- From: Charles Pritchard <chuck@jumis.com>
- Date: Wed, 6 Apr 2011 12:06:30 -0700
- To: Cameron McCormack <cam@mcc.id.au>
- Cc: Rick <graham.rick@gmail.com>, "Dr. Olaf Hoffmann" <Dr.O.Hoffmann@gmx.de>, "www-svg@w3.org" <www-svg@w3.org>
Related to this thread, I'm exploring the possibility of enabling CSS pointer-events to work with canvas. The stroke(), fill() and clearRect properties are sufficient to maintain a 1bit per pixel bitmap as an incremental improvement on the relationship between DOM events and the 2d context. It seems quite manageable within existing specs; but would lump strokeText in with stroke(), or simply not composite strokeText / fillText onto the 1bit per pixel backing store. I'd be ok with that latter behavior, as, if I wanted the text to be clickable, I could composite it onto a temporary layer, first. But it is a little awkward. Thoughts from the SVG crowd? It's my hope that work on canvas accessibility will contribute to svg a11y. -Charles On Apr 6, 2011, at 10:22 AM, Cameron McCormack <cam@mcc.id.au> wrote: > Rick: >> So, this is an answer, and it invalidates the problem presented by my >> use case. I'm straining my brain to think of a case where it wouldn't >> suffice. No luck. > > I think there are certainly situations where you want to have some text > that responds to pointer events but not by performing selection. The > usual way people solve this is to set pointer-events="none" on the text > element and place an invisible rectangle above/below it to capture the > mouse events. This isn’t ideal because you do not always know how big > the text will be, depending on font fallback and so on. > > -- > Cameron McCormack ≝ http://mcc.id.au/ >
Received on Wednesday, 6 April 2011 19:06:59 UTC