remove ref-counting Gc<T> #256
Conversation
|
Big 👍 here. |
|
+1, and we should make the plan known to as many in the community as possible, so people will not assume that Rust is never going to have a proper tracing GC. |
|
Sounds good! |
| implementation). I propose leaving them, but it does not matter | ||
| terribly to me. The important thing is that once `std::gc` is gone, | ||
| then we can remove the support code associated with those two lang | ||
| items, which is the important thing. |
There was a problem hiding this comment.
Take 'em out - keep things clean as possible. Chances are the future GC will end up having different lang items, and in any case it is no bother to add more lang items.
|
👍 |
|
(hmm, the sample program I just added to the RFC seems to consistently cause the playpen timeout to trigger. AFAICT it seems to be triggered by my call to |
|
👍 |
|
@pnkfelix: Would it be possible to just include output from a run of the sample program, so that we don't need to run it in the playpen (or compile it locally) to see what it's demonstrating? |
Revise the printouts as well to be clearer about each cases' intent.
|
Accepted as RFC 67. Discussion. Tracking. |
Centril
added
A-allocation
Summary
Remove the reference-counting based
Gc<T>type from the standardlibrary and its associated support infrastructure from
rustc.Doing so lays a cleaner foundation upon which to prototype a proper
tracing GC, and will avoid people getting incorrect impressions of
Rust based on the current reference-counting implementation.
(rendered)