IRC log of #schooltool for Thursday, 2005-01-06

**** ŽURNALAS PRADĖTAS Thu Jan 6 15:01:51 2005
-->Jūs dabar kalbate #schooltool15:01
---#schooltool tema yra http://schooltool.org15:01
---#schooltool temą nustatė mgedmin Wed Jan 5 20:29:01 200515:01
**** ŽURNALAS BAIGTAS Thu Jan 6 16:32:50 2005
**** ŽURNALAS PRADĖTAS Thu Jan 6 16:32:50 2005
---Atsijungta ().16:42
**** ŽURNALAS BAIGTAS Thu Jan 6 16:42:20 2005
**** ŽURNALAS PRADĖTAS Thu Jan 6 17:16:08 2005
-->Jūs dabar kalbate #schooltool17:16
---#schooltool tema yra http://schooltool.org17:16
---#schooltool temą nustatė mgedmin Wed Jan 5 20:29:01 200517:16
---#schooltool :[freenode-info] please register your nickname...don't forget to auto-identify!
*mgedmin sitting in a cellar in a cafe17:16
-->alga (~alga@ atėjo į #SchoolTool17:16
*mgedmin has internet over wifi to alga's laptop which gets internet over gprs via a bluetooth phone17:17
algahi guys17:18
th1aGive me a minute to blog that.17:19
SteveA"I'll blog that for a dollar!"17:20
tvonactually I think the zodb errors are due to some firefox tweaks I made recently17:21
tvonseems fine in epiphany17:21
SteveAhow on earth could a browser influence getting a zodb error?17:21
SteveAunless something is wrong at a deeper level17:21
tvonlemme see if I can find it17:21
SteveAare you using frames?17:23
tvonthough Brian says he doesnt get the error and he did the same thing I believe17:25
th1aThere's a reason that's off by default.17:25
*bska|mobile nods17:25
mgedminwhat error do you get?17:25
SteveAthat is such a sucky thing to do17:25
SteveAnetscape did the same trick to make navigator seem faster back in the day17:25
mgedminHTTP pipelining is part of the HTTP/1.1 standard, isn't that?  and schooltool claims to support HTTP/1.117:26
---Atsijungta ().17:41
**** ŽURNALAS BAIGTAS Thu Jan 6 17:41:20 2005
**** ŽURNALAS PRADĖTAS Thu Jan 6 17:43:14 2005
-->Jūs dabar kalbate #schooltool17:43
---#schooltool tema yra http://schooltool.org17:43
---#schooltool temą nustatė mgedmin Wed Jan 5 20:29:01 200517:43
---#schooltool :[freenode-info] please register your nickname...don't forget to auto-identify!
SteveAat canonical, we'll be moving launchpad to use 2.4 in the near future17:43
tvonis it an issue with packaging schooltool or with testing 2.4?17:43
-->alga_ (~alga@ atėjo į #SchoolTool17:43
SteveAzope3 should support python 2.4 right now17:44
tvonnot to suggest we would dep on 2.417:44
bska|mobilemgedmin_: you mentioned generating an empty .po file for testing i18n: coverage17:45
bska|mobileafter make extract-translations and update-translations17:46
jintySo then we just follow the default pyhton17:46
mgedmin_not exactly an empty one17:46
mgedmin_use msgen to generate an English translation (each msgstr == msgid)17:46
jintyand change to 2.5 automatically when that comes around.17:46
SteveAjinty:  i think we should promote using 2.4 on hoary for everything except zope 217:46
mgedmin_then use msgfilter to change the translations17:46
mgedmin_then you will see which strings were translated17:46
SteveAi have no idea what python 2.5 will bring17:47
mgedmin_jinty, could you change 'make schooltooltar' so that it builds schooltool.pot and include schooltool.pot in the tarball?17:48
jintyI don't want to specify numbers if we support 2.4 because then the packages for ubuntu and debian have to be different17:48
SteveAthere is no 2.5 yet, so that's not a problem.  there will not be a 2.5 in hoary17:49
tvoncan't we dep on >= 2.3 ?17:49
jintymgedmin: yes.17:49
jintyok, so, for now follow the default - and re-asses the situation when 2.5 comes.17:50
<--alga išėjo (Read error: 110 (Connection timed out))17:50
SteveAthe schooltool project will need to write python at the 2.3 level17:51
<--mgedmin išėjo (Read error: 110 (Connection timed out))17:51
jintyplease, for backportablitity  to sarge17:51
jintyAnyway, my time is up -- so caio17:52
<--jinty išėjo ("Leaving")17:53
