You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The main motivation is that the new failure code is written in Rust so calling it directly makes more sense than indirecting through native runtime code.
This also lets us to have fewer stack-switching mechanisms as upcalls are dealt with differently than other foreign functions.
Probably upcall_call_on_c_stack/call_on_rust_stack needs to remain a native function, same for upcall_new_stack/del_stack (which are not called by generated code but my __morestack).
This has been (partially) completed for fail in de82a9b and malloc/free in d257382. The current implementation is a very thin wrapper around the C runtime, but in the future it should be possible to move more functionality over to rust.
Activity
brson commentedon Jul 10, 2012
Probably upcall_call_on_c_stack/call_on_rust_stack needs to remain a native function, same for upcall_new_stack/del_stack (which are not called by generated code but my __morestack).
brson commentedon Jul 10, 2012
Actually, upcall_call_on_x_stack should be fine as a rust functions. it should just be the upcalls called by __morestack that need to be special.
elliottslaughter commentedon Jul 23, 2012
This has been (partially) completed for fail in de82a9b and malloc/free in d257382. The current implementation is a very thin wrapper around the C runtime, but in the future it should be possible to move more functionality over to rust.
graydon commentedon Jun 6, 2013
Agree, properly tagged and milestoned as of 2013-06-05.
emberian commentedon Aug 5, 2013
@brson is this still relevant?
brson commentedon Oct 10, 2013
@cmr Pretty much obsolete.
Update date references for 2026-05 triage (rust-lang#2861)
Merge pull request rust-lang#2872 from santihdzs/date-ref-triage-2026-05
Merge rust-lang#2861