[Gsoc] Vamsi or Andres: how goes it?
Vamsi Krishna Davuluri
vdavuluri at acm.org
Wed Jul 29 12:25:24 EDT 2009
On Wed, Jul 29, 2009 at 9:36 PM, Vamsi Krishna Davuluri <
vamsi.davuluri at gmail.com> wrote:
> Sorry for the absence. Andres and myself have been meeting on #sugar every
> other day.
> I had been travelling these past few days ( back home). So I couldn't
> concentrate on coding
> for about 4 days.(the food is getting to me too :D). I'm back in action, a
> couple of hours with Andres
> should fix these things:
> On Wed, Jul 29, 2009 at 8:44 PM, Jameson Quinn <jameson.quinn at gmail.com>wrote:
>> I am concerned about the status of the printing project for GSoC. The last
>> update on http://materializingsweetness.wordpress.com/ was a month ago,
>> and the last word from Vamsi on any mailing list I'm on was 3 weeks ago. If
>> there are any hangups or other issues with the project, we need to know
>> about it ASAP. Vamsi and Andres, please respond to let us know.
>> I myself am still away from my own house, so my own connection is
>> intermittent. If there is a response to this email, I'd appreciate it if
>> others could help deal with any issues that are raised, though of course
>> I'll do my best as well.
>> ps. Below is my log from yesterday's meeting.
>> [07:00] <homunq__> Hello, have I timed this right?
>> [07:01] <silbe> i'd say so
>> [07:01] <homunq__> (I'm not quite sure if my UTC calculations are correct
>> from here...)
>> [07:01] <homunq__> Anybody else want to run this meeting
>> [07:01] <homunq__> ?
>> [07:01] == m_stone [n=mstone at 126.96.36.199] has joined #sugar-meeting
>> [07:01] <homunq__> if not, I'll start in 5.. 4...
>> [07:01] <homunq__> #startmeeting
>> [07:02] <silbe> unless there was a summer/winter time changeover the last
>> few weeks it should be correct
>> [07:02] <homunq__> um, no bot
>> [07:02] <homunq__> OK. Roll call first... silbe bemasc lucian mchua ....
>> [07:03] <homunq__> that is not great attendance.
>> [07:03] * mchua listening
>> [07:03] * lucian is here
>> [07:03] <homunq__> but I can't complain, as I missed last week myself.
>> [07:03] <silbe> last week wasn't really better :-/
>> [07:04] * bemasc is here
>> [07:04] <homunq__> OK, well, let's start anyway.
>> [07:04] == subzero [n=Miranda at 188.8.131.52] has joined #sugar-meeting
>> [07:04] <silbe> who do you want first?
>> [07:04] <homunq__> welcome subzero
>> [07:04] <subzero> hi guys
>> [07:05] <silbe> ok, i'd suggest subzero to go first
>> [07:05] <homunq__> any word from bryan if he can come?
>> [07:05] <subzero> homunq__: hi
>> [07:05] <silbe> IIRC he's always having network issues
>> [07:05] <homunq__> OK, subzero, you're on.
>> [07:05] <subzero> homunq__: I talked with him yesterday,, I think he won't
>> come, he is not in nepal
>> [07:05] <subzero> homunq__: ok
>> [07:05] <homunq__> link and progress?
>> [07:06] <subzero> I haven't posed yet (I'll do it today), I finished
>> workin really morning
>> [07:06] <subzero> but you can see my progress on git
>> [07:07] <subzero> I'm currently working around i18n
>> [07:07] <subzero> and the karma plugin
>> [07:07] <subzero> I have dome out karma.localisation version and linked it
>> with Gettext
>> [07:08] <subzero> s/out/our
>> [07:08] <subzero> git.sugarlabs.org/projects/karma
>> [07:09] <subzero> that's what I'm doing, and finally we have a
>> jquery.karma.js (the plugin)
>> [07:10] <homunq__> do you find the standard GetText search algorithm
>> sufficient (ie, where it looks for translations)?
>> [07:10] <subzero>
>> [07:10] <subzero> nope
>> [07:10] <subzero> It won't carry the <img>bla bla </img> elements
>> [07:11] <subzero> I'm in the procress to add <img> supports
>> [07:11] <homunq__> oh, that's not exactly what I was asking, but
>> [07:11] <subzero> also
>> [07:11] <subzero> there are two things
>> [07:11] <subzero> 1. i18n for canvas content
>> [07:11] <subzero> 2. i18n for karma content (outside canvas)
>> [07:12] <homunq__> I meant, not where (text) it finds text for
>> translating, but where (files and directories) it finds the translated
>> [07:12] <subzero> you men something like /dir/dir/file ?
>> [07:13] <homunq__> yeah, gettext's algorithm has seemed too inflexible to
>> me in the past, so I was curious if it worked for you.
