** mgedmin has quit IRC13:25
*** thisfred has joined #schooltool13:25
*** alga has quit IRC13:42
*** srichter has quit IRC14:04
*** srichter has joined #schooltool15:05
ignassrichter, what might be broken if a test using IUserPreferredLanguages(request).getPreferredLanguages() fails ?15:13
ignasi am geting an empty list instead of ('lt',)15:14
*** bskahan has joined #schooltool15:20
*** jinty has joined #schooltool15:27
povbot/svn/commits: * ignas committed revision 5023:15:38
povbot/svn/commits: Merged revisions 4932 and 4968 (fix for a bug that appeared after 4932).15:38
povbot/svn/commits: Move getAllDayEvents() from CalendarViewBase to DailyCalendarView.15:38
povbot/svn/commits: Erradicate it's use in monthly and weekly templates.  Refactor view templates.15:38
povbot/svn/commits: Commands used for merging:15:38
povbot/svn/commits: cd src/schooltool svn merge -r 4931:4932 svn+ssh:// svn merge -r 4967:4968 svn+ssh://
povbot/svn/commits: resolve conflicts in:15:38
povbot/svn/commits: app/browser/templates/ app/browser/templates/ app/browser/tests/test_cal.py15:38
srichterignas: I would have to look as much into the code as you do15:42
ignassrichter, well, apparently it was something with the way tests were being run :/ it disappeared eventually ...15:44
ignassrichter, got some time ?15:50
*** alga has joined #SchoolTool15:57
*** alga has quit IRC15:57
*** th1a has joined #schooltool16:05
povbot/svn/commits: * ignas committed revision 5024:16:32
povbot/svn/commits: Merged revisions 4935, 4936:16:32
povbot/svn/commits: Added caching to dayEvents function in the CalendarViewBase.  Added a missing unit test.  Removed cpu hogging tal:define from the inner loop in the daily template.  Change the name of the calendar cache attribute in CalendarViewBase to make unit tests look better.16:32
povbot/svn/commits: Commands used for merging:16:32
povbot/svn/commits: cd src/schooltool svn merge -r 4934:4936 svn+ssh://
povbot/svn/commits: * ignas committed revision 5025:16:56
povbot/svn/commits: Merged revision 4926:16:56
povbot/svn/commits: Refactored the recurrence expansion from calendar into events.16:56
povbot/svn/commits: Commands used for merging:16:56
povbot/svn/commits: cd src/schooltool svn merge -r 4925:4926 svn+ssh://
povbot/svn/commits: Still don't know why Stephan skipped it ...16:56
*** tiredbones has quit IRC17:12
srichterignas: I only have time for short questions; on Tuesday and Thursday I am doing TA work17:28
ignasTA ?17:32
ignasoh i see17:32
ignaswell by thursday i will be done with merging17:33
th1atits and ass17:33
*** alga has joined #SchoolTool17:34
ignasoh, i see, must be difficult to do any of typing stuff then ;)17:34
th1aOne handed typing.17:35
SteveAterritorial army17:52
SteveAthe british volunteer military reservists17:52
povbot/svn/commits: * ignas committed revision 5026:17:53
povbot/svn/commits: Merged revisions 4959, 4965, 4969, 4970, 4971, 4972, 5016, 5018.17:53
povbot/svn/commits: Benchmarking tools so we could measure the impact of our optimizations (nearly accurately)17:54
povbot/svn/commits: Commands used for merging:17:54
povbot/svn/commits: svn merge -r 4958:4959 svn+ssh:// svn merge -r 4964:4965 svn+ssh:// svn merge -r 4968:4972 svn+ssh:// svn merge -r 4995:499717:54
povbot/svn/commits: svn+ssh:// svn merge -r 5015:5016 svn+ssh:// svn merge -r 5017:5018 svn+ssh://
povbot/svn/commits: And some fiddling to make these scripts actually work.17:54
*** alga has quit IRC17:58
srichterignas: so what was your question?18:14
ignaswhy have you skipped revision 4926 ?18:17
ignaswas there a reason i don't know ?18:17
srichtermaybe I just forgot18:35
ignasstrange, it seems that you worked around it ...18:36
ignaswell, i have merged it now anyway ...18:37
*** tvon has quit IRC18:40
srichteras long as the tests pass :-)18:45
ignaswell, imho it would be kind of nice if you reviewed my checkins18:46
ignasi might have missed some your branch specific things18:46
ignaslike - wrong way of doing test setup and the likes18:46
srichterI usually look at all the checkins18:47
*** tvon has joined #schooltool18:52
povbot/svn/commits: * ignas committed revision 5027:18:53
