Timestamps are in GMT/BST.
[2:04] * jango_ (~jango@67.55.48.9) Quit (Ping timeout: 240 seconds)
[2:05] * jango_ (~jango@173.248.218.155) has joined #basie
[2:12] * jango__ (~jango@dsl-173-248-216-213.acanac.net) has joined #basie
[2:15] * jango_ (~jango@173.248.218.155) Quit (Ping timeout: 240 seconds)
[2:23] * jbalogh (~jeff@c-71-202-47-205.hsd1.ca.comcast.net) Quit (Quit: jbalogh)
[2:37] * jango__ (~jango@dsl-173-248-216-213.acanac.net) Quit (Ping timeout: 240 seconds)
[2:38] * jango__ (~jango@67.55.52.5) has joined #basie
[3:11] * jango__ (~jango@67.55.52.5) Quit (Ping timeout: 240 seconds)
[3:11] * jango__ (~jango@dsl-173-248-219-4.acanac.net) has joined #basie
[5:13] * colin__m (~colin@whit179.ucres.utoronto.ca) has joined #basie
[5:14] * colin__m (~colin@whit179.ucres.utoronto.ca) Quit (Client Quit)
[5:15] <colin_m> I'm shocked at how many people are online at 5 AM. Anyone alive here?
[5:22] <colin_m> I guess everyone's doing the sensible thing and sleeping.
[7:24] <jango__> correcto
[7:33] <colin_m> Blah, I find the impotence of Django's templating language infuriating. Concatenating strings should not be this hard!
[7:33] <jango__> concatenating strings? what are you trying to do?
[7:35] <colin_m> I have two variables that each hold a string. I want to use their concatenation in a url tag.
[7:35] <colin_m> I thought maybe there would be a filter to do it like the add filter, but there doesn't seem to be. Am I missing something obvious?
[7:42] <jango__> hm
[7:42] <jango__> {{ STATIC_URL }}{{ MEDIA_URL }}
[7:43] <jango__> this will give you a concat string of static_url + media_url in your rendered template
[7:48] <colin_m> Hm, doesn't seem to be working. Am I allowed to put that within a url tag?
[7:48] <colin_m> The current line looks like:
[7:48] <colin_m> <td><a href="{% url svnbrowser.views.view_path path={{ pwd }}{{ f.name }} rev=f.rev %}">{{ f.name }}</a></th>
[7:54] <jango__> i don't think this is how you'd use a url template tag
[7:54] <jango__> http://docs.djangoproject.com/en/dev/topics/http/urls/
[7:54] <jango__> scroll down to naming url patterns
[7:54] <jango__> {% url arch-summary 1945 %}
[7:57] <colin_m> To me, that looks pretty much the same as what I'm doing
[7:57] <jango__> yes it does
[7:57] <jango__> lol
[7:57] <jango__> sorry, I misread it ;D
[7:57] <colin_m> Ah, okay, phew
[7:58] <jango__> can you try putting double quotes around {{ pwd }}{{ f.name }}?
[8:01] <colin_m> Hm, that's odd
[8:01] <colin_m> It's interpreting them as literals
[8:01] <colin_m> So I'm getting links like http://127.0.0.1:8000/svn/R12/{{ pwd }}{{ f.name }}
[8:01] <jango__> tsk.
[8:02] <jango__> stubborn django
[8:03] <colin_m> I know there's a strong argument for separating logic from presentation, but I do like templating languages like Mako that just let you insert raw python willy-nilly :/
[8:03] <colin_m> Well I guess I could just make a string concatenation filter of my own. Feels like overkill though... it's such a common thing, you would think there would be some easier way to do it.
[8:04] <jango__> i am still sure you could concatinate w/o applying too much effort
[8:04] <jango__> ask on the mailing list first
[8:04] <colin_m> The Django mailing list?
[8:04] <jango__> and on stackoverflow, if nobody knows
[8:04] <jango__> basie
[8:05] <colin_m> Okay, good idea
[8:05] <jango__> it might save you some time
[8:11] * bwinton_away is now known as bwinton
[8:17] * colin_m (~colin@whit179.ucres.utoronto.ca) Quit (Ping timeout: 276 seconds)
[8:47] <bwinton> Submitted for your consideration: http://blog.medallia.com/2010/06/continuous_integration_at_meda.html
[8:52] * hazure (~cjmuise@red-gw37.cs.toronto.edu) has joined #basie
[9:32] <jango__> @bwinton branch graph looks appealing. have you seen graphs that google code produces for mercurial?
[9:35] <bwinton> Nope. Got a url?
[9:35] <jango__> http://code.google.com/p/elmcity/source/list
[9:36] <jango__> it looks a bit cleaner
[9:36] <jango__> just because the circles are all the same
[9:39] <bwinton> Yeah, although it looks like a different thing???
