This will make @path work with miri and other non-standard entrypoints.
Also since this simplifies librustc_driver::args, move it into a simple source file. Also
remove the tests since they're doing nothing more than checking `str::lines` has the right
behaviour.
Allow git to merge `Cargo.lock`
This commit backs out #46539 in order to fully leverage #63579 where
`git` should be able to merge `Cargo.lock` nowadays with only minimal
conflicts.
Consolidate sigemptyset workarounds
In sys/unix/process, we work around the sigemptyset linking issues
on android in two different ways. This change consolidates these
workarounds, and avoids duplicating bindings from `libc`.
Don't use stage naming in RUSTFLAGS environment variables
This patch supports less behavior than before, since specifiying stage 1
vs stage 2 is no longer possible, but that is presumably a somewhat rare
use case anyway, so not supporting it seems acceptable (and it can be
readded easily if desired).
Fixes#33609
Fix bug in iter::Chain::size_hint
`Chain::size_hint` currently ignores `self.state`, which means that the size hints of the underlying iterators are always combined regardless of the iteration state. This, of course, should only happen when the state is `ChainState::Both`.
avoid unnecessary reservations in std::io::Take::read_to_end
Prevously the `read_to_end` implementation for `std::io::Take` used its
own `limit` as a cap on the `reservation_size`. However, that could
still result in an over-allocation like this:
1. Call `reader.take(5).read_to_end(&mut vec)`.
2. `read_to_end_with_reservation` reserves 5 bytes and calls `read`.
3. `read` writes 5 bytes.
4. `read_to_end_with_reservation` reserves 5 bytes and calls `read`.
5. `read` writes 0 bytes.
6. The read loop ends with `vec` having length 5 and capacity 10.
The reservation of 5 bytes was correct for the read at step 2 but
unnecessary for the read at step 4. By that second read, `Take::limit`
is 0, but the `read_to_end_with_reservation` loop is still using the
same `reservation_size` it started with.
Solve this by having `read_to_end_with_reservation` take a closure,
which lets it get a fresh `reservation_size` for each read. This is an
implementation detail which doesn't affect any public API.
This is no longer used by the index generator and was always an unstable
compiler detail, so strip it out.
This also leaves in RUSTC_ERROR_METADATA_DST since the stage0 compiler
still needs it to be set.
Move token gluing to token stream parsing
work towards #63689, this moves token gluing from the lexer to the token tree layer. This is only a minimal step, but I like the negative diff here.
r? @petrochenkov
rustc_mir: fix miri substitution/"universe" discipline.
Alternative to #61041, based on @RalfJung's own attempt at it.
I haven't done a full audit, but I believe everything is fixed now.
Fixes#61432.
Closes#61336, as a drive-by fix (for a subset of #43408, that is already special-cased).
r? @oli-obk / @RalfJung cc @varkor @yodaldevoid
This makes `rustc` support `@path` arguments on the command line. The `path` is opened and the file is interpreted
as new command line options which are logically inserted at that point in the command-line. The options in the file
are one per line. The file is UTF-8 encoded, and may have either Unix or Windows line endings.
It does not support recursive use of `@path`.
This is useful for very large command lines, or when command-lines are being generated into files by other tooling.
In sys/unix/process, we work around the sigemptyset linking issues
on android in two different ways. This change consolidates these
workarounds, and avoids duplicating bindings from `libc`.
This patch supports less behavior than before, since specifiying stage 1
vs stage 2 is no longer possible, but that is presumably a somewhat rare
use case anyway, so not supporting it seems acceptable (and it can be
readded easily if desired).
