bors [Fri, 15 Jan 2016 06:38:26 +0000 (06:38 +0000)]
Auto merge of #30763 - gchp:issue/30033, r=nagisa
This is achieved by adding the scan_back method. This method looks back
through the source_text of the StringReader until it finds the target
char, returning it's offset in the source. We use this method to find
the offset of the opening single quote, and use that offset as the start
of the error.
Given this code:
```rust
fn main() {
let _ = 'abcd';
}
```
The compiler would give a message like:
```
error: character literal may only contain one codepoint: ';
let _ = 'abcd';
^~
```
With this change, the message now displays:
```
error: character literal may only contain one codepoint: 'abcd';
let _ = 'abcd';
^~~~~~~
```
Nick Cameron [Wed, 6 Jan 2016 20:23:01 +0000 (09:23 +1300)]
Address reviewer comments
[breaking-change]
`OptLevel` variants are no longer `pub use`ed by rust::session::config. If you are using these variants, you must change your code to prefix the variant name with `OptLevel`.
Nick Cameron [Thu, 31 Dec 2015 05:47:14 +0000 (18:47 +1300)]
Implement JSON error emission
[breaking-change]
syntax::errors::Handler::new has been renamed to with_tty_emitter
Many functions which used to take a syntax::errors::ColorConfig, now take a rustc::session::config::ErrorOutputType. If you previously used ColorConfig::Auto as a default, you should now use ErrorOutputType::default().
Steve Klabnik [Thu, 14 Jan 2016 19:58:34 +0000 (14:58 -0500)]
Rollup merge of #30903 - steveklabnik:no_std_stable, r=alexcrichton
This feature is partially stabilized, so describe each part in the appropriate place.
r? @alexcrichton @brson
It would be nice to backport this to beta, since this is the first release where this is true. I try really hard to not do doc backports, but this isn't very large, and might be worth making an exception, I dunno.
Use arena allocation instead of reference counting for `Module`s to fix memory leaks from `Rc` cycles.
A module references its module children and its import resolutions, and an import resolution references the module defining the imported name, so there is a cycle whenever a module imports something from an ancestor module.
For example,
```rust
mod foo { // `foo` references `bar`.
fn baz() {}
mod bar { // `bar` references the import.
use foo::baz; // The import references `foo`.
}
}
```
Rollup merge of #30851 - jonas-schievink:unneeded-dropflags, r=pnkfelix
Apparently we allocate and maintain non-working dropflag hints since June... In anticipation of a working implementation of on-stack drop flag hints, let's not spend even more time on types that don't even need to be dropped.
Rollup merge of #30850 - ranma42:cleanup-io, r=alexcrichton
In 8d90d3f36871a00023cc1f313f91e351c287ca15 `BufStream`, the only
consumer of `InternalBufWriter`, was removed. As implied by the name,
this type is private, hence it is currently dead code.
Rollup merge of #30839 - tomaka:debug-phantomdata, r=nikomatsakis
All the trait implementations of `PhantomData` use `impl<T: ?Sized>` except for `Debug`
https://doc.rust-lang.org/nightly/std/marker/struct.PhantomData.html#implementations
Rollup merge of #30837 - semarie:openbsd-libc, r=alexcrichton
The following PR updates libc version to latest commits for correctly support openbsd.
It corrects several points in rustc to be compatible with libc changes.
Rollup merge of #30823 - pnkfelix:put-back-alloca-zeroing-for-issue-30530, r=dotdash
Put back alloca zeroing for issues #29092, #30018, #30530; inject zeroing for #30822.
----
Background context: `fn alloca_zeroed` was removed in PR #22969, so we haven't been "zero'ing" (\*) the alloca's since at least that point, but the logic behind that PR seems sound, so its not entirely obvious how *long* the underlying bug has actually been present. In other words, I have not yet done a survey to see when the new `alloc_ty` and `lvalue_scratch_datum` calls were introduced that should have had "zero'ing" the alloca's.