[9:40] <bwinton> i.e. that looks like the log, with some extra info, the medallia one looks more like a graph (with no intersecting lines, and using two dimensions for layout instead of just one.)
[9:41] <jango__> true
[9:44] <bwinton> Basically, every VCS viewer I've seen does it the Google Code way (Murky/MagHg/Github/Bitbucket/etc)??? It might be interesting to do something different, if we can figure out something different to do.
[9:50] * colin_m (~colin@whit179.ucres.utoronto.ca) has joined #basie
[10:01] * bwinton is now known as bwinton_away
[10:06] * jharjono (~chatzilla@nat/ibm/x-pfjmfvponjeyenhu) has joined #basie
[10:30] * Sacho (~sacho@213.91.244.15) has joined #basie
[10:30] * Sacho (~sacho@213.91.244.15) has left #basie
[10:39] * Joel_C (~chatzilla@uc059-2.uc.utoronto.ca) has joined #basie
[10:45] * mi_sa (~chatzilla@nat/ibm/x-jvewhwxgzlubeqdl) has joined #basie
[10:51] * jbalogh (~jeff@c-71-202-47-205.hsd1.ca.comcast.net) has joined #basie
[10:55] * gvwilson (~chatzilla@red-gw28.cs.toronto.edu) has joined #basie
[10:56] * DaveWeber (~chatzilla@nat/ibm/x-gmfodktafztzyjbh) has joined #basie
[10:58] <colin_m> Hi guys, we'll be getting started in two minutes.
[11:00] <colin_m> Okay, it's now 11, let's get this ball rolling. Role call:
[11:00] <colin_m> _lance_?
[11:01] <colin_m> bwinton?
[11:01] <colin_m> cvan?
[11:01] <_lance_> I'm here
[11:01] <colin_m> DaveWeber?
[11:02] <colin_m> I'm really striking out here.
[11:02] <DaveWeber> Hi
[11:02] <DaveWeber> sorry im here
[11:02] <colin_m> gvwilson?
[11:02] <gvwilson> you're sorry you're here? :-)
[11:02] <gvwilson> yes, I'm here
[11:02] <colin_m> hazure?
[11:02] <DaveWeber> Sorry (for not answering quicker), I'm here
[11:02] <DaveWeber> sheesh
[11:02] <colin_m> jango__?
[11:02] <jango__> ici
[11:03] <colin_m> jbalogh?
[11:03] <DaveWeber> :)
[11:03] * jango__ is now known as nikita
[11:03] <colin_m> jharjono?
[11:03] <jharjono> here
[11:03] <colin_m> Joel_C?
[11:03] <Joel_C> here
[11:03] <colin_m> jyeung?
[11:03] <jyeung> =)
[11:03] <colin_m> mi_sa?
[11:03] <mi_sa> yep
[11:04] <gvwilson> if this was my daughter's daycare, we would now all sing a happy song
[11:04] <colin_m> Okay, looks like we're just missing cvan, hazure, and jbalogh then.
[11:04] <gvwilson> so thank god we're having this meeting in IRC rather than on Skype
[11:04] <colin_m> Hah. Well let's forge on.
[11:04] <colin_m> Let's start by going over what the teams were supposed to have done this week and what they actually did.
[11:04] <colin_m> The front-end team was supposed to start on taking the mock-ups we made the previous week and integrating them into a Pinax/Django environment, which Johan set up for us. How did that go, front-end team?
[11:05] <colin_m> I worked on building a new (ugly but functional) directory view page from scratch using Django templates and views. To keep the view as close to the real thing as possible, I made up a little module of mocked-up API functions that return bogus but plausible values (either hard-coded, or arrived at by some silly means). It's still very rudimentary though, and I'm working through a few kinks with regard to Django's templating language.
[11:06] <gvwilson> are you doing this solo, or collaborating with...?
[11:06] <colin_m> Incidentally, if anyone knows the arcane art of concatenating strings in a Django template, I'd love to here about it after the meeting.
[11:06] <colin_m> mi_sa was also working on the front end
[11:06] <colin_m> Noam is our third member, but he has exams right now, so he doesn't have time to code.
[11:06] <mi_sa> I have some pages 'working' -- namely the directory view and files view
[11:07] <colin_m> Did you run into any particular roadblocks?
[11:07] <gvwilson> when will you (plural) be ready to sit down with a few real users and do some usability testing?
[11:08] <mi_sa> I had a lot of trouble getting external files to get recognized
[11:08] <colin_m> Well, we could probably do that using the mock-ups that we already have.
