Re: [Sugar-devel] Gtk 3.10 icon size regression

2014-01-19 Thread James Cameron
On Sat, Jan 18, 2014 at 12:07:34PM +0100, Daniel Narvaez wrote:
> On 18 January 2014 11:32, James Cameron  wrote:
> > On Sat, Jan 18, 2014 at 10:41:57AM +0100, Daniel Narvaez wrote:
> > > By the way, XO is our main platform from the number of users
> > > point of view but it seems like a dead end at this point. So IMO
> > > it should not be our main target when designing future
> > > solutions.
> > 
> > Yes, but not for the reason you give.  The OLPC XO is still in
> > production, and is still in deployment.  The best reason for not
> > selecting the OLPC XO as your main target is so that Sugar can be
> > suitable for as many platforms as possible.
> 
> Are new XO being sold? My possibly misguided feeling is that OLPC is
> not even trying anymore.

I'm not in that part of OLPC, I'm in engineering and manufacturing
support.  As far as I know, you can still order new XO laptops if you
have a large enough quantity in mind.  Any organisations wishing to
get a quote can write to countr...@laptop.org.

I may be misguided, but I didn't think Sugar Labs wants to buy any XO
laptops.  ;-) I think developers have enough XO laptops to work on, or
can share them around.

-- 
James Cameron
http://quozl.linux.org.au/
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel


[Sugar-devel] [ASLO] Release Physics-15

2014-01-19 Thread Sugar Labs Activities
Activity Homepage:
http://activities.sugarlabs.org/addon/4193

Sugar Platform:
0.96 - 0.100

Download Now:
http://activities.sugarlabs.org/downloads/file/28882/physics-15.xo

Release notes:
15

ENHANCEMENT:
* Add collaboration support for most tools

BUG FIX:
* Fixed problems with selector palettes


Sugar Labs Activities
http://activities.sugarlabs.org

___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel


Re: [Sugar-devel] mime type error in sugar

2014-01-19 Thread Gonzalo Odiard
On Sun, Jan 19, 2014 at 1:47 PM, Daniel Narvaez  wrote:

> On 19 January 2014 17:04, Gonzalo Odiard  wrote:
>
>> I don't know if this was needed by historical reasons,
>> maybe the Gio functionality was not available.
>>
>> mime.get_for_file do some magic to recognize text files.
>>
>
> Gio seems to handle that just fine (tested with Gio.content_guess_type).
>
>
>> Probably mime.get_for_file could call Gio instead of the SugarExt
>> equivalent.
>>
>
> IMO we should deprecate mime.get_for_file and replace it with Gio.
>


I agree.

This week I will be almost disconnected. Next week will be back working.

Gonzalo



-- 
Gonzalo Odiard

SugarLabs - Learning Software for children
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel


Re: [Sugar-devel] Collaboration support for sugar web activities

2014-01-19 Thread Emil Dudev
Please don't look at the poor javascript code :(

I created a web activity that can act as sugar's mesh for web activities
(with a few lines of code it should work for normal activities, too).

This time I went the other way around. Instead of sending invites, when an
activity is shared, everyone on the mesh can join it (it was easier to
implement). Normal invites could also be implemented, I just have to work
on integrating it with sugar's frame (for the invite notification).

Since this 'new mesh' is a web activity, it can be opened with a normal
desktop browser. Yet it will have little to no functionality. The simplest
explanation: the mesh can not know the path of an activity to open. With a
little work, this problem can be solved.

For the past 2 days I went deep into TogetherJS and I no longer believe
that it will suit sugar's needs.
The first reason is a WebKit bug. I noticed that when a CSS transform
property is used, the screen will go blank (gray). I manually edited
TogetherJS' jQuery modifications to not include CSS transforms, yet this
bug still happens. I have not yet found the problem. Although this is a
WebKit bug (it does not happen outside of sugar), it is one of the reasons
to drop TogetherJS.
The second one is the communication protocol. Although it is quite easy
(the server just replies the messages to all clients and lets them handle
everything), the protocol isn't much secure and may cause more traffic than
needed. I don't expect a kid to look at the code and 'hack' the protocol,
yet I still don't like the idea of leaving everything to the clients.
An other reason, although not much of a problem and more a feature, is the
fact that TogetherJS syncs every single mouse move, mouse click, form
change, input field update... These are lots of unneeded stuff.

As it was discussed earlier, a low-level API can work fine for sugar and it
won't be hard to create such an API. The thing that will be lost (and hard
to implement) are the many neat high-level features of TogetherJS.

About what I made and how to test it:
The sugar-net web activity is the 'mesh'. And a nice activity to test with
is the modified gtd-activity.
Once the sugar-net activity is started, it will communicate with a WebMesh
class in apisocket.py.
When a web activity (like gtd-activity) is shared, it will appear as a
shared activity when you hover the mouse over the XO icon in the sugar-net
activity.
When you click on it from an other mesh activity (an other instance of
sugar), it will automatically start the web activity (gtd-activity) and the
shared session will begin between the two.

