The macro now takes a format string. It no longer defaults to using the
type name. Didn't seem worth going through contortions to maintain. I
also changed most of the debug formats to be `foo[N]` instead of `fooN`.
Also, factor out `do_mir_borrowck`, which is the code that actually
performs the MIR borrowck from within the scope of an inference context.
This change should be a pure refactoring.
This will be important in next commit, since the input types will be
tagged not with `'gcx` but rather `'tcx`. Also, using the region-erased,
lifted types enables better caching.
rustbuild: Don't build with ThinLTO on MIPS
Discovered in #45529 it looks like cross-module TLS imports aren't quite working
today, especially with `hidden` visibility which mostly comes up with multiple
codegen units. As a result this completely disables compiling with ThinLTO and
multiple codegen units on MIPS when bootstrapping.
cc #45654, the tracking issue for this
Discovered in #45529 it looks like cross-module TLS imports aren't quite working
today, especially with `hidden` visibility which mostly comes up with multiple
codegen units. As a result this completely disables compiling with ThinLTO and
multiple codegen units on MIPS when bootstrapping.
cc #45654, the tracking issue for this
incr.comp.: Fix two problems with HIR hashing.
Fixes https://github.com/rust-lang/rust/issues/45469.
This PR fixes two small problems:
* Overflow checks are always enabled in a constant context, so we need to hash spans of potentially overflowing operations. (Eventually I'd like to handle spans differently so we don't have to make HIR hashing know so much about things like this.)
* The HIR map collector had a bug where it would assign the `DepNode::Hir` instead of the corresponding `DepNode::HirBody` in some nested contexts.
r? @nikomatsakis
