Re: [hybi] CML

["Shelby Moore" <[e-mail]>]

And when we return the CML bit to reserved bits, please make it the MSB
next to LenLen, so that if ever we need to support 128-bit sizes, it will
be easy to do so.


>> John Tamplin writes:
>>> Remember that the current JS API is message based, where it receives a
>>> single string containing the entire message.  It seems likely browsers
>>> are going to allocate an entire buffer for the incoming message
>>> (granted
>>> some of them will want that to be in UTF16 characters, which won't be
>>> tied to the number of UTF8 bytes - there was discussion of perhaps
>>> adding that as an extension), filling it as it reads fragments, and
>>> then
>>> handing it off to client code once the last frame is received.
>>
>> 'twasn't my memory that was the problem, but the assumption that
>> reallocation of memory was necessary.  If the message arrives in pieces,
>> you have two choices:
>>   A. re-assemble those pieces into the expected form
>>   B. provide an abstraction that makes it look like a single piece
>>
>> This holds true irrespective of the nature of the API, though there
>> might
>> be some implementation constraints that make one or other harder.
>>
>> A is a fairly unsophisticated approach, it's simple and might be
>> perfectly
>> good if you don't want to spend too much effort on implementation, want
>> to
>> expose a simple array to users.  A is even better if you have a frame
>> size
>> limit.  B follows the old maxim: every problem can be solved by adding
>> another layer of indirection.  B might be more efficient if the
>> implementation needs to be more flexible.
>>
>>> Primarily I suggested it as a way to avoid the apparently endless
>>> arguments we were having about fragmentation.
>>
>> Sadly, I don't think that it necessarily addresses those concerns; nor
>> do
>> I believe that those concerns are especially well grounded.
>
> The above taken together with Thomson's prior post:
>
> http://www.ietf.org/mail-archive/web/hybi/current/msg03525.html
>
> I agree with Thomson's points.  The CML is useless, and perhaps harmful.
>
> CML is not a generalized abstraction of what it attempts to do.
>
> CML abstraction should be in the data not in frame header. For example, if
> the sender wants to signal the size of array, then put it in the data of
> the first frame sent.
> _______________________________________________
> hybi mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/hybi
>
>