[IAEP] Browse version using Webkit2

[tkkang at nurturingasia.com]

Thanks for working on the browser issue and hope more people get involved so that the XO-1 can get the recent software update - all except XO-1 currrently because of the browser issue.

I notice when I point the current browser to www.youtube.com I get kickout and it will return me back to start menu.

Cheers and look forward to more development.  

>-----Original Message-----
>From: Gonzalo Odiard [mailto:godiard at sugarlabs.org]
>Sent: Thursday, April 30, 2015 10:47 PM
>To: 'Sugar-dev Devel', 'iaep'
>Subject: Re: [IAEP] Browse version using Webkit2
>
>Testing old version of Browse in community version images, I can't
>reproduce bug #12866
>If we could solve the connectivity / collaboration issues...
>
>Gonzalo
>
>On Thu, Apr 30, 2015 at 10:34 AM, Gonzalo Odiard <godiard at sugarlabs.org>
>wrote:
>
>> I was trying to find a solution for the problem of Browse in XO-1.
>> After try use a old version of Browse, based in gecko,
>> and recompile the dependencies without success, I started to work in the
>> port of Browse from WebKit to WebKit2.
>>
>> That will not work in Fedora 18 (because Webkit2 crash in F18),
>> but could solve the problem in the new community
>> images based in Fedora 20.
>>
>> I have a basic version working, still with a lot of errors, but is
>> possible navigate,
>> download and see pdf  files and multimedia files.
>> A testing version (not for distribution) is here [1]
>>
>> TESTING:
>> I can't connect to internet with my XO-1 with the community edition image,
>> if anybody can, please test this version and verify if this bug [2] is
>> still present
>> and if the activity crashes.
>> Report any error you find.
>> Known issues:
>> * History is not restored.
>> * Ugly error messages
>>
>> DEVELOPMENT:
>> The porting is done in the branch webkit2 [3]
>> Master will only receive critical fixes. When the port is ready, will be
>> merged in master.
>> Code pending to port is identified by a a comment "# TODO PORT"
>> I found some blockers I couldn't solve yet:
>> * The objects where the information about the history is stored
>> have read only properties (title and url), I can recreate the history
>> with the data stored in the journal metadata.
>> * Need a mechanism to configure where download files, right now
>> is done in the Downloads directory and later copied to the journal.
>> * The error callback function in browser.py receive a int instead of
>> a GError. Looks as a introspection error.
>> * Don't know how store the coockies needed by the xs server with the new
>> API.
>>
>>
>> [1] http://dev.laptop.org/~gonzalo/activities/Browse-158.xo
>> [2] https://dev.laptop.org/ticket/12866
>> [3] https://github.com/sugarlabs/browse-activity/commits/webkit2
>> [4]
>> http://webkitgtk.org/reference/webkit2gtk/stable/WebKitBackForwardListItem.html
>> --
>> Gonzalo Odiard
>>
>> SugarLabs - Software for children learning
>>
>
>
>
>-- 
>Gonzalo Odiard
>
>SugarLabs - Software for children learning
>