SteveAwell, it needs to be an explicit statement of the schooltool project17:53
SteveAthat it will be written to python 2.3, and compatible with python 2.3 and python 2.417:53
th1aI'll add it to the docs.17:55
*mgedmin_ has a theory about conflict errors17:58
SteveAmarius has a good idea what the conflict error is17:58
SteveAfortunately, it is fixed by moving to use more of zope 317:58
SteveAthe current "session" machinery in schooltool, which deals with cookie auth, writes to the database a bit on each access17:59
SteveAthis writing creates contention17:59
SteveAthe zope3 session machinery writes only infrequently18:00
SteveAso avoiding this problem18:00
mgedmin_from now on this will be known as
**** ŽURNALAS BAIGTAS Thu Jan 6 18:03:22 2005
**** ŽURNALAS PRADĖTAS Thu Jan 6 18:03:22 2005
mgedmin_one other thing that needs to be done before a release18:04
mgedmin_make update-translations18:04
mgedmin_then check schoolbell.po for missing translations18:04
mgedmin_although this is probably not necessary in the schoolbell-ui branch18:04
mgedmin_if the code and page templates no longer refer to "teachers" and "pupils"18:04
SteveAth1a: let's talk about meetings18:07
SteveAso, I'd like to see the two teams talking about what they did during the past week,18:08
SteveAwhat issues stood in the way of effectively working on the parts of schooltool they've been working on18:08
SteveAthe plans for the coming week of work18:08
th1aThat certainly sounds reasonable.18:09
SteveAwork out what things depend on other things, especially across teams18:09
SteveAtalk about interesting items of code or UI that were checked in18:10
SteveAmaybe choose one item from each team and look at it in the meeting18:10
SteveAconstructive criticism18:10
SteveAdiscussion of using the code18:10
SteveAof course, it doesn't have to be all of this at once18:10
bska|mobilethat makes sense18:11
---bska|mobile dabar žinoma(s) kaip bskahan18:11
th1aRight.  I suppose I've been expecting the conversation to generate itself more spontaneously and feeling frustrated about it not happening.18:11
SteveAput together a standard agenda18:12
th1aSome structure.18:12
*mgedmin_ thought this specific meeting was an exception18:13
mgedmin_no standard agenda, just focusing on the release18:13
mgedmin_making sure that all things that need to be done will be done18:14
SteveAthat's fine18:14
th1aThe thing you guys need to remember is that I've _never_ done this before.18:14
th1aSo I don't even know exactly what happens in a bug squashing meeting over IRC.18:15
*mgedmin_ neither18:15
SteveAlook through the bugs.  choose the top 5.18:16
SteveAwork on them, different people / groups taking on each one.  talking about progress / problems.18:16
tvonThe biggest that I know of is the rendering issue in KHTMl and the quirks in mozilla18:16
SteveAmaybe get other to review proposed fixes18:16
th1aOK.  I'll keep testing with iCal.  Something is out of whack.18:17
SteveAth1a: tcpwatch can be helpful to see what is going on18:18
th1aI've generally used ethereal.18:18
mgedmin_there's also a mozilla extension (livehttpheaders)18:19
th1aI'll take a quick look at that.18:20
th1aWhen you edit an event in the popup, it doesn't refresh the main window.  Is there a way to do that?18:20
tvonhrm...anyway to sort by priority in our Roundup that I'm missing?18:21
th1aI'm regretting saying we should use a popup for that.  It seems like it introduces way more complication than I'd realized.18:21
tvonah, nm18:21
mgedmin_we're about to go back to the office18:22
mgedmin_we'll be offline for about 20 minutes18:22
-->gintas ( atėjo į #schooltool18:22
*mgedmin_ dislikes popups greatly and has disabled them in his firefox18:22
bskahanth1a: yes18:23
bskahanthe add event popup closes/refreshes18:23
tvonI don't care for popups too much myself, though they can give more of an application feel at times18:23
bskahanthe edit event popup doesn't18:23
bskahanI noticed it yesterday18:23
<--alga_ išėjo ("see you when we get back")18:23
