Alex Crichton [Wed, 31 Dec 2014 00:25:47 +0000 (16:25 -0800)]
rollup merge of #20065: aturon/stab-2-cmp
This patch marks `PartialEq`, `Eq`, `PartialOrd`, and `Ord` as
`#[stable]`, as well as the majorify of manual implementaitons of these
traits. The traits match the [reform RFC](https://github.com/rust-lang/rfcs/pull/439).
In the future, many of the impls should be generalized; see #20063.
However, there is no problem stabilizing the less general impls, since
generalizing later is not a breaking change.
Alex Crichton [Wed, 31 Dec 2014 00:25:45 +0000 (16:25 -0800)]
rollup merge of #20045: alexcrichton/issue-20012
The new semantics of this function are that the callbacks are run when the *main
thread* exits, not when all threads have exited. This implies that other threads
may still be running when the `at_exit` callbacks are invoked and users need to
be prepared for this situation.
Users in the standard library have been audited in accordance to these new rules
as well.
Alex Crichton [Wed, 31 Dec 2014 00:25:44 +0000 (16:25 -0800)]
rollup merge of #19954: michaelwoerister/rust-gdb
This pull request adds the `rust-gdb` shell script which starts GDB with Rust pretty printers enabled. The PR also makes `rustc` add a special `.debug_gdb_scripts` ELF section on Linux which tells GDB that the produced binary should use the Rust pretty printers.
Note that at the moment this script will only work and be installed on Linux. On Mac OS X there's `rust-lldb` which works much better there. On Windows I had too many problems making this stable. I'll give it another try soonish.
You can use this script just like you would use GDB from the command line. It will use the pretty printers from the Rust "installation" found first in PATH. E.g. if you have `~/rust/x86_64-linux-gnu/stage1/bin` in your path, it will use the pretty printer scripts in `~/rust/x86_64-linux-gnu/stage1/lib/rustlib/etc`.
Aaron Turon [Tue, 30 Dec 2014 22:44:26 +0000 (14:44 -0800)]
Stabilize cmp
This patch marks `PartialEq`, `Eq`, `PartialOrd`, and `Ord` as
`#[stable]`, as well as the majorify of manual implementaitons of these
traits. The traits match the [reform
RFC](https://github.com/rust-lang/rfcs/pull/439).
Along the way, two changes are made:
* The recently-added type parameters for `Ord` and `Eq` are
removed. These were mistakenly added while adding them to `PartialOrd`
and `PartialEq`, but they don't make sense given the laws that are
required for (and use cases for) `Ord` and `Eq`.
* More explicit laws are added for `PartialEq` and `PartialOrd`,
connecting them to their associated mathematical concepts.
In the future, many of the impls should be generalized; see
since generalizing later is not a breaking change.
Alex Crichton [Fri, 19 Dec 2014 19:29:39 +0000 (11:29 -0800)]
std: Re-enable at_exit()
The new semantics of this function are that the callbacks are run when the *main
thread* exits, not when all threads have exited. This implies that other threads
may still be running when the `at_exit` callbacks are invoked and users need to
be prepared for this situation.
Users in the standard library have been audited in accordance to these new rules
as well.
bors [Tue, 30 Dec 2014 17:51:21 +0000 (17:51 +0000)]
auto merge of #20307 : nikomatsakis/rust/assoc-types-normalization-extend-bound, r=nrc
Rewrite associated types to use projection rather than dummy type parameters. This closes almost every (major) open issue, but I'm holding off on that until the code has landed and baked a bit. Probably it should have more tests, as well, but I wanted to get this landed as fast as possible so that we can collaborate on improving it.
The commit history is a little messy, particularly the merge commit at the end. If I get some time, I might just "reset" to the beginning and try to carve up the final state into logical pieces. Let me know if it seems hard to follow. By far the most crucial commit is "Implement associated type projection and normalization."
