On 28/06/07, Anne van Kesteren <[EMAIL PROTECTED]> wrote:
On Thu, 28 Jun 2007 05:57:09 +0200, Cameron McCormack <[EMAIL PROTECTED]> wrote: >> For sequence<octet> it would be really nice if we could have a more >> native representation of a byte array than a UTF-16 string. > > Strange, I really should have used sequence<unsignedshort> there. > Fixed. > > I still am a little unsettled by the special casing of > sequence<unsignedshort> for string values. > > We could revisit the decision not to use wstring, and just state that > wstrings must be sequences of UTF-16 code units, and that they can > include 0x0000.It would still be very good to have an octet / byte representation in ECMAScript. I'm aware of a couple of implementations of such a thing, but I haven't been able to play with them or figure out how they work exactly myself: 1. XMLHttpRequest has a responseBody member in Internet Explorer 7 that returns an array of unsigned integers representing octets. Likewise, the send() method accepts such an array. 2. I heard Adobe Flex has some notion of a byte array to represent files.
Maybe ES4 can come to the rescue here: <uri:http://developer.mozilla.org/es4/proposals/bytearray.html> -- David "liorean" Andersson
