From: Eric M. Ludlam <eric@si...> - 2003-02-27 03:23:28
The token access code has now all moved. The new semantic-token.el
file just has all the good ol' code from before.
You can read in the commentary of that file a little bit about how I
think it could be organized.
Next up is to prototype a new API. This is a bit of a hybrid of what
David had proposed a while back.
A generic token may be created like this:
(semantic-token name type-symbol &rest plist)
;; Note, doc-string could be found in the plist, and properties and
overlay are set up after creation.
Special tokens could be created like this:
(semantic-token-new-variable name type default-value &rest extra-specifiers)
(semantic-token-new-function name type arg-list &rest extra-specifiers)
(semantic-token-new-type name type part-list parents &rest extra-specifiers)
(semantic-token-new-include name system-flag &rest plist)
(semantic-token-new-package name detail &rest plist)
where the extra-specifiers are a plist. In the existing storage
mechanism, they would be put into the extra-specifier slot, and
reduced (removing nils).
I think the special token constructors would be useful for language
authors who need encouragement to fill in all the right pieces.
Eric Ludlam: zappo@..., eric@...
Home: http://www.ludlam.net Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net GNU: http://www.gnu.org