>> [07:13] <subzero> (btw this is the last version jquery.karma.js:
>> [07:14] <subzero> nope, it work's for Image an Sound (inside canvas)
>> [07:14] <subzero> s/work's/works/
>> [07:15] <subzero> the second point is the problem (outside canvas)
>> [07:16] <subzero> the truth is that I have to modify the Gettext file (I
>> have done some changes)
>> [07:17] <homunq__> that kind of change is the kind of thing that doesn't
>> exist until it's fully documented.
>> [07:18] <subzero> I'm agree. until now all the changes I've done to
>> Gettext.js are in order to link karma plugin with it
>> [07:18] <subzero> I will be working on that all this week
>> [07:18] <homunq__> OK. Is there anything you need more feedback on?
>> [07:19] <subzero> homunq__: well, I have a basic version of layers within
>> [07:20] <homunq__> so, you want more people to try it out, but no specific
>> [07:21] <subzero> I'm testing redrawing all the canvas tree (a bunh of
>> drawings in z-order) vs several canvases
>> [07:21] <subzero> well bascically I'm deciding what is the best: to cosume
>> processor or to consume memory ?
>> [07:22] <subzero> opinions? (related to the xo)
>> [07:22] == dwmw2_gone has changed nick to dwmw2
>> [07:22] <bemasc> depends how big your canvas is
>> [07:23] <bemasc> canvas is 32-bit RGBA. 1200*900*4 = 4.3 MB per image at
>> fullscreen on an XO.
>> [07:23] <subzero> bemasc: uuh, thats a good point
>> [07:23] <homunq__> I think we have to move on
>> [07:23] <bemasc> so keeping 10 around would probably be a bad idea, but 3
>> might be ok.
>> [07:24] <subzero> the problem around a simple canvas is that it will do a
>> lot of math
>> [07:24] <homunq__> not to belittle this question
>> [07:24] <subzero> homunq__: I understand, thanks for reading
>> [07:24] <homunq__> bemasc, do you want to go next?
>> [07:25] <bemasc> ok
>> [07:25] <subzero> bemasc: thanks for your advice
>> [07:26] <bemasc> I released a new version of SharedTextDemo.
>> [07:26] <bemasc> It's faster. I wrote some very fancy data structures.
>> Now I'm looking for actual use cases.
>> [07:27] <bemasc> I think I'm likely to start by adding live shared editing
>> to Pippy.
>> [07:27] <bemasc> That's all.
>> [07:28] <homunq__> OK, cool
>> [07:28] <homunq__> Silbe, let's go next.
>> [07:28] <silbe> ok
>> [07:28] <lucian> somewhat related, is psyco desirable on an XO ?
>> [07:28] <lucian> (sorry silbe, go ahead)
>> [07:29] <silbe> i have more or less finished changing sugar* to use the
>> new datastore API
>> [07:29] <silbe> started working with Tomeu on getting my changes into
>> [07:30] <silbe> some small fixes are already in, for the larger piece i've
>> submitted for review he has requested some changes (mostly stylistic)
>> [07:31] <homunq__> links?
>> [07:31] <silbe> need to work on merging Tomeus latest changes (tree-view
>> etc.) back into my tree so i get a working version again (the metacity stuff
>> broke a lot)
>> [07:32] <silbe>
>> http://wiki.sugarlabs.org/go/Version_support_for_datastore/Progress(still updating it, shame on me!)
>> [07:34] <homunq__> OK.
>> [07:34] <homunq__> Do you need feedback from anybody besides Tomeu right
>> [07:35] <silbe> homunq__: i don't think so. of course suggestions from
>> anyone else are welcome, but i don't block on them.
>> [07:36] == sdziallas_ [n=sebastia at p5B047270.dip.t-dialin.net] has joined
>> [07:36] <homunq__> OK, then I think you're done. (To say "we" at this
>> point would be hubris. Good job.)
>> [07:37] <homunq__> Lucian?
>> [07:37] <lucian> homunq__: here
>> [07:37] <homunq__> You're on. Link and summary.
>> [07:37] <lucian> i haven't gotten almost any word done the past few days,
>> there are some complications about my accomodations for the next year
>> [07:38] <silbe> homunq__: i don't think i'm really done yet, but thanks
>> anyway. :)
>> [07:38] <lucian> http://honeyweb.wordpress.com
>> [07:38] <homunq__> silbe: I didn't mean done with the project, I meant
>> done with your part of this meeting.
>> [07:38] <homunq__> If you disagree, definitely speak up.
>> [07:39] == sdziallas [n=sebastia at fedora/sdziallas] has quit [Nick
>> collision from services.]
>> [07:39] == sdziallas_ has changed nick to sdziallas
>> [07:39] == m_stone [n=mstone at 184.108.40.206] has quit [Read error: 104
>> (Connection reset by peer)]
>> [07:39] <lucian> i pretty much have the SSB done
>> [07:39] <silbe> homunq__: oh, ok. :)
>> [07:39] <lucian> all future work is related to fixing ugly bugs (Gears
>> warning, data URI downloading)
>> [07:40] == m_stone [n=mstone at 220.127.116.11] has joined #sugar-meeting
>> [07:40] <lucian> with stuff like window.sugar.alert
>> [07:40] <bemasc> don't forget gears-journal integration!
