On 14/09/2007, Brendan Eich <[EMAIL PROTECTED]> wrote: > "Almost all" excludes IE JScript, right?
Well, it doesn't agree with me on what is the right approach, but it does agree with me that the ES3 approach is wrong. > Those implementations are > JavaScriptCore, based on PCRE, and what else? The only implementations that I know of that specifically implement the ES3 algorithms are those in SpiderMonkey, linear_b and futhark (I have no idea about SEE, Rhino, and a lot of other fringe JS engines however). And the ES4 refimpl I guess. I only said "almost all" because I haven't tested each and every one. It could be possible that the only ones behaving according to the ES3 algorithm *at all* are indeed those written to match the ES3 algorithms. It's also possible that there are those that do match the ES3 algorithms but are not written specifically for them - I just don't know of any such implementation. JavaScriptCore uses PCRE, so of course it's not written specifically for the ES3 algorithms. The JScript one is shared with at least VBScript, and I wouldn't be surprised if it was shared with VB, VBA and other COM environments as well. I don't think it's written specifically for the ES3 algorithms. It's peculiar, and I think it doesn't agree with anybody else on a few edge cases besides those I've brought up. > Again, it's the Perl connection. Not only. Both Perl and those POSIX flavour regex that implement capturing submatches... -- David "liorean" Andersson _______________________________________________ Es4-discuss mailing list Es4-discuss@mozilla.org https://mail.mozilla.org/listinfo/es4-discuss
