Ensure derive(PartialOrd) is no longer accidentally exponential
Previously, two comparison operations would be generated for each field, each of which could delegate to another derived PartialOrd. Now we use ordering and optional chaining to ensure each pair of fields is only compared once, addressing https://github.com/rust-lang/rust/issues/49650#issuecomment-379467572.
Closes#49505.
r? @Manishearth (sorry for changing it again so soon!)
Close#50755
typeck: Save the index of private fields
Save the index of all fields regardless of their visibility. Problems
could occur later when attempting to index fields in error recovery if
they are not inserted.
Fixes: #50493
It was introduced in #50240 to avoid an allocation when creating a new
BTreeMap, which gave some speed-ups. But then #50352 made that the
default behaviour for BTreeMap, so LazyBTreeMap is no longer necessary.
On growth, Vec does not require to exactly double its size for correctness,
like, for example, VecDeque does.
Using reserve instead better expresses this intent. It also allows to reuse
Excess capacity on growth and for better growth-policies to be provided by
RawVec.
Previously, every `open64` was accompanied by a `ioctl(…, FIOCLEX)`,
because some old Linux version would ignore the `O_CLOEXEC` flag we pass
to the `open64` function.
Now, we check whether the `CLOEXEC` flag is set on the first file we
open – if it is, we won't do extra syscalls for every opened file. If it
is not set, we fall back to the old behavior of unconditionally calling
`ioctl(…, FIOCLEX)` on newly opened files.
On old Linuxes, this amounts to one extra syscall per process, namely
the `fcntl(…, F_GETFD)` call to check the `CLOEXEC` flag.
On new Linuxes, this reduces the number of syscalls per opened file by
one, except for the first file, where it does the same number of
syscalls as before (`fcntl(…, F_GETFD)` to check the flag instead of
`ioctl(…, FIOCLEX)` to set it).
update libcompiler_builtins
let's OpenBSD to use libcompiler_rt.a from system library. it unbreaks
the build from source on OpenBSD.
see https://github.com/rust-lang-nursery/compiler-builtins/pull/241 for related PR
note this PR brings some other changes (on `floatdisf`/`floatundisf`) with compiler-builtins update.
r? @alexcrichton
Fix self referential impl Trait substitutions
A high impact bug because a lot of common traits use a `Self` substitution by default. Should be backported to beta.
There was a check for this which wasn't catching all cases, it was made more robust.
Fixes#49376Fixes#50626
r? @petrochenkov
It only has a single use, within code handling indented block comments.
We can replace that with the new `FileMap::col_pos()`, which computes
the col position (BytePos instead of CharPos) based on the record of the
last newline char (which we already record).
This is actually an improvement, because
`trim_whitespace_prefix_and_push_line()` was using `col`, which is a
`CharPos`, as a slice index, which is a byte/char confusion.
Save the index of all fields regardless of their visibility. Problems
could occur later when attempting to index fields in error recovery if
they are not inserted.
This PR improves the span of eager resolution type errors referring to indexing and field access to use the base span rather than the whole expression.
Also a note "Type must be known at this point." is added to where we at some point in the past emitted the "type must be known at this context" error, so that early failures can be differentiated and will hopefully be less surprising.
Fixes#50692 (or at least does the best we can for the moment)
r? @estebank
