- From: Cameron McCormack <notifications@github.com>
- Date: Sun, 05 Jul 2015 23:14:00 -0700
- To: heycam/webidl <webidl@noreply.github.com>
- Message-ID: <heycam/webidl/pull/52/r33908055@github.com>
> </p> > - <div class='note'> > - <p> > - Interfaces using <a class='xattr' href='#ArrayClass'>[ArrayClass]</a> will > - need to define a “length” <a class='dfnref' href='#dfn-attribute'>attribute</a> > - of type <span class='estype'>unsigned long</span> that exposes the length > - of the array-like object, in order for the inherited <span class='estype'>Array</span> > - methods to operate correctly. Such interfaces would typically also > - <a class='dfnref' href='#dfn-support-indexed-properties'>support indexed properties</a>, > - which would provide access to the array elements. > - </p> > - </div> > + <ol class='algorithm'> > + <li>Let <var>values</var> be the result of > + <a class='dfnref' href='#dfn-convert-ecmascript-to-idl-value'>converting</a> Hmm, so in what way are they different, apart from doing the iterable check up front (before doing it again in the "convert to sequence<T>" algorithm)? Maybe it's better to just call into the "convert to sequence<T>" and "convert to FrozenArray<T>" algorithms directly here from the union conversion algorithm. That would still have some duplicate processing (we'd check for Date/RegExp twice now, rather than checking for iterability twice) but at least would capture any changes we make to those two algorithms without needing to change anything here. --- Reply to this email directly or view it on GitHub: https://github.com/heycam/webidl/pull/52/files#r33908055
Received on Monday, 6 July 2015 06:14:27 UTC