povbot/svn/commits: Merged revisions 4947 and 4952 to 4958:18:53
povbot/svn/commits: ACL view refactoring and optimization.18:53
povbot/svn/commits: Removed an unused import from18:53
povbot/svn/commits: src/schooltool/app/rest/relationships.py18:53
povbot/svn/commits: Commands used for merging:18:53
povbot/svn/commits: cd src/schooltool svn merge -r 4946:4947 svn+ssh:// svn merge -r 4951:4958 svn+ssh://
povbot/svn/commits: * ignas committed revision 5028:19:04
povbot/svn/commits: Merged revision 4937:19:04
povbot/svn/commits: Optimized weekly and monthly recurrence rules to fast-forward to the interesting time period instead of expanding all recurrences from the very start.19:04
povbot/svn/commits: Commands used for merging:19:04
povbot/svn/commits: cd src/schooltool svn merge -r 4936:4937 svn+ssh://
povbot/svn/commits: * ignas committed revision 5029:19:16
povbot/svn/commits: Merged revision 4966:19:16
povbot/svn/commits: ACL view optimisation: speed it up roughly 10 times.19:16
povbot/svn/commits: Previously the ACL view computed the set of available permissions for all principals, and then applied batching to show only a subset.  The computation was expensive.  Now the ACL view first does batching, and only then performs the computation.  End result: ACL view rendering time went from 3 to 0.3 seconds, and form handling went from 5 to 0.5 seconds on my laptop (Pentium M, 1.6 GHz).19:16
povbot/svn/commits: Commands used for merging:19:16
povbot/svn/commits: cd src/schooltool svn merge -r 4965:4966 svn+ssh://
povbot/svn/commits: * ignas committed revision 5030:19:48
povbot/svn/commits: Merged revisions 4974, 4975, 4978, 5017:19:48
povbot/svn/commits: Added a function that pigeonholes CalendarDay objects into time intervals.19:48
povbot/svn/commits: Optimized CalendarViewBase.getYear so it would pass a day cache to the monthly view thus saving us 11 getDays calls.  Benchmarks indicate a 2x speedup.19:48
povbot/svn/commits: Added "smart" caching for CalendarViewBase.getDays.  ~10% performance gain for nonrecurrent event benchmark.  A view knows which time range is of interest, so it can call calendar.expand only once for this range, instead of calling calendar.expand five times for smaller, overlapping time ranges.19:49
povbot/svn/commits: Marius thinks that local caches in getMonth/getYear, and perhaps in dayEvents also, are now not necessary and could be removed.19:49
povbot/svn/commits: Commands used for merging:19:49
povbot/svn/commits: cd src/schooltool svn merge -r 4973:4975 svn+ssh:// svn merge -r 4977:4978 svn+ssh:// svn merge -r 5016:5017 svn+ssh://
*** pips has joined #schooltool19:57
*** mgedmin has joined #schooltool20:00
*** pips has left #schooltool20:07
*** pcardune has joined #schooltool20:20
pcardunehello all, I am just now trying out the new CanDo svn server... what is my password?20:22
povbot/svn/commits: * ignas committed revision 5031:20:26
povbot/svn/commits: Merged revisions 5006, 5007 and 5022:20:26
povbot/svn/commits: Timezone cleanups in schooltool.calendar:20:26
povbot/svn/commits: - Require `first` and `last` arguments to expand to have explicit timezones20:26
povbot/svn/commits: (instead of falling back to UTC).20:26
povbot/svn/commits: - Do not insist that `first` and `last` have the same timezone.20:26
povbot/svn/commits: - Changed parse_datetimetz to accept ISO 8601 dates in UTC (i.e. with a 'Z' at20:26
povbot/svn/commits: the end).20:26
povbot/svn/commits: - Moved argument checks from CalendarEventMixin.expand to CalendarMixin.expand20:26
povbot/svn/commits: (optimisation; about 10-15% speedup in benchmark/
povbot/svn/commits: - Removed timezones and all-day events from "Future goals" section in the README20:26
povbot/svn/commits: (they're implemented, although I have some suspicions as to the correctness of the implementation).20:26
povbot/svn/commits: - Fixed a corner case description in the CalendarMixin.expand doctest.20:26
povbot/svn/commits: Partial bugfix for issue379: PDF calendar views always use UTC.20:26
povbot/svn/commits: Fixed SchoolTool unit tests after the changes applied by the optimization branch merge broke them.20:26
povbot/svn/commits: Commands used for merging:20:26