---Atsijungta ().18:23
**** ŽURNALAS BAIGTAS Thu Jan 6 18:23:57 2005
**** ŽURNALAS PRADĖTAS Thu Jan 6 18:45:16 2005
-->Jūs dabar kalbate #schooltool18:45
---#schooltool tema yra http://schooltool.org18:45
---#schooltool temą nustatė mgedmin Wed Jan 5 20:29:01 200518:45
th1aSo let's pick a reasonable number (like 20) and go on.18:47
*bskahan nods18:47
*mgedmin was just looking at checkins on schooltool-ui branch18:48
th1aIn the month view I'd just like each day to be a fixed size.18:48
mgedminbskahan, is the 'portlet_calendars' attribute used anywhere?  I think it should be just removed18:48
bskahanI'm working on the form to modify it now18:51
mgedminI thought we had agreed to replace it with a relationship18:51
bskahanits a list of additional calendars (other than groups your a member of) to add to the portlet18:52
bskahanthe displayed calendars use the subscription relationship now18:52
tvonth1a: I'm working on that one.  The monthly view is still a table which is causing problems when trying to fix the size18:53
bskahanportlet_calendars lets you add calendars of resources/people/groups-you-arent-a-member of to your portket so that you can subscribe/unsusbscribe to them18:53
mgedminthere are two pretty similair things:18:54
mgedmin- a list of objects whose calendars you want merged with your calendar18:54
mgedmin- a list of objects that you want shown in the portlet so that you can easily merge them with your calendar18:54
mgedminI think it would be better if they were implemented in the same way18:54
mgedminit is possible that defining a new relationship is too hard now18:55
mgedminif it is so, then it should be fixed18:55
mgedmindefining new relationships should be trivial18:55
bskahanits alot of boilerplate18:56
mgedminI've noticed18:56
bskahanits not difficult though18:56
<--SteveA išėjo ("Leaving")18:56
mgedminI agree that it there should be less18:56
mgedminI think that just defining three URIs and using the 'relate' function will work18:57
mgedminrelationship schemas add some syntactic sugar and some type checking18:57
-->SteveA ( atėjo į #schooltool18:58
tvonth1a: can we push the fixed width till after the release (when we can use tableless layouts for the montly views)?18:58
th1aHow does PHP iCalendar do it?19:00
th1aI'm not sure what the hard part here is.19:00
tvonits not doing what I tell it to, in short19:01
th1aIn what sense?19:01
th1aWhat's happening?19:01
tvonone sec19:03
th1amgedmin:  We need to do a little more ugly trickery to get iCal to be able to post calendars.19:04
Aisteth1a: did you read my note on the iCal bug you filed?19:04
th1aLet me check.19:04
Aisteit is possible to do that without any tricketry19:05
mgedminth1a, let me get this straight: in iCal you can only do calendar synchronization in one direction19:05
-->alga ( atėjo į #SchoolTool19:05
mgedminthat is, either you subscribe to a web calendar and have it read-only19:05
mgedminor you have a local calendar and publish it on the web19:05
mgedminwhen you publish a calendar you have to manually enter the calendar name and URL, right?19:06
Aistethat is correct as far as I understand it19:07
mgedminwe should document that the calendar name must be 'calendar' and the URL must end with the user's username (http://servername:7080/persons/username)19:07
th1aThat only works if you're managing one calendar.19:07
mgedminwhat do you mean?19:08
th1aSince the calendar's name has to be 'calendar'19:08
mgedminschoolbell currently supports only one calendar per person19:08
th1aIf you're managing, say, the school calendar and you're personal calendar.19:09
mgedminwe might consider allowing several calendars per person, but I suggest doing that after schoolbell is a zope3 component19:09
th1aYeah, not now.19:09
mgedminif you're managing the school calendar and your personal calendar19:09
mgedminI assume that you will have two separate local calendar19:09
