The recent Cargo update changed filenames, which broke a lot of incremental
rustbuild builds. What it thought were the output files were indeed no longer
the output files! (wreaking havoc).

This commit updates this to stop guessing filenames of Cargo and just manage
stamp files instead.

Apparently some versions of git don't support the `-C` flag, so let's use the
guaranteed-to-work `current_dir` function.

…bleImpl.

…criminant.

Fixes rust-lang#29361.

…rner

Update E0049 to new error format

Fixes rust-lang#35210.
Part of rust-lang#35233.

r? @jonathandturner

…r-docs, r=GuillaumeGomez

Tweak std::marker docs

Fixes rust-lang#29361.

r? @GuillaumeGomez

rustbuild: Fix dependency tracking with new Cargo

The recent Cargo update changed filenames, which broke a lot of incremental
rustbuild builds. What it thought were the output files were indeed no longer
the output files! (wreaking havoc).

This commit updates this to stop guessing filenames of Cargo and just manage
stamp files instead.

…r=jonathandturner

rustbuild: Use current_dir instead of -C

Apparently some versions of git don't support the `-C` flag, so let's use the
guaranteed-to-work `current_dir` function.

…p-glue, r=eddyb

trans: Let the collector find drop-glue for all vtables, not just VTableImpl.

This fixes rust-lang#36260. So far, the collector has only recorded drop-glue for insertion into a vtable if the vtable was for an impl. But there's actually no reason why it shouldn't do just the same for closure vtables, afaict.

r? @eddyb
cc @rust-lang/compiler

…bnik

doc: make that sound better

Don't ignore a doc code-block we can compile.

…eddyb

Workaround rust-lang#34427 by using memset of 0 on ARM to set the discriminant.

Workaround rust-lang#34427 by using memset of 0 on ARM to set the discriminant.

Default RUST_MIN_STACK to 16MiB for now

Temporary(?) workaround to rust-lang#35408.

cc @nikomatsakis if you want to veto the idea.

r? @eddyb