From: Donal K. F. <don...@ma...> - 2006-06-29 10:34:02
|
miguel sofer wrote: > To test this matter, I have written a patch that implements this on HEAD > and submitted it as FR #1512138 to sourceforge. I've reviewed the patch and I think it looks largely OK. But we should *not* document the use of -1 as a signalling value. Keep that info inside the Tcl library and instead have people call a function (probably in the internal stubs table) to find out whether the object is being deleted. That future-proofs us instead of spraying the results of the hack across some unknown number of libraries. ;-) (There's one or two comment formatting issues too, but they're trivial.) Donal. |