- From: liorean <liorean@gmail.com>
- Date: Sun, 20 May 2007 03:33:35 +0200
> > Ian Hickson wrote: > > >>Speaking of setTimeout, where is this defined? > > > http://whatwg.org/specs/web-apps/current-work/#settimeout > On 4/21/05, Dean Edwards <dean at edwards.name> wrote: > > OK. That's twice in one day. I'm off to read the WA1 spec.... On 21/04/05, Jim Ley <jim.ley at gmail.com> wrote: > It's rather odd though, as it's been defined such that the mozilla > implementation will be non-conformant, either the mozilla > implementation will need to change to be conformant - breaking > compatibility with existing scripts. Or mozilla will not be able to > be conformant. As I understand it, there are currently two different signatures, the Microsoft one and the Netscape one. The Netscape one adds a virtually unlimited number of optional arguments to send to the function when the timer triggers. (IIRC in reality it's limited to a maximum of 2^16-3 arguments, so it's not really unlimited. But how many scripts in the wild use 65 533 arguments for any function, anyway?) The Microsoft one adds a single optional argument specifying which language the string is to be parsed as. The Netscape signature is for when the first argument is a function reference, the Microsoft signature for when the first argument is string containing the timer function source code. Exactly how is that incompatible with the Mozilla implementation? I can, without actually testing code, only see one possible incompatibility: If the Mozilla implementation also allow the Netscape signature for when the first argument is a string. If that is the case, it is a pretty obscure usage that I can't imagine has found much use on the web considering how it is incompatible with Microsoft's implementation. If this is the incompatibility you're speaking of, is there any data on how common this practice is, as well as how common the incompatible Microsoft variant is? -- David "liorean" Andersson
Received on Saturday, 19 May 2007 18:33:35 UTC