[11:09] <colin_m> Who are our real users? Aren't they basically people like us? Or are we too closely involved with the project to see the forest for the trees?
[11:09] <mi_sa> colin and I had a disagreement about the mockups to be created. I went about modifying our existing mockups and getting the links/buttons to work, so that we could have more pages done quicker
[11:09] <gvwilson> @mi_sa what are "external files"?
[11:09] <mi_sa> js, css, etc
[11:09] <gvwilson> @colin_m our real users are second- and third-year undergrads, so we can test on your fellow students at St. George
[11:10] <gvwilson> @mi_sa @colin_m what was the other option? (i.e., the source of the disagreement?)
[11:10] <mi_sa> colin thought it made more sense to have projected API calls and return hard-coded data
[11:10] <mi_sa> that seemed to take more time/effort than was necessary if we wanted users to test out the GUI
[11:10] <mi_sa> because users dont' care about the api
[11:11] <colin_m> I was pretty stubborn about not wanting to spend any more time putting layers of shellac on our hard-coded HTML pages, since I figured you wouldn't have us doing any proper usability tests with real users (but obviously I was wrong)
[11:11] <mi_sa> and also because api hasn't even been agreed on yet by the two back-end teams
[11:11] <gvwilson> agreed: you can do a lot of usability testing with hand-drawn sketches on paper to see if the workflow's going to work before you write any code
[11:11] <gvwilson> although in this case, we're pretty confident that UI and expectations will match
[11:12] <colin_m> Okay, back-end teams?
[11:12] <colin_m> The back-end teams were supposed to redo the API descriptions they did the previous week so that they were, wherever possible, as similar as possible. What did you come up with, back-end teams? (I'm addressing the HG and SVN teams as one, since it's my understanding that you worked together this week)
[11:13] <_lance_> I came up with the class diagram which I'll pull up in a moment
[11:13] <_lance_> Minor change to it: I think the Hg team is going to merge HgFile and HgDirectory
[11:13] <gvwilson> have the other back-end devs seen it/commented on it/agreed to it
[11:13] <_lance_> http://dl.dropbox.com/u/2179828/ClassDiagram.png
[11:14] <_lance_> One rep from the Hg team has seen it and commented on it, have not agreed on it yet
[11:14] <gvwilson> VcsDiff and its children need some work :-)
[11:14] <gvwilson> but to clarify: no code yet?
[11:15] <gvwilson> and you haven't mapped out the sequence of calls that will be triggered in response to particular HTTP requests?
[11:15] <_lance_> Afraid not
[11:15] <colin_m> Can you tell us about any asymmetries in the design between SVN and HG?
[11:16] <_lance_> The way the pathing is done, and the way the diffs are done
[11:16] <_lance_> Let me pull up the stubs
[11:16] <gvwilson> @entire-back-end-team the coming week will need to be a lot more productive... :-(
[11:17] * cvan (~cvan@c-76-102-14-98.hsd1.ca.comcast.net) Quit (Quit: cvan)
[11:17] <_lance_> I don't know how the hg team is going to represent the diff
[11:17] <_lance_> The method string is very vague
[11:17] <colin_m> Is there a reason why the representation should be different between HG and SVN?
[11:19] <_lance_> I don't think so
[11:19] <colin_m> What about the pathing? Why the difference?
[11:19] <_lance_> Oh, I think both teams are changing the diff signature to allow for a second file to be input
[11:19] <colin_m> So we're going to allow arbitrary diffs?
[11:19] <gvwilson> can anyone from the Hg team speak to that?
[11:20] <_lance_> well, sort of- only because it makes it easier to diff renamed files
[11:20] <jyeung> From the discussion with _lance_ is not really arbitrary diffs that we want
[11:20] <_lance_> we don't intend to be able to diff a.txt with b.txt, but a.txt with a_new_name.txt
[11:20] <jyeung> yes to support diff renames for svn and hg
[11:21] <_lance_> since the code used to track down the old name of a renamed file ought to be in a different method
[11:21] <gvwilson> what happens in Svn and/or Hg if you ask the VCS to diff two things that aren't related? because our interface should just display whatever they do
[11:22] <mi_sa> doesn't that mean that the user has to know both names?
[11:22] <colin_m> Well presumably they'd be getting at it from the file history view, so they wouldn't be manually entering the filenames.
[11:23] <colin_m> They'd be selecting radio buttons.
[11:23] <mi_sa> and what happens in the revision log of the file tha has been renamed? Does it show all or just since after being renamed?
[11:24] <_lance_> between two unrelated files, I'm not sure, I don't know what svn diff would output with -old and -new pointing to unrelated urls
[11:24] <colin_m> I think it would make sense to include the history pre and post rename in the history view.
