Re: [Erlyaws-list] argh, chrome 14 broke yaws winsockets :(

[Jan B. <jb...@gm...>]

On 7 October 2011 01:38, Wes James <com...@gm...> wrote:
> ignore the last message.....
>
> I forgot to go back to the regular functions in yaws_api.erl.  Your
> work is doing what I was wanting all along.  It gets the whole chunk
> back to the client properly!
>
> Awesome!!
>
> Thanks Jan!

Excellent :D

I'm now looking at making a test suite to be able to regression test
as the rest of the protocol gets implemented. This will also help as
the protocol gets finalised. I read that Chromium 16 will bring
Websocket version 16 [1]. I think the closing handshake changes a bit
between  versions 8 and 16, and as you found, I haven't even
considered handling closing yet :)

Best
JD

[1] http://groups.google.com/a/chromium.org/group/chromium-extensions/browse_thread/thread/9f8214e3a9ed037f

>
> -wes
>
> On Thu, Oct 6, 2011 at 5:24 PM, Wes James <com...@gm...> wrote:
>> On Thu, Oct 6, 2011 at 2:35 PM, Jan Bothma <jb...@gm...> wrote:
>> <snip>
>>
>>> Haven't looked in detail, but if you're trying to send two frames with
>>> the data fragmented between them, the first bytes should be different
>>> because of the Fin bit.
>>>
>>> Anwyay, I've improved long messages (although there's a bug for
>>> multibyte characters I think)
>>> https://github.com/jbothma/yaws/commit/9d34f72ffde4ea6ec33c399b203c74e986c83c0d
>>> so all your data might work with my updated frame and unframe.
>