mgedminyou will publish one to http://servername:7080/groups/root and another to http://servername:7080/persons/me19:10
th1aThe problem is that literally the first two good test sites I have both have one person managing multiple calendars in iCal, so this isn't an edge case.19:10
mgedminboth will be published as "calendar"19:10
mgedminhowever the publication name can be different from the local calendar name19:10
th1aAh.  Let me check that on mine.19:10
tvonth1a: I think I can get the width.19:10
<--thisfred išėjo ("Farewell, cruel channel...")19:11
th1aAh.  Brilliant.19:14
th1aOK.  Yeah, you can make the publish name different than the "display" name.19:14
th1aSo we're good.19:14
th1aI'll take care of adding that to the docs.19:14
th1aiCal does do link hooking properly.19:16
th1aDoesn't seem to work at all over https at all, but that's not our problem.19:23
th1aApparently... 'The "normal" workaround for this bug is first subscribe to http:// calendar (insecure), then manualy edit ~/Library -> Preferences -> and replace http:// with https://.'19:24
th1aOr so I'm told.19:24
Aisteoh :( that is not very nice19:25
Aisteand I was wondering the other day, why it wouldn't connect over https...19:25
th1aApple is on the naughty list.19:25
Aiste:)19:25 is only available over https19:25
th1aThat's how I discovered that issue.19:25
tvonthe Firefox calendar extension seems to work...on the brighter side of life19:29
tvonoh, wait...I lied19:30
th1aCalendar extension?19:30
bskahanmozilla calendar for firefox19:31
th1aNot Sunbird?19:31
bskahansunbird segfaults on me19:31
bskahanthe last 3 releases19:31
tvonwait, no, it does work19:33
th1aOne would hope.19:34
*bskahan wishes evolution did iCal19:35
algait almost does19:35
mgedminevolution can subscribe to a web calendar, but it can't publish19:36
mgedminas a nice side effect, if you subscribe to a web calendar with evolution, those calendar events will be visible in the calendar popup in your GNOME desktop19:36
mgedminhowever evolution is a stinking piece^W^Wnot really ready for prime time yet19:37
th1aLet's see... other issues.19:43
th1aAre the merge calendar checkboxes going to be removed from the person info page?19:43
tvonI have a q relating to overlays19:45
bskahanassuming I get the relationships working for URICalendarListing there are 3 boxes to add at the bottom of the page for updating the portlet listing19:45
tvonWe need to store colors for each overlay for use in the UI.  Any suggestions on where to actually store this information?19:45
bskahan3 boxes - groups, resources, people19:45
tvonI figure the colors need to be consistant on a per user basis..meaning that Foo group is always Mauve when Joe views his calendar....beyond that I'm not sure if any consistancy is required (Foo needn't always be Mauve for everyone, as that gets complicated).19:46
th1aRight.  Consistent per user.19:47
tvonmy kneejerk reaction is to store a dict on the user with {calendar.__parent__.__name__:color}...but that doesnt sound like a "right way"19:48
mgedmintvon, a very good question19:48
mgedmineasy workaround: a_list_of_standard_colors[hash(getPath(calendar)) % len(a_list_of_standard_colors)] or something19:49
mgedminthe "right way" would be attaching annotations to relationships, however we cannot do that now19:50
algaor introduce several rels:19:52
*mgedmin shudders19:52
mgedminyou can have a PersistentKeysDict with calendars as keys19:53
mgedmin(from schooltool.db import PersistentKeysDict)19:53
algahm, register a utility that chooses a colour for a calendar?19:53
*mgedmin thinks19:53
mgedminperhaps relationships were not such a good idea after all...19:54
*bskahan stops typing19:54
mgedminok, the advantage of relationships:19:55
algain principle, facets are annotations on steroids19:55
mgedmin- no need to change interfaces19:55
mgedmin- when a group or a person is removed, the relevant relationships disappear automatically19:55
mgedmindisadvantage of relationships:19:55