povbot/svn/commits: cd src/schooltool svn merge -r 5005:5007 svn+ssh://
povbot/svn/commits: Applied 5022 manualy.20:26
pcardunetvon, ayt?20:39
tvonpcardune: hey20:39
pcarduneI'm just now trying out the svn server... what is my password?20:39
tvonwith your ssh keym it should just ask you for your ssh passphrase20:40
tvonbskahan: ayt?20:40
pcarduneis there a particular way to tell svn to use my ssh key? because at the moment it asks for the password of pcardune@
tvonpcardune: is it asking for your password or a key passphrase?20:42
tvonit should figure it out20:42
th1asrichter:  I posted my problem with IPluggableAuthenticationService to the Zope3-dev list as you suggested.20:43
tvonpcardune: you are on the same computer that has your ssh info on it?20:44
tvonpcardune: erm, you don;t seem to have a user on the box20:45
pcardunethat would do it... I will resend my ssh key to bskahan20:46
povbot/svn/commits: * ignas committed revision 5032:20:53
povbot/svn/commits: Merged revisions 4949, 4950, 4951, 4979, 4993:20:53
povbot/svn/commits: Removing superfluous ftest -- this should fix issue377 (ConnectionStateError).  Removing unused imports.  Refactored the hasPermission implementation Removed dead code from CalendarViewBase.getEvents. (The variable was used before this function has been refactored) Clarified a comment.20:53
povbot/svn/commits: Commands used for merging:20:53
povbot/svn/commits: cd src/schooltool svn merge -r 4948:4951 svn+ssh:// svn merge -r 4978:4979 svn+ssh:// svn merge -r 4992:4993 svn+ssh://
*** thisfred has quit IRC20:56
ignasth1a, ping20:59
*** pips1 has joined #schooltool21:02
*** pips1 has left #schooltool21:04
bskahanpcardune, hey can you resend your ssh key as an attachment, it got linebroken when it was inlined21:17
ignascould you check out the most up to date trunk and use it a bit ?21:17
ignasi have just merged all the optimisation changes ...21:18
*** pips1 has joined #schooltool21:18
ignasoh, and Marius merged them into the release branch ...21:18
th1aWhich should I try?21:18
pips1hi all21:19
ignasth1a, well ... any ;)21:20
th1aHi pips.21:20
pips1hi th1a21:20
pips1jinty, ping!21:21
pips1is jinty around?21:28
pips1I understand he is preparing another bug fix release...21:29
pips1I'd like to contribute a translation21:30
pips1but maybe you can answer my questions too! :-)21:31
th1aVia Rosetta?21:31
pips1I was going to use a local app like kbabel...21:31
pips1I exported the po file from rosetta.21:32
pips1and I was planning to import it after i'm done.21:32
pips1is that a good way of doing it?21:33
th1aNote that as a monolingual American, my understanding of i18n is weak.21:33
th1aBut yes, that should work.21:33
pips1(btw, I am having a terrible time with my keyboard at the moment and that's why I am typing so slow. I am running breezy-preview and it has a bug with keyboard layout. ugh.)21:34
pips1I was wondering whether jinty might be able to include my translation into the bugfix release if I hurry up :-)21:39
th1aignas: python -w -s src/schooltool21:39
th1aTraceback (most recent call last):21:39
th1a  File "", line 31, in ?21:39
th1a    from schooltool.testing import test21:39
th1a  File "/Users/hoffman/Desktop/schooltool/src/schooltool/testing/", line 244, in ?21:39
th1a    DocFileCase_classes = (doctest.DocFileCase,)21:39
th1aAttributeError: 'module' object has no attribute 'DocFileCase'21:39
th1amake: *** [test] Error 121:39
th1apips1:  We are planning on updating the translations, so I think the answer is yes, if you can get it done in the next couple days.  However, we haven't discussed the exact timeline for the release other than ASAP.21:40
* pips1 will send jinty an email.21:41
pips1cu l8r21:41
*** pips1 has left #schooltool21:41
th1aignas:  otoh, the server seems to start.21:42
*** ignas has quit IRC21:45
*** pips1 has joined #schooltool21:49
pips1I forgot to ask: do I need to translate both schoolbell and schooltool po files?21:50
* pips1 is still confused about the two packages21:50
th1aI'm not sure.21:51
th1aignas:  Much better!21:53
mgedminpips1, I think the answer is yes21:53
mgedminschooltool.po definitely contains strings that aren't in schoolbell.po21:53