Niko Matsakis [Tue, 30 Dec 2014 11:29:59 +0000 (06:29 -0500)]
Normalize associated types in bounds too. Also, make the workaround
for lack of impl-trait-for-trait just a bit more targeted (don't
substitute err, just drop the troublesome bound for now) -- otherwise
substituting false types leads us into trouble when we normalize etc.
Niko Matsakis [Mon, 29 Dec 2014 16:22:16 +0000 (11:22 -0500)]
Adjust tests for inferenceGet more conservative about inference for now. Seems better to err on the side of being more correct rather than less. Fix a bug in typing index expressions that was exposed as a result, and add one type annotation that is not required. Delete some random tests that were relying on old behavior and don't seem to add anything anymore.
Niko Matsakis [Sun, 28 Dec 2014 00:42:27 +0000 (19:42 -0500)]
Remove the def-id from type parameters. Having this def-id was bad for several reasons:
1. Produced more unique types than is necessary. This increases memory consumption.
2. Linking the type parameter to its definition *seems* like a good idea, but it
encourages reliance on the bounds listing.
3. It made pretty-printing harder and in particular was causing bad error messages
when errors occurred before the `TypeParameterDef` entries were fully stored.
Niko Matsakis [Tue, 23 Dec 2014 10:52:47 +0000 (05:52 -0500)]
Rename `Polytype` to `TypeScheme` to differentiate type schemes (early bound) from higher-ranked things (late-bound), which also use the `Poly` prefix.
Niko Matsakis [Thu, 18 Dec 2014 14:26:10 +0000 (09:26 -0500)]
Move the TypeContents-based "Sized" queries into trans, where the full
types are always known and hence the ParameterEnvironment is not
necessary. For other `Sized` queries, use the trait infrastructure
just like `Copy`.
bors [Tue, 30 Dec 2014 11:11:07 +0000 (11:11 +0000)]
auto merge of #19941 : alexcrichton/rust/issue-19767, r=brson
This commit adds support for the compiler to distinguish between different forms
of lookup paths in the compiler itself. Issue #19767 has some background on this
topic, as well as some sample bugs which can occur if these lookup paths are not
separated.
This commits extends the existing command line flag `-L` with the same trailing
syntax as the `-l` flag. Each argument to `-L` can now have a trailing `:all`,
`:native`, `:crate`, or `:dependency`. This suffix indicates what form of lookup
path the compiler should add the argument to. The `dependency` lookup path is
used when looking up crate dependencies, the `crate` lookup path is used when
looking for immediate dependencies (`extern crate` statements), and the `native`
lookup path is used for probing for native libraries to insert into rlibs. Paths
with `all` are used for all of these purposes (the default).
The default compiler lookup path (the rustlib libdir) is by default added to all
of these paths. Additionally, the `RUST_PATH` lookup path is added to all of
these paths.
Brian Anderson [Mon, 29 Dec 2014 19:15:40 +0000 (11:15 -0800)]
Install copyright information and package docs
This distributes docs in a separate package called rust-docs. The rust-packaging
project will combine it with Rust and Cargo into a single installer in a variety of formats.
bors [Tue, 30 Dec 2014 00:42:13 +0000 (00:42 +0000)]
auto merge of #20160 : nick29581/rust/ranges2, r=nikomatsakis
The first six commits are from an earlier PR (#19858) and have already been reviewed. This PR makes an awful hack in the compiler to accommodate slices both natively and in the index a range form. After a snapshot we can hopefully add the new Index impls and then we can remove these awful hacks.
r? @nikomatsakis (or anyone who knows the compiler, really)
Alex Crichton [Tue, 30 Dec 2014 00:36:52 +0000 (16:36 -0800)]
rollup merge of #20306: alexcrichton/second-pass-string
This commit performs a second pass over the `std::string` module, performing the
following actions:
* The name `std::string` is now stable.
* The `String::from_utf8` function is now stable after having been altered to
return a new `FromUtf8Error` structure. The `FromUtf8Error` structure is now
stable as well as its `into_bytes` and `utf8_error` methods.