mgedmin- no way to attach extra information to a relationship...19:55
mgedminactually that's not true19:55
mgedminwe have an (insanely complicated) way to attach extra information to a relationship19:56
mgedminit involves facets19:56
tvonrelate it to a note19:56
algacome on, we can use Zope annotations!19:56
mgedminnot in schoolbell-ui19:56
alganot yet, you mean19:56
mgedminwe can use zope annotations in the trunk19:57
mgedminbut the relevant zope packages are not present in the schoolbell-ui branch19:57
tvonheh, unless we merge before the release :)19:58
tvonfor the future, it would be good to have a way to generate a years worth of an 'intersting calendar' for importing20:07
tvonfor testing20:07
<--bskahan išėjo ("Leaving")20:08
-->bska|mobile ( atėjo į #schooltool20:08
---bska|mobile dabar žinoma(s) kaip bskahan20:09
<--tvon ( paliko #schooltool ("Leaving")20:16
-->tvon ( atėjo į #schooltool20:16
th1aWhere do we end up in the discussion of storing colors?20:16
th1aThat was a good discussion for a while.20:17
th1aMeanwhile, here's a little question:  if I was to add a link to a calendar's acl page to the left toolbar, where would I do that?20:19
th1aLeft column.20:19
tvonyou could fill the "column-left-extra" slot in the template20:19
th1aWhich template?20:20 I believe20:20
tvonstick it between the h1 slot and the content slot20:20
tvon<metal:slot fill-slot="column-left-extra">something</metal:slot>20:20
th1aWhich template do I edit?20:20
tvonoh oh20:21
tvona link "to" the acl page, not "on" the acl page, sorry20:21
tvonhrm...from all the calendar views?  probably in the calendarpage macro20:22
bskahanthe last macro in macros.pt20:23
th1aThen make a little macro for it?20:24
bskahanthat will put it on all calendar views20:24
th1aAnd I don't need to do a test for that kind of thing, right?20:25
bskahanthere are places where the rendered output is tested, so its possible to break existing tests that way20:25
bskahannormally it won't20:26
th1aI'm going to take a shower.20:27
th1aHopefully alga and mgedmin aren't lying unconscious in a cellar somewhere in Vilnius.20:27
*bskahan agrees20:28
mgedmindo we need user-selectable colours for calendars in this release?20:30
bskahanI don't think we need user selectable colors20:32
bskahantvon disagrees with me20:32
bskahanat least not for schooltool20:32
bskahanfor schoolbell its a different story20:32
mgedminI think we need user selectable colors for schoolbell 1.020:32
bskahanbut either way, not for this release20:32
SteveAth1a: did you agree with jeff waugh what status schoolbell will have in hoary?20:32
mgedminI as asking about 0.920:32
tvonthey werent in the plan20:33
tvonfor 920:33
tvonplan was to provide 20 or so pre-selected colors20:33
tvon(though I think thats a lot)20:33
<--tvon ( paliko #schooltool ("Leaving")20:37
-->tvon ( atėjo į #schooltool20:37
th1aSteveA:  I didn't hear back from Jeff.  Did you?  What's his email?20:40
th1aYeah, 20 is a lot of colors, having thought about it a bit.20:41
bskahanhow many calendars do you expect people to display at one time?20:43
bskahanin  a school setting my thought model is:20:43
tvonI really think 4-6 colors would be enough...10 at an absolute maximum20:44
tvonits mostly a matter of finding colors that are easily distinguishable and don't need a changed font color20:44
bskahana teacher with 1 personal, 5 classes, 1 dept., 1 ext carricular, and 1 school20:44
SteveAjeff.waugh at should reach him20:44
tvonand ideally that still fit the theme20:44
th1aA student could have six or seven classes, with one calendar for each class.20:44
th1aSports schedules...20:44
tvonth1a: true, but thats not needed for schoolbell is it?20:44
bskahanI lean towards 1020:45
th1aAnyway, we just need a palette of colors.  10-ish is enough for now.20:45
th1aMy discussion of colors with my sister got theoretical pretty quickly.20:46
tvonin the long run we can try mixing patterns in too20:46