[11:24] <_lance_> there's a svn renamed history command
[11:24] <gvwilson> these issues should all have been resolved by now --- we should be writing and testing code, not figuring out what our corner cases (much less how to deal with them)
[11:24] <gvwilson> would a face-to-face session one evening next week help move things along?
[11:25] <_lance_> yes
[11:25] <mi_sa> i agree
[11:25] <jyeung> yes
[11:25] <colin_m> sure
[11:25] <jharjono> yep
[11:25] <Joel_C> sounds good
[11:25] <gvwilson> Tuesday June 22 starting at 6:00? can the IBMers get downtown in time for that?
[11:26] <gvwilson> or does it have to be 6:30?
[11:26] <nikita> I am okay with 6
[11:26] <_lance_> That time works for me
[11:26] <jyeung> Where are we meeting?
[11:26] <Joel_C> good for me
[11:26] <jyeung> Bahen?
[11:26] <mi_sa> 6:30 is better for me
[11:26] <gvwilson> bahen seems the logical place --- G20 madness won't start 'til the next day
[11:26] <_lance_> robarts?
[11:27] <DaveWeber> im good for 6
[11:27] <_lance_> where can we meet in bahen since the great hall is closed?
[11:27] <gvwilson> @mi_sa if we split the difference, can you be here for 6:15?
[11:27] <gvwilson> @_lance_ i have magic powers
[11:27] <mi_sa> i can leave early, it's not really that big of a problem
[11:27] <mi_sa> if jyeung can give me a ride to the station :)
[11:27] <mi_sa> or schol
[11:27] <gvwilson> OK, let's make it 6:00 pm
[11:27] <jyeung> mi_sa I will be going to no worries.
[11:28] <colin_m> Okay, sounds like that's settled.
[11:28] <gvwilson> I've booked Bahen 5256 for 6:00-8:00 pm
[11:28] <gvwilson> I'll provide snacks, but if you want to grab a sub or something and bring it to stay fueled, please do so
[11:29] <DaveWeber> [What sort of snacks are we talking about?]
[11:29] <gvwilson> Timmies
[11:29] <colin_m> Moving briskly along...
[11:29] <colin_m> Before we decide on our homework for next week, I'd like to bring up the idea of team reshuffling. Right now the front-end team comprises me and Misa. Noam can't work on the project right now because he has exams. Because this constitutes a bit of an imbalance in team size, and because I think the front-end's work is highly parallelizable, I'd like to suggest we move one or two people from one of the back-end teams to the front-end team. Thoughts
[11:29] <colin_m> ? Volunteers?
[11:30] <gvwilson> does this mean someone from the front end will move to the back end?
[11:30] <gvwilson> and if so, which one?
[11:30] <mi_sa> I am switching
[11:30] <colin_m> Well, that would mean we'd still just have two people in the front-end team.
[11:30] <mi_sa> I emailed you
[11:30] <colin_m> Which is pretty lonely.
[11:30] <gvwilson> :-)
[11:31] <gvwilson> @DaveWeber, interested in doing some front-end work?
[11:31] <gvwilson> or @jharjono ?
[11:31] <colin_m> Sorry, so mi_sa, you're switching to the back-end?
[11:31] <mi_sa> yes
[11:31] <DaveWeber> I'm all for it, but I'm terrible at UI design/graphics so I don't know how useful I'll be
[11:31] <_lance_> can I also get a hg team member delegate to stick with me post-meeting to get some more details hammered out before tuesday?
[11:32] <gvwilson> johan?
[11:32] <colin_m> Okay, then I think we should aim for at least two ship-jumpers.
[11:32] <jharjono> hmm i just messed up some UI stuff at work big time, so I think i'm really bad at UI
[11:32] <jyeung> _lance_ Me
[11:32] <_lance_> okay
[11:32] <gvwilson> @jharjono @DaveWeber well, one of the reasons for this course is to give you a chance to practice things when your paycheck isn't on the line :-)
[11:32] <mi_sa> Colin, Noam's exams are going to be over soon
[11:32] <mi_sa> i don't think it will affect too much
[11:32] <gvwilson> how about both of you move to front end to get some practice, and mi_sa moves to the SVN back end?
[11:33] <gvwilson> anyone desperately unhappy with that?
[11:33] <mi_sa> do we need 4 ppl on the front-end?
[11:33] <colin_m> We can move someone back when Noam comes back to us, if necessary.
[11:33] <DaveWeber> I'm good with it
[11:33] <jharjono> same
[11:33] <colin_m> Like I said, I think the front-end work is highly parallelizable.
[11:33] <colin_m> Okay, good. The next order of business is what we're going to aim for for next week.