* The `String::from_utf8_lossy` function is now stable.
* The `String::from_chars` method is now deprecated in favor of `.collect()`
* The `String::from_raw_parts` method is now stable
* The `String::from_str` function remains experimental
* The `String::from_raw_buf` function remains experimental
* The `String::from_raw_buf_len` function remains experimental
* The `String::from_utf8_unchecked` function is now stable
* The `String::from_char` function is now deprecated in favor of
`repeat(c).take(n).collect()`
* The `String::grow` function is now deprecated in favor of
`.extend(repeat(c).take(n)`
* The `String::capacity` method is now stable
* The `String::reserve` method is now stable
* The `String::reserve_exact` method is now stable
* The `String::shrink_to_fit` method is now stable
* The `String::pop` method is now stable
* The `String::as_mut_vec` method is now stable
* The `String::is_empty` method is now stable
* The `IntoString` trait is now deprecated (there are no implementors)
* The `String::truncate` method is now stable
* The `String::insert` method is now stable
* The `String::remove` method is now stable
* The `String::push` method is now stable
* The `String::push_str` method is now stable
* The `String::from_utf16` function is now stable after its error type has now
become an opaque structure to carry more semantic information in the future.
A number of these changes are breaking changes, but the migrations should be
fairly straightforward on a case-by-case basis (outlined above where possible).
Alex Crichton [Tue, 30 Dec 2014 00:36:41 +0000 (16:36 -0800)]
rollup merge of #20276: alexcrichton/deprecate-term
This library is now published on crates.io as the `term` crate, so the in-tree
version is now deprecated. Once stability warnings are enabled, this library
will automatically be gated.
Alex Crichton [Tue, 30 Dec 2014 00:36:38 +0000 (16:36 -0800)]
rollup merge of #20271: vhbit/datalayout-fix
According to http://llvm.org/docs/LangRef.html#data-layout correct syntax
for data layout is `a:<abi>:<pref>` so it looks like `a0:<abi>:<pref>` is
either a typo or outdated syntax (as it goes back pretty deep in time)
Alex Crichton [Tue, 30 Dec 2014 00:36:30 +0000 (16:36 -0800)]
rollup merge of #20265: nicholasbishop/bishop_add_missing_bitflags_methods
The methods `from_bits` and `from_bits_truncate` were missing from the
list of generated methods. Didn't see a useful way to abbreviate, so
added with the same docstrings used in the macro definition.
Alex Crichton [Tue, 30 Dec 2014 00:36:27 +0000 (16:36 -0800)]
rollup merge of #20263: crhino/lifetime-elision-help
Fixes #19707.
In terms of output, it currently uses the form `argument #1`, `argument #2`, etc. If anyone has any better suggestions I would be glad to consider them.
Alex Crichton [Tue, 30 Dec 2014 00:36:22 +0000 (16:36 -0800)]
rollup merge of #20250: ipetkov/deriving
* Both enums already derived `Copy`, but storing them in any
struct/container would prevent implementing `Clone` for said
struct/container even though they should be clonable.
* Also add PartialEq and Eq for good measure.
Alex Crichton [Tue, 30 Dec 2014 00:36:18 +0000 (16:36 -0800)]
rollup merge of #20245: fhahn/make-lexer-tests-runable-again
I would like to look into some issues related to the model lexer #15883.
I stumbled upon 2 minor problems when I tried running the lexer tests:
* antlr did not put the generated files in the correct directory
* grammer/verify.rs did not work with the most recent version of rust
With these changes (and setting CLASSPATH=/usr/share/java/antlr-4.4-complete.jar:$CLASSPATH) I was able to execute the tests.
Note that I just fixed the syntax errors and added `None` as 2. argument of `Literal`. I am not sure if this is correct however. I still have to take a closer look at what verify.rs actually does. Are there any helpful pointers?