th1aVarious ways to generate a large variety of distinct yet harmonious colors.20:46
th1aShe was also thrown off by the transparency issues.20:47
th1aShe's kind of a perfectionist.20:47
th1aSo anyway, just pick 10 colors for now.20:48
th1aWe can tweak that later.20:48
tvonthe idea needs some testing though20:48
*tvon snuggles the unloved little overlapping translucent events20:48
tvonokay, 10 colors is no problem20:49
th1aThe translucency is nice.20:53
th1aSteveA:  I sent an email to Jeff.20:53
bskahanth1a: I've pondered the automatic color generation problem a bit (being colorblind and all) it gets ugly pretty quickly20:54
th1aYou can pick a harmonious set of colors and then vary them in different ways to create more variations.20:55
*bskahan nods20:55
bskahanadobe has published some whitepapers on it20:55
th1aAh.  Makes sense.20:56
mgedminthat's interesting, any urls?20:57
bskahanI'll look for them20:58
mgedminwow, I'm famous21:17
mgedmin1. create a weblog21:18
mgedmin2. ???21:18
mgedmin3. profit!21:18
th1aSo did we reach a conclusion about how to store the colors?21:20
mgedminI think the conclusion was to assign them on the fly21:20
th1aThat'll do for now, I guess.21:20
tvonI'm not sure that will work...that gives each calendar an assigned color basically, right?21:21
mgedminif you use hashing on the groups name or path, the colours will be stable21:21
*mgedmin thinks a bit more21:21
mgedmin(keyboard accident21:22
mgedminhashing may result in collisions, though21:22
tvonif each calendar has an assigned color and there are only 10 colors then there is a very good chance that some user will have overlaid calendars of the same color21:22
bskahanthat could lead to one person having all calendars the same color (depending on their selection) and another person having all unique colers21:22
mgedminif you have N calendars overlaid and select an extra one21:23
mgedminI suppose you want those N calendars to keep the same colour21:23
tvonthough we could overlook that for this release...  21:23
tvonI'm open to suggestions21:24
bskahanmy initial thought was to assign a color to each color added to your portlet in succession21:25
bskahanbut that doesn't handle removing one in the middle later on21:25
mgedminit would be best to store colours21:27
bskahancolor facet on the calendar, updated by URICalendarListing relate events?21:28
mgedmindifferent users may have different colours for the same calendar21:29
bskahanI meant on the users calendar21:30
mgedminoption 1 (ugly dirty hack): set an attribute on the relationship link21:30
mgedminoption 2 (large architectural investment): devise a framework for attaching arbitrary annotations on relationships21:30
mgedminor perhaps a specific schema of attributes for a given relationship type21:30
bskahanwhat would you set the attribute on?21:31
mgedminoption 3: forget about relationships, use a named facet on a person for storing a list of tuples (calendar, overlay, colour)21:31
mgedminoption 4: add an attribute to ICalendarOwner with this attribute instead of using a facet21:31
bskahanI like option 3 at this point21:32
mgedminbtw ICalendarOwner has a method makeCompositeCalendar21:34
mgedminhaving an attribute that explains how the composite calendar is made is somewhat fitting21:34
mgedminI think 3 or 4 makes more sense21:36
mgedmin4 might be easier to implement21:36
*bskahan agrees21:36
mgedminthere are some downsides21:36
mgedminif you delete a group, now you must go through *all* objects in the system and remove that group from their subscribed_to_calendars_or_something attribute21:37
bskahantrue, with relationships you don't have to21:38
mgedminotherwise the group object will be kept in the ZODB and never garbage collected21:38
mgedminalso, attributes that are lists of items are somewhat tricky to use correctly with ZODB21:38
bskahanhow so21:38
mgedminit either must be a PersistentList or you must never modify its items directly21:39
