[webkit-dev] Place For new Core Module

Eric Seidel eric at webkit.org
Tue Aug 24 15:59:30 PDT 2010


Secret ports have an absolutely horrible track record of ever catching
up with public WebKit.

Only one has ever been successful to my knowledge (Chromium) and I'm
not even sure we could call it full success yet.  (They've spent 2
years attempting to fully catch up, and yet you can't build a useable
binary out of webkit.org -- although that's very close!)

Apple's iPhone fork has (or had, maybe this has changed) one full-time
person who's job is constantly merging.  I'm not sure how close to
tip-of-tree they're able to stay.


It's possible that you work with the most fantastic engineers WebKit
will ever see.  But let me caution you: if CableKit forks in secret,
you're very likely to always be months if not years behind and riddled
with security vulnerabilities. :)

Furthermore, specs take a long time and a lot of buy-in.  Spec-work
done in secret is unlikely to end up ever getting the buy-in it needs
to become a spec.

Best of luck.

-eric

On Tue, Aug 24, 2010 at 3:50 PM, Chinmaya Sn <chinmaya at gmail.com> wrote:
> Thanks All. I think I am getting the general idea of what can get into
> to WebKit and how things fit.
>
> Right now both the standard and implementation are WIP, spec is intended to
> be a Web standard. At this point, neither the standard nor the
> implementation are public.
>
> My primary goal was to keep my changes around the WebKits coding and
> development rules,
> such as: I am already using WTF library extensively to keep the code
> compatible with WebKit.
>
> In future, if I can convince you folks to take these changes into
> WebKit it will be lot easier.
> What this means to me today, is just enforcing WebKit's disciplines.
> If I can't get these changes into WebKit then may be I need to fork as
> CableKit or something :)
>
>
> May be I pushed the button little early, I should wait till 108th minute.
>
>
> cheeers
> Chinmaya
>
> On Tue, Aug 24, 2010 at 3:39 PM, Maciej Stachowiak <mjs at apple.com> wrote:
>>
>> Given your description, I'm not sure this functionality is appropriate for WebKit. We are generally reluctant to take feature patches for features that are experimental or of niche interest - we prefer to focus on technologies that are part of the standards-based Web technology stack, that are likely to become Web standards, or that are valuable general-purpose complements to Web standards.
>>
>> It sounds like the feature you are talking about does not fit this template very well currently, though it may in the future if it does get on track to become a Web standard.
>>
>> Regards,
>> Maciej
>>
>> On Aug 24, 2010, at 2:26 PM, Chinmaya Sn wrote:
>>
>>> Let me put it this way, I am working on a new standard, which
>>> leverages some of HTML5 standards but tries to extend them
>>> (not enhance) to answer some of the Cable industry challenges.
>>>
>>> I started by analysing WebKit layer, it has very strong assumption of
>>> layering browser-engine, and mostly has APIs which helps you
>>> build a html browser, why not, WebKit is a browser-engine.
>>>
>>> In my first case, of rendering a special kind of "app" on top of HTML5 video;
>>> these apps don't speak html/javascript, they speak primitives such as
>>> drawRectangle, drawLine etc. and has own events and processing
>>> functions (its a language in itself) which needs to be translated into
>>> either C++ or HTML/JavaScript.
>>>
>>> My primary requirements is to add new modules which represent new
>>> standards which work in conjunction with HTML5 (very closely) without
>>> any changes inside WebCore.
>>>
>>>
>>>
>>> Today WebKit would not gain anything, this is new standard work in
>>> progress. Tomorrow things might change, and if someone says I
>>> want this Cable specific feature, one could easily tell 'hey build
>>> WebKit with this flag and you will get that feature for free',
>>> that is my motto.
>>>
>>> I am driving this as as future proof solution.
>>> Currently I'll be performing these changes only in my own branch.
>>>
>>> Makes sense?
>>>
>>>
>>>
>>> On Tue, Aug 24, 2010 at 2:09 PM, Ariya Hidayat <ariya.hidayat at gmail.com> wrote:
>>>> I guess you also need to supply the justification: what would WebKit
>>>> gain from such modules? Are the modules part of HTML5 and other web
>>>> standards?
>>>>
>>>>
>>>> Regards,
>>>>
>>>
>>>
>>>
>>>
>>> On Tue, Aug 24, 2010 at 2:17 PM, Simon Fraser <simon.fraser at apple.com> wrote:
>>>> I'll be the first to ask this question: what do you need to do that you can do
>>>> with existing WebKit APIs, and existing web technology?
>>>>
>>>> With video, you could use http streaming to deliver custom content,
>>>> and make sure of accelerated compositing to ensure that you can
>>>> efficiently layer HTML content on top of video.
>>>>
>>>> If your platform does not have support for these, then the "right"
>>>> way to address this is to add support for these missing pieces,
>>>> rather than hacking at WebCore internals.
>>>>
>>>> Simon
>>>>
>>>> On Aug 24, 2010, at 12:31 PM, Chinmaya Sn wrote:
>>>>
>>>>> The current module I have implemented works closely with
>>>>> RenderingEngine/GraphicsContext and HTML5 video player, which
>>>>> tries to deliver some special interactive Apps (In Cable World) as an
>>>>> overlay over HTML5 Video,
>>>>> the apps themselves are delivered as one of the private streams in the
>>>>> Video, so some work has gone inside implementation
>>>>> of video player itself.
>>>>>
>>>>> In near future, I will be poking around a lot  with HTML5 player. I
>>>>> will be implementing lot more modules like above,
>>>>> which will be linked to WebCore. Each of the new enhancements I
>>>>> provide needs to be use WebCore and should be
>>>>> accessible to HTML page author. So I will be drilling some holes to
>>>>> expose these modules as a finite set of JavaScript objects.
>>>>>
>>>>> --
>>>>> Chinmaya
>>>>>
>>>>>
>>>>>
>>>>> On Tue, Aug 24, 2010 at 1:23 PM, Peter Kasting <pkasting at google.com> wrote:
>>>>>> On Tue, Aug 24, 2010 at 12:21 PM, Chinmaya Sn <chinmaya at gmail.com> wrote:
>>>>>>>
>>>>>>> I have special case in my hand; currently I maintain a custom branch
>>>>>>> of WebKit, I already have
>>>>>>> implemented a module which sits closely with WebCore and it works just
>>>>>>> fine. One of my
>>>>>>> goals is to deliver these changes back to community and even may be
>>>>>>> back to WebKit repository. In near
>>>>>>> future I will be implementing at least 2-3 such modules, which work
>>>>>>> closely with WebCore and
>>>>>>> JavaScriptCore but they may not have to anything to do HTML browser
>>>>>>> engine as such.
>>>>>>
>>>>>> Why don't you say more about what you're trying to do.  It's hard to answer
>>>>>> in the abstract.
>>>>>> PK
>>>>>
>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Chinmaya
>>> _______________________________________________
>>> webkit-dev mailing list
>>> webkit-dev at lists.webkit.org
>>> http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
>>
>>
>
>
>
> --
> --
> chinmaya sn
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> http://lists.webkit.org/mailman/listinfo.cgi/webkit-dev
>


More information about the webkit-dev mailing list