mgedminI'm not sure if there are many overlapping strings21:54
mgedminth1a, I wrote that line of code21:54
th1aWhich line?21:54
mgedminI think I know how to fix it21:54
mgedmin DocFileCase_classes = (doctest.DocFileCase,)21:54
mgedminquick fix: change the if above to check whether hasattr(doctest, 'DocFileCase')22:04
mgedminif there isn't an if, add one22:04
* mgedmin busy firefighting a server that hasn't come up22:04
pips1mgedmin: ok. another question: in rosetta, there are already several (partially) translated po files. In s.bell, there are de, de_DE, de_CH. In s.tool, there is only 'de'. How do I add/get de_CH for s.tool?22:05
* mgedmin knows not22:07
jintyhoi pips122:07
th1aThat's a Rosetta question?22:07
pips1hoi jinty :-)22:08
th1aYou can ask on hm... #launchpad?  Perhaps there's a rosetta channel?  I thought there was...22:09
pips1jinty, do you know how it works with the po files for the various dialects of a language?22:09
jintyIf you import the translation into rosetta, I can use it with the 0.11.2 debian/ubuntu packages.22:10
jintyI think you use the locale22:10
jintyi.e. es22:10
jintyi.e. es_AR22:10
jintyso es contains the most neutral translations22:10
jintyes_AR contains argentinian specific stuff22:11
pips1hm I guess I'll just go for 'de' rather than 'de_CH' to start with...22:11
jintyyep, normally it is better to get a good de before de_CH22:12
pips1'de' is my mother tongue, 'de_CH' is what I spoke with my local chums :-)22:13
pips1ok, I'll start with 'de'22:13
jintyif you set the locale to de_CH, it'll first try de_CH, then use de as a fallback22:13
jintyIf i remember correctly22:14
jintywill you be using the packages?22:14
jintyok, then ping me when you are done...22:15
*** ignas has joined #schooltool22:17
jintyanyone know anything about/has used sqlos?22:21
*** jelkner has joined #schooltool22:21
jelknerjinty: hi brian22:21
jelkneri just sent you an email22:21
jintyyeah, got it.22:22
jintybusy sending a very positive reply;)22:23
jelkneri have a question for anyone here22:24
jelknerwe are having *serious* performance issues with CanDo22:24
jelknerit is big time slow22:24
jelkneron the order of taking 35 seconds to display a screen22:24
th1aWell, we just fixed the performace problems with SchoolBell.22:24
th1aBy "we" I mean POV.22:25
jelknerhi tom!22:25
jelknerso we need to get that into cando22:25
th1aSo it is probably similar issues.22:25
jelknerthat will be the first item on the agenda at 4...22:25
jintyI should be relasing a new schooltool bugfix release with these optimizations shortly.22:26
th1aOh, yeah, it is tuesday.22:26
th1aBut CanDo will probably need its own set of optimizations.22:26
th1aYou're going to have to figure out exacly what is causing your poor performance.22:28
jelknerso we can talk about that too22:28
jelkneri want to make sure paul knows how to go about that in a clear, systematic way22:28
th1aIt is a bit late in the day to ask POV about it today, but they're the ones with experience doing profiling and optimization in SchoolBell.22:28
jelkneras long as we can tell paul who to ask, he should be able to take it from there22:29
th1aI think it is safe to say that the problems you're having are solvable.22:29
jelknerme too22:29
jelkneri told dave not to worry about it22:29
th1aI just sent an email to the SchoolTool list outlining the improvements in the upcoming release.22:29
jelknerafter all, good xp practice says "no premature optimization"...22:30
jelknerit is not premature now22:30
jelknerwe know what we need and we have real data to test it with22:30
th1aThese are the (very unscientific) before and after times:22:31
th1a> 2005: 2 seconds22:31
th1anow: .5 seconds22:31
th1a> 2006 10 seconds22:31
th1anow: 1 second22:31
th1a> 2007: 19 seconds22:31
th1anow: .4 seconds22:31
th1a> 2008: 30 seconds22:31
th1anow: .6 seconds22:31
th1a> 2009: 40 seconds22:31
th1a1 second22:31
th1aYear view for 2010:22:31
th1a1 second22:31
th1aMonth view for July 2005:22:31
th1a.75 seconds22:31
th1aThat's for April 12 of each year.22:31
th1aWith a bunch of repeating events.22:31
jelknerwhat do the years mean?22:32
ignasyour computer is fast :)22:32