[11:34] <colin_m> Front-end team: What do we think our priority should be? What's a concrete goal we could set ourselves?
[11:34] <colin_m> Here's what I said (I cheated and wrote it out ahead of time): I think we should have a working Pinax app that generates at least one of the browser's views. It doesn't have to look pretty, but it should be generated dynamically (using proper view functions, templates, and mocked-up back-end calls) rather than just a static HTML page like the ones we made two weeks ago. In addition to that, we should look at some existing Pinax applications and h
[11:34] <colin_m> ave an understanding of how a Pinax app differs from a regular Django app (I think this primarily boils down to what it means for an app to be 'group aware').
[11:35] <gvwilson> p.s. http://www.cs.utoronto.ca/~gvwilson/ca-plane-pour-moi.mp3 for those who want to listen along...
[11:35] <colin_m> But it sounds like gvwilson had in mind more UI-oriented stuff?
[11:35] <gvwilson> @colin_m we *must* have a complete and consistent set of function/method calls traced out for each back end in response to every HTTP request the front end is going to send
[11:36] <gvwilson> and we *must* have a complete set of prototypes of the front end so that we can start user testing
[11:36] <gvwilson> if this were a regular course, I'd say "...and they're worth 25% of your mark each", but it would be fair to drop that on you on such short notice
[11:36] <colin_m> My understanding, based on a post from the pinax mailing list (which I may be taking out of context), is that the front-end of a Pinax app is kind of peripheral.
[11:37] <colin_m> This is the thread: http://groups.google.com/group/pinax-users/browse_thread/thread/57eb361ce9ba3111/b4d385d838b57a48?lnk=gst&q=JoeCodeswell#b4d385d838b57a48
[11:37] <colin_m> (talk about horrible URL schemes)
[11:37] <gvwilson> the former is the only thing that will give us confidence we've got a working design for the glue between our app and our back ends
[11:37] <colin_m> Relevant quote from James Tauber: Templates are almost always site-specific, or at least theme-specific.
[11:37] <colin_m> So we provide default templates in a theme (and hope to provide more themes) but expect site developers will override them for their needs.
[11:37] <colin_m> There may be cases where template fragments make sense in an app, but page-level templates in an app means the app making too many assumptions about the structure of a site and the layout of its pages.
[11:37] <gvwilson> (apologies --- have to run to another meeting that was supposed to start 7 minutes ago --- @colin_m, please post a summary to the Basie blog when we're done)
[11:37] <gvwilson> (I'll try to get back online while in my other meeting)
[11:38] <gvwilson> thanks everyone --- look forward to seeing you on Tuesday
[11:38] * gvwilson (~chatzilla@red-gw28.cs.toronto.edu) has left #basie
[11:38] <colin_m> Okay, well, anyone have an opinion on what the front-end team should aim for?
[11:38] <Joel_C> well, the views aren't front-end code
[11:39] <Joel_C> so I think this is more an issue of mocking up the core UI features so they can be tested
[11:39] <colin_m> So is the back-end implementing the API and the views?
[11:39] <nikita> and no python code is allowed in templates, so we are good
[11:40] <colin_m> Okay, then, front-end team: Do we think the current mock-ups are good enough to do usability testing with?
[11:41] <colin_m> I'm going to say 'yes, with some sanding around the edges'
[11:41] * bwinton_away is now known as bwinton
[11:42] <colin_m> Okay, let's say the front-end team's task is to get the mock-ups in a state that's ready to do demos/testing.
[11:42] <colin_m> And to find a testing audience.
[11:42] <DaveWeber> Sure
[11:42] <colin_m> Back-end team, what do you think your priority should be, and what (concrete) goal do you think you could achieve by next week?
[11:42] <colin_m> *Back-end teams
[11:43] <_lance_> First I think we need a mini-deadline for Tuesday so we can work together
[11:43] <_lance_> And well, our priority is going to be sorting out the controller, so mapping the HTTP requests to actual method calls
[11:43] <Joel_C> also - have we solidified the method calls yet?
[11:44] <Joel_C> between HG and SVN?
[11:44] <jyeung> almost there..
[11:45] <colin_m> Okay, mini-deadline of Tuesday for the back-end teams to finalize their methods and inheritance hierarchies?
[11:45] <colin_m> What about for Friday?
[11:45] <colin_m> Do you think by then you can have started writing code?
[11:45] <jyeung> As for the http requests, are the urls going to specify which repository we are looking it like /svn/view/source/path/ or we are going to abstract it?
[11:47] <colin_m> Sounds like that's a decision you'll have to make this week
[11:48] <colin_m> Okay, let's say by Friday each of the back-end teams should have started implementing at least one class?
[11:49] <colin_m> Sounds like there are no objections.