Alex Crichton [Tue, 30 Dec 2014 00:36:12 +0000 (16:36 -0800)]
rollup merge of #20231: fhahn/issue-20226-eexist
I've created a patch for #20226, which maps `EEXIST` to the `PathAlreadyExists` error on Unix. To test this, I use `mkdir`, which raises `EEXIST` if the directory already exists.
On Windows, I map `ERROR_ALREADY_EXISTS` to `PathAlreadyExist`, but I am note sure if `mkdir` on Windows raises `ERROR_ALREADY_EXISTS` and do not have a Windows installation handy for testing.
And I noticed another thing. No error seems to map to `IoErrorKind::PathDoesntExist` and I am wondering what the difference to `FileNotFound` is?
Alex Crichton [Tue, 30 Dec 2014 00:35:53 +0000 (16:35 -0800)]
rollup merge of #20160: nick29581/ranges2
The first six commits are from an earlier PR (#19858) and have already been reviewed. This PR makes an awful hack in the compiler to accommodate slices both natively and in the index a range form. After a snapshot we can hopefully add the new Index impls and then we can remove these awful hacks.
r? @nikomatsakis (or anyone who knows the compiler, really)
Alex Crichton [Tue, 30 Dec 2014 00:35:51 +0000 (16:35 -0800)]
rollup merge of #20042: alexcrichton/second-pass-ptr
This commit performs a second pass for stabilization over the `std::ptr` module.
The specific actions taken were:
* The `RawPtr` trait was renamed to `PtrExt`
* The `RawMutPtr` trait was renamed to `PtrMutExt`
* The module name `ptr` is now stable.
* These functions were all marked `#[stable]` with no modification:
* `null`
* `null_mut`
* `swap`
* `replace`
* `read`
* `write`
* `PtrExt::is_null`
* `PtrExt::is_not_null`
* `PtrExt::offset`
* These functions remain unstable:
* `as_ref`, `as_mut` - the return value of an `Option` is not fully expressive
as null isn't the only bad value, and it's unclear
whether we want to commit to these functions at this
time. The reference/lifetime semantics as written are
also problematic in how they encourage arbitrary
lifetimes.
* `zero_memory` - This function is currently not used at all in the
distribution, and in general it plays a broader role in the
"working with unsafe pointers" story. This story is not yet
fully developed, so at this time the function remains
unstable for now.
* `read_and_zero` - This function remains unstable for largely the same
reasons as `zero_memory`.
* These functions are now all deprecated:
* `PtrExt::null` - call `ptr::null` or `ptr::null_mut` instead.
* `PtrExt::to_uint` - use an `as` expression instead.
Alex Crichton [Tue, 30 Dec 2014 00:35:50 +0000 (16:35 -0800)]
rollup merge of #19661: alexcrichton/mutex-result
All of the current std::sync primitives have poisoning enable which means that
when a task fails inside of a write-access lock then all future attempts to
acquire the lock will fail. This strategy ensures that stale data whose
invariants are possibly not upheld are never viewed by other tasks to help
propagate unexpected panics (bugs in a program) among tasks.
Currently there is no way to test whether a mutex or rwlock is poisoned. One
method would be to duplicate all the methods with a sister foo_catch function,
for example. This pattern is, however, against our [error guidelines][errors].
As a result, this commit exposes the fact that a task has failed internally
through the return value of a `Result`.
All methods now return a `LockResult<T>` or a `TryLockResult<T>` which
communicates whether the lock was poisoned or not. In a `LockResult`, both the
`Ok` and `Err` variants contains the `MutexGuard<T>` that is being returned in
order to allow access to the data if poisoning is not desired. This also means
that the lock is *always* held upon returning from `.lock()`.
A new type, `PoisonError`, was added with one method `into_guard` which can
consume the assertion that a lock is poisoned to gain access to the underlying
data.
This is a breaking change because the signatures of these methods have changed,
often incompatible ways. One major difference is that the `wait` methods on a
condition variable now consume the guard and return it in as a `LockResult` to
indicate whether the lock was poisoned while waiting. Most code can be updated
by calling `.unwrap()` on the return value of `.lock()`.