th1aApril 12 of each year.22:32
th1aThis thing usually runs SchoolTool very slowly.22:32
th1aIt runs the test suites SUPER slowly.22:32
th1aAs you can see, the expansion of repeating events used to be brute force.22:33
ignaswell benchmark of recurrent events (we have benchmarks now) used to run like 15-30 minutes on my laptop22:34
ignasbefore optimization22:34
th1aOh, this isn't your benchmarks.  This is just ab of one page request.22:35
ignasth1a, if you want to compare the old release with the new one you should just copy the benchmarks folder into the old release and run it there ...22:38
ignasand compare the results with benchmarks executed on the optimized version22:38
th1aHow does one run the benchmarks?22:38
jelknerpcardune: hi paul!22:46
jelknerglad you're here22:46
jelkneri've already started asking about a few things, and i really needed your input22:46
pcardunejelkner, hello22:52
jelknerwe were discussing the speed issue22:54
jelknerthe top item on our agenda22:54
jelknerare you familiar with the problem?22:54
pcarduneyes, i saw that... I have a few ideas up my sleeve22:54
pcardunei did some optimization before for different things22:54
pcarduneI guess it would just be useful to know what specific stuff to watch out for...22:55
jelknerthey've recently made some schoolbell changes that dramatically improved performance on that end22:55
jelknerthat might be part of the problem22:55
pcardunefor example, I assume that tales loops run slower than python loops...22:55
pcarduneI don't think so... I think the problem has to do with a lot of relationship look ups22:56
jelknerthla: tom, paul is here22:56
tvonHello all22:56
jelknerwho can he talk to about optimizations for speed?22:56
jelknertvon: hi tom!22:57
pcardunebskahan, ayt?22:57
tvonjelkner: heya22:57
th1ath1a:  Sorry I was distracted by some hot gossip about my old school.22:58
th1aThe relationship lookups are brute force at this point, I think.22:58
*** ffsnoopy has joined #schooltool22:59
th1aSo that would be a good place to do some work.22:59
povbot/svn/commits: * jinty committed revision 5033:22:59
povbot/svn/commits: Use the latest zpkgtools feature to fix loading of resources via subversion over http.22:59
jelknerffsnoopy: hi mitchell!22:59
ffsnoopyjelkner: hiya22:59
pcarduneyeah, it is deffinitly having to do with relationship look ups22:59
th1aThere was a long thread on the Zope3-dev list about optimizing libraries that are similar to our relationship library.23:00
jelknerok, it's 4 pm23:00
jelknerpcardune: do you know where dave is?23:01
pcardunehe is on skype... downloading mirc at the moment23:01
jelknerhe could just use his ubuntu-live cd ;-)23:01
pcardunehe says, "yeah, but I don't want to"23:02
jintyis there an agenda for the meeting somewhere?23:04
jelkneri'm about to post it here23:04
jelkner1. the performance issue and what to do about it23:04
pcardune(I still can't check out from svn... it still asks me for a password)23:04
jelkner2. the general development environment and making sure all is working23:04
*** Welsh has joined #SchoolTool23:05
jelkner3. involving mitchell foral (ffsnoopy) in the development process23:05
Welshwelsh here23:05
jelknerWelsh: hey dave, you made it23:05
WelshmIRC, I'm a believer!23:05
tvonWelsh: hello23:05
jelknerubuntu is better!23:05
WelshAlso believing in Carduner about right now!23:05
jelknerhow is that for an agenda?23:06
tvonjelkner: sounds good23:06
*** pips1 has left #schooltool23:06
jinty4. place to put the tarballs23:07
jelknerlet's start with 1. then: the performance issue23:07
jelknerwe need a way to fix this fast23:07
Welshquick resent of agenda for23:07
ffsnoopyhow about describing it briefly?23:07
ffsnoopyso i have some idea of whats going on23:07
jelknerdave, you want to handle that23:07
bskahanpcardune, try now23:07
jelkner1. the performance issue and what to do about it23:08
jelkner2. the general development environment and making sure all is working23:08
jelkner3. involving mitchell foral (ffsnoopy) in the development process23:08
jelknerwe are on 123:08
WelshPerformance relates to lookups of student spreadsheets23:08
jelknerdo you have any other issues?23:08