[11:49] <mi_sa> what exactly are we discussing on tuesday
[11:49] <jyeung> thats reasonable
[11:49] <colin_m> Before we adjourn, the floor is open.
[11:49] <_lance_> It'd make more sense to start with the controller
[11:49] <colin_m> Why do you say that?
[11:50] <_lance_> The front end guys need it to do something with
[11:51] <_lance_> We can have the backend spit out dummy data in the meantime
[11:51] <jyeung> I agree
[11:51] <colin_m> Okay then, let's say at least two views each by Friday.
[11:52] <colin_m> Floor is open. Any questions or comments before we go?
[11:52] <_lance_> Views?
[11:52] <mi_sa> what are we discussing on tuesday again?
[11:52] <colin_m> Yeah, isn't that what Django calls them?
[11:53] <_lance_> Well.. django is a Model/View/Controller framework so yes
[11:53] <_lance_> I asked about having a controller done and you mentioned views, so I got confused
[11:54] <colin_m> Sorry, what django calls 'views' are what most other MVC frameworks call the controller.
[11:54] <colin_m> At leas that's my understanding.
[11:54] <DaveWeber> I have to go - lunch time. See you all on tuesday!
[11:54] <jharjono> Model/View/Template
[11:54] <DaveWeber> have a great weekend everyone
[11:54] <jharjono> view is django term for controllers, and template is django term for views, confusing eh :)
[11:54] <_lance_> later Dave
[11:55] <_lance_> Oh, okay
[11:55] * DaveWeber (~chatzilla@nat/ibm/x-gmfodktafztzyjbh) Quit (Quit: ChatZilla 0.9.86 [Firefox 3.5.9/20100315083431])
[11:55] <colin_m> Before we go, I have a question about Review Board.
[11:55] <colin_m> For those of us who aren't familiar with it, could we get a very brief explanation of how it works?
[11:55] <colin_m> And is it necessary for every commit to go through it?
[11:55] <Joel_C> yes that would be helpful
[11:56] <colin_m> nikita? This seems to be your bag
[11:57] <nikita> yep
[11:57] <nikita> one sec
[11:57] <Joel_C> have to go for lunch. bye everyone
[11:58] <colin_m> Bye
[11:58] <nikita> so review board is just a tool to review your code
[11:58] <nikita> or actually anything you want to commit to svn
[11:59] <nikita> this goes for the initial commit and all subsequent commits
[11:59] <nikita> the reason we have it in place
[11:59] <_lance_> What about non-code resources being committed?
[11:59] <nikita> is because it saves a lot of time and prevents us from commiting buggy code
[12:00] <_lance_> And what about committing to the sandbox?
[12:00] <nikita> the reason I asked you to put the diagram on the RB is because it gives others to review and comment on it
[12:00] <colin_m> That was going to be my question. Is it not possible for commits to the sandbox to bypass the review process, if they're just intended to be provisional?
[12:00] <nikita> before the final diagram will go into svn
[12:00] * cvan (~cvan@nat/mozilla/x-jirjtddvfbmulvrj) has joined #basie
[12:01] <nikita> even though this particular folder is called a sandbox, we started working on the project and from now on all commits should go through the review board
[12:01] <colin_m> I feel like if every commit has to be reviewed, we're going to fall into the habit of e-mailing code back and forth (which my team has already been doing) and no good can come of that.
[12:01] <nikita> people still have an opportunity to get the patch and see your diagram
[12:01] <nikita> tell you what they think is missing
[12:01] <nikita> let you fix it
[12:01] <nikita> and reupload the next patch
[12:02] <nikita> plus it sort of makes you the owner of the diagram - changes go through you, and it's not like everyone is doing work on it
[12:03] <nikita> when you assign work to a group of people it's nice to have one person responsible for a piece of work within the team
[12:03] <colin_m> Okay, it's past noon, so if you guys have other places to be, I won't keep you.
[12:03] <colin_m> (But if you want to stay and continue the discussion, feel free)
[12:04] <nikita> i am at work, I will stick around if there are any questions, otherwise i am accessible by email :)
[12:05] <colin_m> nikita, is it possible to have a place in the repository where it's possible to commit code that is explicitly _not_ bug-free, with the understanding that it's a work in progress?
[12:05] <nikita> create your own branch, if you want
[12:05] <colin_m> Oh, okay. Where should we do that?
[12:06] <colin_m> Directly under the sandbox directory?
[12:06] <nikita> under branch directory in the repo ;D
[12:06] <colin_m> Oh, okay
[12:07] <_lance_> sounds like we should have input from the front end team about whether it's feasible for them to tell us which vcs we're pulling from
[12:07] <_lance_> so can we have a request like /diff/svn/file1/etc
[12:07] <_lance_> or just /diff/file1/etc/
[12:08] <colin_m> Is it reasonable to expect that a project will have 1 or less repositories?
