On Sunday, 11 October 2015 at 04:35:03 UTC, Jonathan M Davis
wrote:
On Sunday, 11 October 2015 at 04:16:11 UTC, deadalnix wrote:
If we go these DIP road, there is no coming back and this will
get in the way of a principled approach.
Then come up with an alternative DIP which shows a better way
to solve this. As it stands, it looks likely that we'll end up
with some form of DIP 74, and if you have a better proposal,
then now is probably the time to do it.
I have, other have as well, in fact there was a lengthy
discussion in private between Walter, Andrei, myself and some
other where very precise proposal has been made.
I'm sorry but that's bullshit. I'm tired of reexplaining the same
thing again and again while is it clear that nobody cares about
facts here. If people would care about facts, the DIP25 fiasco
would have been enough to put the ideas back on the table.
The one place in-language where I'm sure that something like
RefCounted doesn't do it is exceptions, since we really should
have a way to make those reference counted, but you can only
throw something derived from Throwable - which means a class
and not a wrapper around one. So, we need a tweak of some kind
to the language there, but that's pretty specific, whereas it
_should_ be possible to get something like RefCounted to at
least solve the normal cases. Clearly though, Walter and Andrei
have come to the conclusion that it's not.
Proposal included solution for this problem. Exception ownerhip
is transferred to the runtime when thrown, and the runtime
transfer it back to the catch block that effectively catches it.