Anne van Kesteren wrote:
On Thu, 20 Apr 2006 23:38:56 +0200, Julian Reschke
<[EMAIL PROTECTED]> wrote:
In the end, we want to have these clients/libraries fixed, right?
I guess that's the question. Is it sensible to implement it as
case-sensitive knowing that you probably break content? (Such as the
examples in the specification...) Should we specify that methods MUST be
treated as case-sensitive knowing that we will than never exit CR state
given the way it has been implemented and will stay implemented because
of the content that's out there? Those type of questions...
I think a clear signal should be sent to implementors that they should
get things right.
Right now, an XHR implementation uppercasing method names probably is
harmless, as (IMHO) all HTTP method names defined in IETF specs indeed
use uppercase letters.
On the other hand, an implementation that isn't able to transmit any
legal method name that already *is* uppercase clearly is broken, and the
spec should say that. Otherwise XHR by definition is useless for
anything that does beyond GET/PUT/POST/DELETE & friends.
Best regards, Julian
(who wants to be able to implement WebDAV client functionality using XHR)