- From: Aryeh Gregor <ayg@aryeh.name>
- Date: Tue, 2 Aug 2011 13:29:24 -0400
- To: Anne van Kesteren <annevk@opera.com>
- Cc: Maciej Stachowiak <mjs@apple.com>, Boris Zbarsky <bzbarsky@mit.edu>, public-html@w3.org
On Tue, Aug 2, 2011 at 1:02 PM, Anne van Kesteren <annevk@opera.com> wrote: > Why the rush? If Selection.toString() can wait, so can innerText, no? innerText is far more widely used, so it's more important to get interop. It also doesn't work reliably for pretty-printing, since Opera doesn't do pretty-printing with it and Gecko doesn't support it at all, so it's much more practical to get it working with a minimal spec. In fact, a very large fraction of users only get it and don't set it, so implementing it as an alias to textContent would be better than nothing (unless it causes you to go down IE codepaths, of course). But you're right that there's an option (4) here: leave it unspecified indefinitely, until we have the bandwidth to write a spec for Selection.toString(), and then match that spec. I think that option is worse than having a minimal spec for now, but it's on the table. FWIW, Selection.toString() is currently defined by reference to Range.toString(), which behaves basically like textContent: http://html5.org/specs/dom-range.html#dom-selection-stringifier The spec just has an XXX saying that it's known to be wrong. Perhaps we should spec innerText in the simple way I suggest, but leave a warning in the spec explaining the compat situation and saying we might want to change it once we get a good definition for Selection.toString().
Received on Tuesday, 2 August 2011 17:30:21 UTC