mgedmin(or if you do modify its items directly, you have to change some attribute on the object to make sure ZODB notices the change)21:39
bskahanI didn't know that (/me wonders about bugs out in the wild) thanks21:43
th1aSo is the problem with that line 148 in shouldn't be using tal:condition="view/isManager"?21:47
tvonth1a: yeah, that seems to be it21:49
th1aTa da!21:50
tvonshould check to see if authenticated user has permissions to modify the acl21:50
bskahanmgedmin: thanks21:50
bskahanth1a: the link for manager is in calendarpages content, iirc21:52
bskahanto edit calendar acl21:53
th1aActually, is that a bug in the first place.21:54
th1aIt might be a bug for SchoolBell but a feature for SchoolTool.21:54
th1aBecause student's shouldn't be able to make their calendar public.21:55
th1aBut in most organizations it would be ok to decide...21:55
th1aAlthough employees (for example) shouldn't be able to arbitrarily decide to make their calendar private, if their employer purposely set up a public calendar for him.21:57
th1aI think this isn't a bug.21:58
tvonneed ModifyCalendarACLPermission of sorts21:58
tvonbecause in a "normal" calendaring environment a person should be able to share/hide their calendar21:59
th1aIt is fine for now.21:59
bskahanI think its a bug, whether or not that link appears is a (possibly complex) ACL issue21:59
th1aWhat does "Nosy List" mean, by the way?22:00
bskahanI expected it to mean I would get mails when changes happen to the bug22:00
tvonala zilla's CC list I think22:01
bskahanthat would be the bugzilla CC: equivalent22:01
bskahanbut it doesn't seem to22:01
bskahanmgedmin: on balance I think option 4 wins22:03
bskahanusing relationships is probably the easiest, but I don't know how we'd do the colors in that case22:04
mgedminNosy List in roundup contains comma separated usernames, not emails22:05
mgedminI think we need to rethink our relationship APIs a little bit22:06
tvonmgedmin: I assumed the usernames were supposed to represent account holders, which have emails22:06
mgedminI meant that the nosy list field contains usernames, not email addresses22:07
bskahanafter saying that, I got an email on a bug I was nosy on22:08
bskahanmgedmin: can you add us to the default nosy list22:08
*mgedmin tries to remember where that one is defined... probably in the python source file somewhere22:10
bskahannot a huge deal, but if you think of it some time22:10
th1aalga:  I see you found SchoolBell on Rosetta.  I'm afraid I couldn't make heads or tails of what I was supposed to do with it.22:17
th1aMakes a little more sense now.22:19
th1aSo should we direct people who are interested in doing translations to Rosetta?22:19
th1aIt's here, btw:
SteveAthe rosetta team will be very receptive to feedback from translators22:27
th1aDo they have an email address.22:29
SteveAthere's a mailing list for rosetta users22:30
SteveAbut also, mail carlos at or daf at canonical.com22:30
th1aSo, Rosetta isn't exclusive of other translation tools, right.  If someone wants to use KBabel, or something like that, we'll be able to import their work into it?22:31
tvonmgedmin: for this release I'm leaning towards option #1 for colors (hack an attribute onto the relationship).  After we merge with trunk we can make the relationships annotatable and do it that way.22:35
tvon#2 requires significant effort/time and restructuring.  #3 and #4 both have issues when something is deleted22:36
tvonthats my thinking anyways22:37
th1aI'm going to shovel the walk.22:56
bskahanits 50 F here23:01
bskahansupposed to be 60s next week23:01
SteveAth1a: i don't know kbabel, but i'd expect so, so long as everything uses po files23:02
th1aOK.  I'll announce it to the list.23:02
*mgedmin was trying to think of a more or less clean way to implement option #1, but couldn't23:03
mgedminit's getting late23:03
mgedminsee you tomorrow23:03
---Atsijungta ().23:03
**** ŽURNALAS BAIGTAS Thu Jan 6 23:03:33 2005

Generated by 2.15.1 by Marius Gedminas - find it at!