| Age | Commit message (Collapse) | Author | Lines |
|
structural comparison from equal_type_structures()
|
|
|
|
|
|
|
|
They are now created by the resolve pass, which is the only pass that
needs them, and kept internal to that pass.
|
|
|
|
|
|
|
|
|
|
|
|
According to --time-passes, resolution went from 2 to 0 seconds. Not
really the bottleneck... but if we want to be crazy fast, just
consider this a future bottleneck that was fixed very timely.
|
|
* Cleans up the algorithm
* Move first pass to walk (second still folds)
* Support part of a type/value namespace split
(crate metadata and module indices still need to be taught about this)
* Remove a few blatant inefficiencies (import tables being recreated for
every lookup, most importantly)
|
|
|
|
|
|
|
|
|
|
|
|
of direct calls to metadata.Encode.ty_str() in trans
|
|
|
|
|
|
|
|
|
|
translation time.
|
|
|
|
|
|
stack now.
|
|
|
|
|
|
|
|
|
|
rust_native_rust_* stubs
|
|
|
|
matches. Add a test case.
|
|
|
|
and introduce an "either value or alias" mode
|
|
module passes, along with the TargetData passes.
|
|
|
|
|
|
|
|
|
|
supports 8 arguments.
|
|
(Have fun mergining your stuff with this.)
|
|
|
|
|
|
from the parameter tydescs. This allows SROA, and consequently inlining, to work.
|
|
|
|
generated for tydescs
|
|
|
|
|
|
|