[12:08] <_lance_> no
[12:08] <colin_m> Then I would say it's necessary to specify it in the URL
[12:08] <_lance_> okay
[12:08] <colin_m> Actually, if we allow more than one repo of the same type, don't we need an additional identifier?
[12:09] <colin_m> Because what if a project has two SVN repos with a /trunk directory? The url /view-source/svn/trunk becomes ambiguous.
[12:10] <mi_sa> have the project name in the url?
[12:10] <_lance_> well, if the project has two svn repos..
[12:10] <colin_m> Well it would be, but _lance_ is saying a project can have more than one repository.
[12:10] <mi_sa> oh sorry misread
[12:12] <mi_sa> when projects set up their repository for browsing, you can ask them to give the repository a name
[12:12] <mi_sa> and that name could be in the url
[12:12] <colin_m> I'm trying to imagine why one project would need more than one repository of the same type, and I'm drawing a blank.
[12:13] <_lance_> One repo for the students, one repo for the project they're pulling from
[12:13] <_lance_> Like when we were working on selenium and used svn, and selenium had svn
[12:14] <colin_m> So you're saying you'd want to be able to browse Selenium's svn repo from within your project?
[12:14] <colin_m> As well as the project's repo?
[12:14] <_lance_> Yes, especially to do things like diff across repos
[12:14] <_lance_> So that we can send them a patch
[12:15] <colin_m> My intuitive feeling is that the project should somehow have ownership of the repo(s) viewable under its browser.
[12:15] <colin_m> But I'm really not sure one way or the other.
[12:17] <_lance_> Afk, making lunch and checking on dinner
[12:28] <jyeung> _lance_ so what should vcsDiff contains?
[12:29] <_lance_> from svndiff i like the idea of a list of tuples
[12:30] <_lance_> where a 1-tuple means no difference, and a 2-tuple will contain the contents of file A in the first half, contents of B in the second half, etc
[12:33] <jyeung> I see
[12:33] <_lance_> what about get_download?
[12:34] <jyeung> and we'll be merging the file and directory together. No reason no to
[12:36] <jyeung> for get download you mean in the repo or in the revision?
[12:36] <_lance_> both
[12:37] <_lance_> can output a string containing binary data, and spit that out in the http response
[12:37] <_lance_> or we can just spit out a URL as a string
[12:39] <jyeung> I like the first one but how much more work is require?
[12:39] <_lance_> about the same
[12:40] <_lance_> hm.. though might not be a good idea if the diff is really big
[12:42] <jyeung> for revision should we change the get_download to like get_diff or something?
[12:42] <_lance_> don't think so.. might want to get the entire project at that revision
[12:43] <jyeung> there does diff comes in?
[12:44] <_lance_> get_patch ?
[12:44] <jyeung> you said that it might not be a good idea if the diff is really big when we are talking about get_download
[12:44] <_lance_> oh
[12:44] <_lance_> the file was really big i meant
[12:45] <jyeung> oh okay okay
[12:46] * nikita (~jango@dsl-173-248-219-4.acanac.net) Quit (Remote host closed the connection)
[12:46] <jyeung> for the hprepo should add is_exists() or file_exists()
[12:46] <jyeung> with parameters of the file? and revsion?
[12:47] <_lance_> file_exists I'd say
[12:49] <jyeung> for get_dir show it would be something like get_dir(path, revno) and return a RepoFile etc right?
[12:50] <_lance_> a list of HgFiles I'd imagine
[12:52] <jyeung> ahh yes I got it. for the UI to display the list of dir and files
[12:52] <jyeung> for SvnRepo what is get_properties returning?
[12:55] * nelle_ (~nelle@91.104.45.150) has joined #basie
[12:59] <_lance_> that's svn specific
[12:59] <_lance_> result of svn properties command
[13:04] <jyeung> I see. for FileConstraits should the File in SVNFileConstrainit a SvnFile?
[13:06] <_lance_> Yes, it has to be
[13:11] <jyeung> looking from the class diagram. what else is missing? other than VCSDiff the method parameters?
[13:12] <_lance_> not sure
[13:17] <_lance_> We still need to talk about the mapping of http requests to methods too
[13:17] <_lance_> though not sure if i want to talk about that right now
[13:18] <jyeung> we could
[13:19] <jyeung> but can we assume 1 repo per projecT?
[13:19] <jyeung> and they could have different projects?
[13:19] <jyeung> like if they want more than 1 repo they need to recreate another project
[13:25] <_lance_> i vote for no but i guess we can ask around
[13:28] <jyeung> k. that will change the controller design.
