Map invalid Spans to DUMMY_SP during crate metadata encoding.
This mirrors what we do for stabilizing the incr. comp. cache and is necessary for reproducible builds. For the incr. comp. cache, we *have* to do this because the encoding there cannot represent broken Spans. Metadata encoding has to be in sync with that as not to get unexpected interactions when compiling incrementally.
This should help with fixing issue https://github.com/rust-lang/rust/issues/47066.
r? @alexcrichton
ci: use a shared script to build musl
The dist-x86_64-musl, dist-various-1 and dist-i586-gnu-i686-musl builders had different scripts to build musl. This PR creates an unified script, which makes it easier to add new musl targets and update musl and libunwind (used in the musl targets).
The libunwind is update from 3.7 to 3.9 for dist-x86_64-musl and dist-i586-gnu-i686-musl (dist-various-1 already used 3.9 version).
The ref was actually always necessary.
This clarifies some incorrect thinking in the original code that might
have led to errors in the future to do with unsizing.
Previously, on a type mismatch (and if this wasn't preëmpted by a
higher-priority suggestion), we would look for argumentless methods
returning the expected type, and list them in a `help` note.
This had two major shortcomings. Firstly, a lot of the suggestions didn't
really make sense (if you used a &str where a String was expected,
`.to_ascii_uppercase()` is probably not the solution you were hoping
for). Secondly, we weren't generating suggestions from the most useful
traits!
We address the first problem with an internal
`#[rustc_conversion_suggestion]` attribute meant to mark methods that keep
the "same value" in the relevant sense, just converting the type. We
address the second problem by making `FnCtxt.probe_for_return_type` pass
the `ProbeScope::AllTraits` to `probe_op`: this would seem to be safe
because grep reveals no other callers of `probe_for_return_type`.
Also, structured suggestions are preferred (because they're pretty, but
also for RLS and friends).
Also also, we make the E0055 autoderef recursion limit error use the
one-time-diagnostics set, because we can potentially hit the limit a lot
during probing. (Without this,
test/ui/did_you_mean/recursion_limit_deref.rs would report "aborting due to
51 errors").
Unfortunately, the trait probing is still not all one would hope for: at a
minimum, we don't know how to rule out `into()` in cases where it wouldn't
actually work, and we don't know how to rule in `.to_owned()` where it
would. Issues #46459 and #46460 have been filed and are ref'd in a FIXME.
This is hoped to resolve#42929, #44672, and #45777.
This `horizontal_trim` function strips the leading whitespace from
doc-comments that have a left-asterisk-margin:
/**
* You know what I mean—
*
* comments like this!
*/
The index of the column of asterisks is `i`, and if trimming is deemed
possible, we slice each line from `i+1` to the end of the line. But if, in
particular, `i` was 0 _and_ there was an empty line (as in the example
given in the reporting issue), we ended up panicking trying to slice an
empty string from 0+1 (== 1).
Let's tighten our check to say that we can't trim when `i` is even the same
as the length of the line, not just when it's greater. (Any such cases
would panic trying to slice `line` from `line.len()+1`.)
Resolves#47197.
Use name-discarding LLVM context
This is only applicable when neither of --emit=llvm-ir or --emit=llvm-bc are not
requested.
In case either of these outputs are wanted, but the benefits of such context are
desired as well, -Zfewer_names option provides the same functionality regardless
of the outputs requested.
Should be a viable fix for https://github.com/rust-lang/rust/issues/46449
Port libpanic_abort and libpanic_unwind to CloudABI
This change ports both the libpanic* libraries to CloudABI.
The most interesting part of this pull request, however, is that it imports the CloudABI system call API into the Rust tree through a Git submodule. These will also be used by my port of libstd to CloudABI extensively, as that library obviously needs to invoke system calls to implement its primitives.
I have taken the same approach as libc: `src/libcloudabi` + `src/rustc/cloudabi_shim`. If some other naming scheme is preferred, feel free to let me know! As `libcloudabi` is pretty small, maybe it makes sense to copy, instead of using a submodule?
rustc: use {U,I}size instead of {U,I}s shorthands.
`Us`/`Is` come from a time when `us` and `is` were the literal suffixes that are now `usize` / `isize`.
r? @nikomatsakis
doc: improve None condition doc for `checked_div` and `checked_rem`
This commit improves the condition mentioned in the docs for which `checked_div` and `checked_rem` return `None`.
For signed division, the commit changes "the operation results in overflow" to "the division results in overflow", otherwise there is room for misinterpretation for `checked_rem`: Without considering overflow, `MIN % -1` would be simply zero, allowing the misinterpretation that "the operation" does not result in overflow in this case. This ambiguity is removed using "when the division results in overflow".
For unsigned division, the condition for `None` should be simply when `rhs == 0`, as no other overflow is possible.
Bump to 1.25.0
* Bump the release version to 1.25
* Bump the bootstrap compiler to the recent beta
* Allow using unstable rustdoc features on beta - this fix has been applied to
the beta branch but needed to go to the master branch as well.
Issue 46976
ICE is due to an empty path segments, so I set the path to be the same as the in band ty params symbol. (I think this is how regular generics end up being handled?)
Pinging @cramertj, this is your code I'm editing here.
Restore working debuginfo tests by trimming comments from non-header directive lines
I noticed when adding a debuginfo test that nothing I did caused the test to fail. Tracing back this seems to have been caused by 3e6c83de1d which broke parsing of the command/check lines, leaving all tests passing without any checking. This commit provides a basic (although still not very robust) restoration of tests and a should-fail test which checks the parser is running
[unix] Don't clone command-line args on startup
Fixes part of #47164 and simplifies the `args` code on non-Apple Unix platforms.
Note: This could change behavior for programs that use both `std::env::args` *and* unsafe code that mutates `argv` directly. However, these programs already behave differently on different platforms. The new behavior on non-Apple platforms is closer to the existing behavior on Apple platforms.
Minor rewrite of env::current_exe docs; clarify symlinks.
- Update example in ‘security’ section to use hard links, like the
linked securityvulns.com example.
- Weaken language on symbolic links – indicate behavior is
platform-specific
Fixes https://github.com/rust-lang/rust/issues/43617.
Ideally, we should make use of CloudABI's internal proc_raise(SIGABRT)
system call. POSIX abort() requires things like flushing of stdios,
which may not be what we want under panic conditions. Invoking the raw
CloudABI system call would have prevented that.
Unfortunately, we have to make use of the "cloudabi" crate to invoke raw
CloudABI system calls. This is undesired, as discussed in the pull
request (#47190).