Welshand student competency records23:08
jelknerhow bad is it?23:09
WelshPaul says there are a lot of relationship lookups on those two pages.23:09
WelshI'm concerned because the wait is 35 secs.23:09
Welshway, way, way too much23:09
pcardunebskahan, it works, thanks23:09
Welshcould we get it down to something like 10 secs.23:09
Welshthat's a long wait, but at least bearable23:10
th1aWe should be able to do much better than that.23:10
WelshThe good news is that the main spreadsheet lookup is ok23:10
WelshMaybe 5-10secs.23:11
WelshBuy the student spreadsheets also are important23:11
jelknerso, what to we do?23:11
WelshWhat can we do in the near term to improve?23:11
jelknergood question23:12
WelshMy nearterm is next Tuesday, when I demo all this at Virginia Beach Advanced Tech Center23:12
WelshThe nearterm after that is the sooner the better.23:12
jelknerthe main thing is for paul to know who to talk to23:12
jelknerthla: who do you suggest?23:12
pcarduneI have a few things up my sleeve to try beforehand (I'm not at a total dead end yet)23:13
th1aalga is less gumpy than mgedmin.23:13
jelknerpcardune: do you know alga? (less grumpy is better ;-)23:13
jelknerthat's great, but you need a backup plan, someone to turn to... it is crunch time!23:14
jelknerso, do you know alga?23:15
jelknerok, i want to move this forward and not waste folk's time, so...23:15
pcarduneno i don't23:15
pcardunei can hear you23:16
jelknerthla: could you email paul his contact info?23:16
jelknerthanks!  next item...23:16
jelknerwe need to figure out how the early adopters should run CanDo so that developer fixes can be easily incorporated into their systems23:17
jelknerapt-get update is great, but only if we build a new package everytime we fix something23:17
jelkneris that the way to go, or would something else be better?23:18
jintyin the near term?23:19
tvondeb packages are pretty standard for this23:19
jelknerand certainly the easiest from the users point of view23:19
WelshVA Beach would like a list of steps of whatever we come up with23:19
WelshThey've got pretty savvy tech folks over there23:19
jintyfor very early adopters with technical knowledge a subversion checkout is good23:20
jelknerthat works for me, it would work for dave, and it sounds like it would work for va beach23:20
tvonDevelopers should use subversion... the trunk checkout gives them a working setup after a 'make'23:20
th1apcardune:  This thread may be relevant to you:
jintythe next level down is tarballs. nightly builds would be good.23:21
pcardunetvon, my make fails with error: invalid Python installation: unable to open /usr/lib/python2.4/config/Makefile (No such file or directory)23:22
jelknerkeep in mind that we are in a working environment, a school, which is a pressure cooker23:23
jelknerteachers come to dave and say, "this isn't working"23:23
jelknerhe tells paul23:23
jelknerpaul fixes it23:23
jintypcardune: apt-get install python{,2.4,2.3}-dev23:23
jelknerand then dave needs to apply the fix23:23
jelknerwe need this to be *easy*23:23
jelknerthe great thing about jinty's package is that after apt-get install cando the thing is running23:24
jelknerthat's great23:24
jelknerand it runs after a reboot of the machine23:24
*** flint_ has joined #schooltool23:24
pcardunehmm... i think the Breezy upgrade changed some things23:24
jelknerso a power outage won't break it23:24
jelknerit makes the admin's life bareable23:25
jintyjelkner: it is possible to apply patches to a package and re-build, without the effort of re-building the tarball23:26
jelkneron your own machine?23:26
jelkneror do we do this and update the package?23:26
jintyno, to the package, it is stored in the .diff.gz23:27
jelknerand then apt-get update applies it, yes?23:27
jelkneri like that the best23:27
jelknerso if no one objects, the next question is: can we make this easy for paul to do?23:28
jintypcardune: know how to use arch?23:28
pcarduneafraid not23:28
tvonis CanDo going to follow the move to arch/baz?23:29
jelknerCanDo will go where thla and the rest of you folks says it goes23:29
tvonth1a: is ST moving to baz along with the rest of canonical/ubuntu/etc?23:30
