- From: Aryeh Gregor <ayg@aryeh.name>
- Date: Thu, 17 Nov 2011 19:13:38 -0500
- To: James Graham <jgraham@opera.com>
- Cc: Dominique Hazael-Massieux <dom@w3.org>, public-test-infra@w3.org, public-script-coord@w3.org
On Thu, Nov 17, 2011 at 9:06 AM, James Graham <jgraham@opera.com> wrote: > I think there has been no announcement yet because Aryeh only started > hacking on it yesterday :) I think there is still quite some work to do. Yeah, I haven't announced it because it's still in heavy flux. In particular, I just rewrote most of it today in a completely incompatible way, so any scripts using it will have broken. It should be considered highly unstable for the time being, but I expect to get a relatively stable API and a lot more features by sometime next week, if all goes well. However, comments or suggestions would still be great at this point, with respect to what API people would like to see and/or what things it should test. My plan is that every spec that defines any IDLs should be able to generate basic WebIDL tests with only a few easily-automate-able steps, suitable for incorporating into a Makefile or such so that the tests can be updated whenever the spec changes with no extra effort. For more in-depth testing, some extra one-time manual work will be necessary -- such as providing objects that implement the IDLs (for IDLs with no constructors defined), or identifying "safe" methods or attributes that can be called/get/set with various values to check return values and such without having to worry about side-effects. (Or maybe I'll default to assuming things are safe, and leave it up to the caller to provide an object that can be freely trashed, plus perhaps a way to flag *unsafe* methods.) Anyway, it's definitely not suitable for use yet.
Received on Friday, 18 November 2011 00:14:30 UTC