>> [07:41] <lucian> bemasc: that should mostly work right now, i just can't
>> test it properly
>> [07:41] <bemasc> nice.
>> [07:41] <lucian> there is one major enough area that i haven't worked on
>> enough, collaboration
>> [07:41] <homunq__> why can't you test it?
>> [07:41] <lucian> homunq__: because of the Gears security warning bug, you
>> can't accept it
>> [07:42] <homunq__> oh. that's a bad bug.
>> [07:42] <lucian> homunq__: i think i have a hack for that, but for some
>> reason the buttons don't show in my jhbuild
>> [07:42] <subzero> lucian: the freaky winwdow without buttons?
>> [07:42] <lucian> subzero: yes
>> [07:42] <bemasc> lucian: Google Docs is pretty good collaboration. So is
>> [07:43] <lucian> bemasc: but it's server-side. i want to offer some
>> collaboration options for client-only html apps
>> [07:43] <lucian> bemasc: like karma apps
>> [07:43] <lucian> bemasc: XHR isn't well suited at all for this
>> [07:44] <lucian> i think i should do like gears and simply offer a JS api
>> for it, that uses the existing Sugar collaboration APIs in the background
>> [07:44] <homunq__> so the Gears bug is in Gears itself... are you in touch
>> with your upstream?
>> [07:44] <lucian> homunq__: no, it's a pyxpcom bug
>> [07:45] <bemasc> lucian: If I were doing this, I would separate out the
>> collaboration logic, hide it behind a server running on localhost, and then
>> use XHR.
>> [07:45] <lucian> homunq__: tomeu found it and he has a fix, but it's not
>> included yet
>> [07:45] <lucian> bemasc: the main problem with that is that you need a way
>> to find the localhost server
>> [07:46] <lucian> bemasc: collaboration over the mesh works nicely in sugar
>> right now
>> [07:46] <lucian> bemasc: i'm not sure i even can duplicate it in JS with
>> [07:46] == sdziallas_ [n=sebastia at p5B047123.dip.t-dialin.net] has joined
>> [07:47] == sdziallas [n=sebastia at p5B047270.dip.t-dialin.net] has quit
>> [Nick collision from services.]
>> [07:47] == sdziallas_ has changed nick to sdziallas
>> [07:48] <lucian> bemasc: if i offer just stuff like window.sugar.send and
>> let Browse handle the rest (like it does with bookmarks), it should be very
>> simple to implement
>> [07:48] <bemasc> lucian: well, I'll be curious to see what you come up
>> [07:48] <homunq__> OK, sounds good.
>> [07:48] <homunq__> Anything more to add?
>> [07:49] <lucian> not really
>> [07:49] <homunq__> Has anyone seen Vamsi or Andres?
>> [07:51] <homunq__> I take that as a no.
>> [07:51] <homunq__> Since my connectivity is really poor, I'd like help
>> tracking them down.
>> [07:52] <homunq__> (I'm actually in #@$$!$ vista right now for this
>> meeting, and my step-father-in-law is breathing down my neck, that's how bad
>> I have it.)
>> [07:52] <silbe> homunq__: how about just sending them an email?
>> [07:53] <homunq__> 'silbe: I will do that, and I'll copy the list.
>> [07:53] <silbe> sounds good
>> [07:54] <homunq__> Heh... I guess that one advantage of poor attendance is
>> that the meetings wrap up in time.
>> [07:55] <homunq__> subzero: I guess that means you have 5 more minutes to
>> finish what I cut you off about :)
>> [07:55] <silbe> not sure the marketing team meeting is going to happen
>> [07:55] <silbe> the last few times it wasn't
>> [07:57] <homunq__> :)
>> [07:57] == walterbender [n=chatzill at 18.104.22.168] has joined
>> [07:58] <homunq__> subzero, you were saying a simple canvas does a lot of
>> [07:58] <subzero> yep
>> [07:59] <subzero> what I'm doing is to create a elements-tree
>> [07:59] <subzero> and redrawing all the tree from time to time
>> [07:59] <subzero> this way we have z-index
>> [08:00] <lucian> subzero: i'd really go for multiple canvases, if only for
>> the cleaner separation
>> [08:00] <subzero> yeah, I thought it was a better way, but Bemasc is rght
>> about the huge size
>> [08:01] <subzero> in multiple canvases
>> [08:01] <silbe> subzero: IIUC you won't get hardware acceleration as well
>> [08:02] <silbe> subzero: that might suck on slow processors
>> [08:02] == lucian [n=lucian at sunjammer.sugarlabs.org] has quit
>> [08:02] <subzero> yes, I don't have hardware acceleration
>> [08:03] <homunq__> OK, I have to go right now. I'll write that email later
>> [08:03] <homunq__> Thanks for attending, people.
>> [08:03] <subzero> the solution needs to be suitable for XO
>> [08:03] <silbe> subzero: is xulrunner able to provide accelerated drawing
>> at all?
>> [08:04] <subzero> silbe: afaik, no
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the GSoC