[squeak-dev] self copyEmpty better than self species new [was: Why
is Heap>>#species => Array?}
Klaus D. Witzel
klaus.witzel at cobss.com
Thu Feb 21 22:20:20 UTC 2008
On Thu, 21 Feb 2008 22:19:43 +0100, nicolas cellier wrote:
...
>
> opened http://bugs.squeak.org/view.php?id=6943 for this one
Well, hm, I have as yet not thought about the subject line in that report.
But now I think that Heap #species problem is more subtle: (self species
new) cannot pass the sortBlock of Heap (nor any other clone-invariant
instVar) for users like #collect: and friends; therefore Array is passed
by Heap #species.
But #copyEmpty can pass sortBlock and *whatsoever* to clones ;-) With
#copyEmpty in place of (self species new), sending #species would be
mainly in support of #= and #copyEmpty, throughout the collection
hierarchy.
BTW hey Traits guys, did you discover something like the above during your
reengineering of the collection hierarchy, that would be interesting to
know !
>
>
More information about the Squeak-dev
mailing list
|