Improve extension APIs for TB

David Bienvenu dbienvenu at mozilla.com
Wed Oct 26 14:40:36 UTC 2011


On 10/26/2011 6:34 AM, Wayne Mery (vn) wrote:
> Does "no further effort" mean that STEEL is the current state of the art, but no longer a strategic direction?
Iirc, Steel was abandoned in favor of supporting Jetpack, though I think eventually there was a realization that Jetpack wasn't going to be a panacea because so much of 
what TB extensions do requires access to the core mail interfaces.  asuth can probably express that more clearly...

stdlib sounds more like a Steel replacement. I think fleshing this out and sanitizing it (not depending on current selection, etc) would be very helpful.

I'm all for extension authors getting involved in describing the issues they have with the current interfaces and proposing new and/or improved interfaces, instead of going 
to great lengths to workaround the issues. Our interfaces are *not* frozen, and we want to improve them.

I've been working on making the backend more scriptable and extensible (e.g., making msg sending scriptable, and not require an nsIEditor, paving the way for rkent's 
SkinkGlue, etc) but it would be great to figure out other high leverage improvements, to enable extensions to do things they could not do before.

- David




More information about the tb-planning mailing list