[ENH] ContextCleanup ([er] start, request for help)
Anthony Hannan
ajh18 at cornell.edu
Thu Jun 5 17:41:59 UTC 2003
Thanks, Daniel, for reviewing this. I have made some changes to the
package per your suggestions. The new version is on its home page:
http://minnow.cc.gatech.edu/squeak/ClosureCompiler.
danielv at netvision.net.il wrote:
> About the first change set that includes various changes -
> * I think the use of primitiveChangeClassTo: in order to weasel our way
> out of displaying the right inspector is a spectacular abuse of an
> unspiked fence. Anybody see a reason not to rewrite this as a nice,
> simple dispatch creating the right subclass instance in the class side?
The reason I just change the inspector's class is because the debugger
holds the inspector and has dependents on it. It was easier to change
the class then substitute a new inspector and re-establish all those
pointers.
> * BasicObject seems to be unused, in any of these changesets. Seems to
> me we don't need it.
> * Two of the methods added to Object, add an idiom which reads strange
> to me. Is it just me? they're inCollection and inCollectionDo:. Other
> opinions?
> * Various protocol added to OrderedCollection, including a clear stack
> protocol. The stack protocol is more intention revealing, so it's good,
> IMO, but I'm not sure about the rest.
I removed BasicObject and other protocols like #inCollection that are no
longer used by the Closure Compiler.
> * Various protocol added to Stream subclasses, which I don't follow. Any
> takers? elucidation from Anthony?
The new protocols are enumeration, insert, remove, and backup. This
allows more flexible manipulation of streams.
> * Protocol added to SequenceableCollection is generally good,
> #collectArray: is a good idea, but should probably be in Collection, so
> set gets it too (that's where I've needed it, anyway). There's so much
> perculiar added protocol, I'm wondering where the need for it is all
> coming from...
I've removed peculiar protocol and moved #collectArray: to Collection.
Cheers,
Anthony
More information about the Squeak-dev
mailing list
|