Both sugar-net and gtd-activity still use Mozilla's hub server for the
communication.

sugar-toolkit-gtk3:
https://github.com/edudev/sugar-toolkit-gtk3/tree/webcollaboration
sugar: https://github.com/edudev/sugar/tree/webcollaboration
gtd-activity: https://github.com/edudev/gtd-activity/tree/collaboration
sugar-net: https://github.com/edudev/sugar-net

Emil Dudev



On Sat, Jan 18, 2014 at 4:20 PM, Emil Dudev  wrote:

> I'm not quite sure.
>
> One way to do it is to use 1 class as the client and integrate the class
> into the current mesh box (neighbourhood.py). This would probably mean that
> normal sugar activities could use this 'new protocol'.
> An other way would be to use a web activity as the client. I'm not sure
> how exactly will I have to handle the activity -> sugar communication and
> pass the invites.
>
> If going with the second approach I'd have to rewrite the whole
> neighbourhood view in HTML/javascript. And with this an other question
> arises: How much of sugar's core can be rewritten in HTML/javascript.
> One of TogetherJS's features is to follow your friends around pages. I
> think it's limited to a single domain, but I believe I can use it.
>
> I believe that the first approach would be easier, but with all this talk
> about new frameworks for web activities and going multi platform, I'll
> choose the second approach.
>
> Emil Dudev
>
>
>
> On Sat, Jan 18, 2014 at 1:19 PM, Daniel Narvaez wrote:
>
>> Interesting! Are you planning to use a js actitvity or a python one as
>> client?
>>
>>
>> On 18 January 2014 01:21, Emil Dudev  wrote:
>>
>>> Busy month...
>>>
>>> Anyway, I checked out telepathy and tried to fix the many protocol
>>> problems. Sadly, I gave up pretty easily. I suppose it would be easier to
>>> rewrite the whole process from scratch.
>>>
>>> Today I was able to recreate the basic functionality of TogetherJS'
>>> server. From a nodejs server, I rewrote it in python (using gwebsockets).
>>> This will give me the ability to customize it freely.
>>>
>>> Tomorrow I plan on building the basic invitation process using web
>>> sockets. I suppose it will be more reliable than telepathy. And it would
>>> possibly be better for an Android app or other systems with limited
>>> functionality.
>>>
>>> About WebRTC: I really think that peer-to-peer connections are needed.
>>> But at this point, I'll go with client-server co

Re: [Sugar-devel] [DESIGN] Better journal icons

2014-01-19 Thread Daniel Narvaez
Adding [DESIGN].


On 19 January 2014 21:45, Sam Parkinson  wrote:

> Hi
>
> Here is my PR that makes the journal icons better. It now only shows the
> white outline. You can check it out here:
> https://github.com/sugarlabs/sugar-toolkit-gtk3/pull/88 .
>
> Sam
>
> ___
> Sugar-devel mailing list
> Sugar-devel@lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel
>
>


-- 
Daniel Narvaez
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel


[Sugar-devel] Better journal icons

2014-01-19 Thread Sam Parkinson
Hi

Here is my PR that makes the journal icons better. It now only shows the
white outline. You can check it out here:
https://github.com/sugarlabs/sugar-toolkit-gtk3/pull/88 .

Sam
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel


Re: [Sugar-devel] mime type error in sugar

2014-01-19 Thread Daniel Narvaez
On 19 January 2014 17:04, Gonzalo Odiard  wrote:

> I don't know if this was needed by historical reasons,
> maybe the Gio functionality was not available.
>
> mime.get_for_file do some magic to recognize text files.
>

Gio seems to handle that just fine (tested with Gio.content_guess_type).


> Probably mime.get_for_file could call Gio instead of the SugarExt
> equivalent.
>

IMO we should deprecate mime.get_for_file and replace it with Gio.
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel


Re: [Sugar-devel] mime type error in sugar

2014-01-19 Thread Gonzalo Odiard
I don't know if this was needed by historical reasons,
maybe the Gio functionality was not available.

mime.get_for_file do some magic to recognize text files.

Probably mime.get_for_file could call Gio instead of the SugarExt
equivalent.

Gonzalo


On Sat, Jan 18, 2014 at 6:49 AM, Daniel Narvaez  wrote:

> Why we need that method at all? Without knowing the history of this code,
> it seems like we should Gio, both if the file exists and if it doesn't (if
> I remember correctly guess if for when the file doesn't exist).
>
>
> On 16 January 2014 20:42, Gonzalo Odiard  wrote:
>
>> I have filled #4715 [1] with information about a problem with the mime
>> database in sugar.
>> A workaround is available, but I would like to know what other think
>> about using Gio to identify the mime types instead of our own
>> SugarExt.mime_get_mime_type_for_file
>> Regards,
>> --
>> Gonzalo Odiard
>>
>> SugarLabs - Learning Software for children
>>
>> [1] http://bugs.sugarlabs.org/ticket/4715
>>
>> ___
>> Sugar-devel mailing list
>> Sugar-devel@lists.sugarlabs.org
>> http://lists.sugarlabs.org/listinfo/sugar-devel
>>
>>
>
>
> --
> Daniel Narvaez
>



-- 
Gonzalo Odiard

SugarLabs - Learning Software for children
___
Sugar-devel mailing list
Sugar-devel@lists.sugarlabs.org
http://lists.sugarlabs.org/listinfo/sugar-devel