* jinty imagines mgedmin becoming even more grumpy23:31
jelknerwell, this is all well and good, but it won't effect the next few weeks, so what do we do now?23:32
jelkneri'm willing to assume the roll of package builder23:32
jelknerif that makes sense23:32
th1atvon:  Eventually.  There is no timeline.23:32
jelknersomeone needs to do it23:32
jelknerand pcardune is busy fixing bugs23:32
tvonjinty: can you get the changes you made for your package into svn?23:32
jelknerjinty: so, does that make sense?23:32
jintytvon: arch makes some things really easy, like keeping track of the package specific changes23:33
flint_we can refer to him as "package-roll elkner" (like the bluesman Jelly Roll Morton)23:34
* tvon hasnt messed with arch/baz in a while23:34
jintyjelkner: I can teach you how to update the packages, yes23:34
jelknergreat, then i'll assume that responsibility23:35
jelkneri'll start by reading the docs you sent me, and then get back to you23:35
jelknerso, one last thing23:35
flint_give a man a package and he can do, teach a man to package and he cando more....23:35
tvonso we need to get the packaging bits into SVN, probably make a branch for our current release and jelkner needs to know how to update packages from that branch23:35
tvonsound right?23:36
jelknerthat sounds great23:36
jelkneri could do an svn update and build the package from that, then just put it where it needs to go23:36
tvonthats the idea23:36
* jinty thinks23:37
jelknerthen users can apt-get update and we are good to go23:37
tvonjelkner: you'll need to get a shell acct setup with bskahan23:37
jelknerhow do i do that?23:37
bskahanjelkner, send me your ssh public key23:37
jintythat can work23:38
*** bskahan has quit IRC23:39
th1aI have to go buy food.  Catch you later.23:40
jelknerok, that takes care of item 2 for now23:40
flint_Is Elkner yet another package manager (YAPM)?23:40
*** th1a is now known as th1a|way23:40
jelknerpcardune: what is up with your svn access?23:40
pcarduneit works23:40
jelknerok, item 323:40
jelknermitchell is interested in working with you on this23:41
jelkneri was thinking of asking him to complete eldar's import stuff as a first step23:41
jelknerdoes that make sense to you?23:41
jelknerand can you point him at the existing code?23:41
pcarduneyes, i can do that23:42
jelknerffsnoopy: mitchell, you still here?23:42
tvonif he wants to dev, have him send a ssh key to bskahan23:43
jelknerffsnoopy: do you have a public ssh key you can send to bskahan?23:43
jelknerok, i'll set that up with you tomorrow23:44
Welshcan we make a decision on ports23:44
Welshi.e. the ports SchoolTool/CanDo will run on23:44
ffsnoopyjelkner: sounds good23:44
jelknerWelsh: done23:44
Welshfor the nearterm23:44
jelknerit runs on 728023:44
Welshis that the port I should ask to be opened, along with ww23:45
jelknerbut i can make it run behind apache using port 80 for you23:45
Welsher, 2223:45
jelkneractually, yes23:45
jelknerthat would make things even easier23:45
Welshso, 7280 plus 2223:45
Welshwill try.23:45
Welshdone did23:45
Welshgot paid23:45
Welshhad fun23:45
jelknersounds like we are finished here then23:46
WelshI mean, did for FREE, remember audience23:46
jelknerno, this audience wants to get paid to23:46
Welshyou guys deserve it23:46
jelknerwe want freedom, not poverty23:46
Welshmoney only creates freedom to a very limited point23:47
Welshbut that's another chat23:47
jelknerok then, it's a wrap23:47
flint_or more or less another rant...23:47
jelknerpcardune: can i talk to you on gaim?23:47
jintywe want to do something that makes us feel good, but not starve...23:47
pcardunesure thing23:47
* tvon tips his hat23:48
flint_believe me brian I am thinking hard about the getting paid part.  Talking with mdz.23:48
jelknersee ya all next week, same time, same #schooltool channel23:48
*** tiredbones has joined #schooltool23:48
*** jelkner has quit IRC23:49
*** ffsnoopy has quit IRC23:49
*** jinty has quit IRC23:49
Welshelkner, what phone are you at?23:49
*** Welsh has quit IRC23:50
flint_ghost flint flint23:53
*** flint_ has left #schooltool23:56
*** flint_ has joined #schooltool23:57

Generated by 2.15.1 by Marius Gedminas - find it at!