[13:28] <jyeung> whats the advantage of having muti repo for a single project?
[13:29] <jyeung> ahh I see nvm that
[13:29] <_lance_> well, like in - yeah :)
[13:29] <_lance_> it would have helped in 302 when giving patches back to mainline
[13:29] <_lance_> or importing patches from mainline
[13:30] <jyeung> yes. but i hope not mutiple of different types of repo
[13:30] <_lance_> yes, multiple different types of repos.. at least for the browser
[13:30] <_lance_> but not necessarily for things like diff
[13:31] <_lance_> although.. hm.. you can't exactly just do 'svn diff' between two different repos, can you?
[13:31] <jyeung> i am not sure
[13:32] <jyeung> we could have something like /thunder/view-source/r.554 etc. where "thunder" is the game of the repo given during setup. but multiple repo wasnt talked about during our mockup
[13:32] <_lance_> maybe we could just use difflib
[13:34] <jyeung> yes for creating the diff.. but is multiple repo per project a MUST have for the initial version or that can be added later for additional feature?
[13:35] <_lance_> i don't think it's essential for version 0.1
[13:35] <_lance_> and apparently yes you can diff from two arbitrary repos using svn diff command
[13:38] <jyeung> so how about we do 1 repo per project for now and add it later in future releases?
[13:39] <_lance_> yeah, but try not to leave anything that would make it difficult down the road
[13:39] <jyeung> yes I agree
[13:40] <jyeung> as for the http request implicit or explicit svn/hg urls?
[13:41] <_lance_> i don't know. i don't see either one being particularly bad for svn.
[13:42] <_lance_> and it's something that i don't necessarily see the need for consistency on
[13:46] * colin_m (~colin@whit179.ucres.utoronto.ca) Quit (Ping timeout: 260 seconds)
[13:57] * Nix (~Nix@opensuse/member/peternixon) has joined #basie
[13:58] <Nix> hi guys
[14:03] <_lance_> hey
[14:04] <jyeung> hi
[14:04] <jyeung> we will have to set the controller to handle the request and decide which function to call either the hg or the svn
[14:05] * jharjono (~chatzilla@nat/ibm/x-pfjmfvponjeyenhu) Quit (Quit: ChatZilla 0.9.86 [Firefox 3.6.3/20100401080539])
[14:13] <jyeung> we can specific what kind of repository it is in the project settings and let the controller handles it. or put it in the url?
[14:55] * Nix (~Nix@opensuse/member/peternixon) Quit (Ping timeout: 272 seconds)
[15:29] * mi_sa (~chatzilla@nat/ibm/x-jvewhwxgzlubeqdl) Quit (Quit: ChatZilla 0.9.86 [Firefox 3.5.9/20100315083431])
[15:47] * Nix (~Nix@opensuse/member/peternixon) has joined #basie
[15:56] * Nix (~Nix@opensuse/member/peternixon) Quit (Ping timeout: 265 seconds)
[16:05] * Nix (~Nix@opensuse/member/peternixon) has joined #basie
[16:36] * bwinton is now known as bwinton_away
[17:07] * jyeung_ (~chatzilla@CPE001d73111113-CM000a7366632e.cpe.net.cable.rogers.com) has joined #basie
[17:08] * jyeung (~chatzilla@CPE001d73111113-CM000a7366632e.cpe.net.cable.rogers.com) Quit (Ping timeout: 260 seconds)
[17:08] * jyeung_ is now known as jyeung
[17:31] * bwinton_away is now known as bwinton
[17:36] * bwinton is now known as bwinton_away
[17:51] * Nix (~Nix@opensuse/member/peternixon) Quit (Remote host closed the connection)
[17:58] * hazure (~cjmuise@red-gw37.cs.toronto.edu) Quit (Quit: Leaving)
[19:07] * nelle_ (~nelle@91.104.45.150) Quit (Ping timeout: 258 seconds)
[19:08] * nelle_ (~nelle@91.106.48.23) has joined #basie
[19:39] * nelle_ (~nelle@91.106.48.23) Quit (Ping timeout: 258 seconds)
[20:04] * cvan (~cvan@nat/mozilla/x-jirjtddvfbmulvrj) Quit (Quit: cvan)
[20:04] * cvan (~cvan@nat/mozilla/x-ycgtcsdmqctkwhaj) has joined #basie
[20:12] * cvan (~cvan@nat/mozilla/x-ycgtcsdmqctkwhaj) Quit (Quit: cvan)
[23:58] * colin_m (~colin@whit179.ucres.utoronto.ca) has joined #basie
These logs were automatically created by BasieBot on irc.freenode.net using the Java IRC LogBot.