----
I first fixed #30018, then decided to do a survey of `alloc_ty` calls to see if they needed similar treatment, which quickly led to a rediscovery of #30530.
While making the regression test for the latter, I discovered #30822, which is a slightly different bug (in terms of where the "zero'ing" needs to go), but still relevant.
I haven't finished the aforementioned survey of `fn alloc_ty` calls, but I decided I wanted to get this up for review in its current state (namely to see if my attempt to force developers to include a justification for passing `Uninit` can possibly fly, or if I should abandon that path of action).
----
(*): I am putting quotation marks around "zero'ing" because we no longer use zero as our "dropped" marker value.
Rollup merge of #30801 - Amanieu:oom_print, r=alexcrichton
This adds the ability to override the default OOM behavior by setting a handler function. This is used by libstd to print a message when running out of memory instead of crashing with an obscure "illegal hardware instruction" error (at least on Linux).
bors [Thu, 14 Jan 2016 01:26:54 +0000 (01:26 +0000)]
Auto merge of #30466 - alexcrichton:move-wrapping-and-fill-out, r=aturon
This commit migrates all of the methods on `num::wrapping::OverflowingOps` onto
inherent methods of the integer types. This also fills out some missing gaps in
the saturating and checked departments such as:
bors [Wed, 13 Jan 2016 22:27:46 +0000 (22:27 +0000)]
Auto merge of #30870 - Eljay:issue-30477, r=alexcrichton
Fixes #30477, #30213.
The loop over reexports used to be a closure before #30043 but it's an iterator now so it should just continue instead of exiting the loop and skipping stuff.
(Note that it might be a good idea to replace *all* calls of
`alloc_ty` with calls to `alloc_ty_init`, to encourage programmers to
consider the appropriate value for the `init` flag when creating
temporary values.)
* Only `call_lifetime_start` for an alloca if the function entry does
not itself initialize it to "dropped."
* Remove `schedule_lifetime_end` after writing an *element* into a
borrowed slice. (As explained by [dotdash][irc], "the lifetime end
that is being removed was for an element in the slice, which is not
an alloca of its own and has no lifetime start of its own")
bors [Wed, 13 Jan 2016 08:15:45 +0000 (08:15 +0000)]
Auto merge of #29498 - wthrowe:replace-pattern, r=alexcrichton
It appears this was left out of RFC rust-lang/rfcs#528 because it might be useful to
also generalize the second argument in some way. That doesn't seem to
prevent generalizing the first argument now, however.
This is a [breaking-change] because it could cause type-inference to
fail where it previously succeeded.
Also update docs for a few other methods that still referred to `&str` instead of patterns.
bors [Wed, 13 Jan 2016 02:05:02 +0000 (02:05 +0000)]
Auto merge of #30639 - rkruppe:dec2flt-fastpath-tables, r=alexcrichton
Add tables of small powers of ten used in the fast path. The tables are redundant: We could also use the big, more accurate table and round the value to the correct type (in fact we did just that before this commit). However, the rounding is extra work and slows down the fast path.
Because only very small exponents enter the fast path, the table and thus the space overhead is negligible. Speed-wise, this is a clear win on a [benchmark] comparing the fast path to a naive, hand-optimized, inaccurate algorithm. Specifically, this change narrows the gap from a roughly 5x difference to a roughly 3.4x difference.
Anton Blanchard [Fri, 8 Jan 2016 06:42:51 +0000 (06:42 +0000)]
Incorrect getrandom() system call for PowerPC Linux
Michael Ellerman pointed out that the system call for getrandom()
on PowerPC Linux is incorrect. This bug was in the powerpc32 port,
and was carried over to the powerpc64 port too.
Robin Kruppe [Wed, 30 Dec 2015 13:01:42 +0000 (14:01 +0100)]
Speed up dec2flt fast path with additional tables.
Add tables of small powers of ten used in the fast path. The tables are redundant: We could also use the big, more accurate table and round the value to the correct type (in fact we did just that before this commit). However, the rounding is extra work and slows down the fast path.
Because only very small exponents enter the fast path, the table and thus the space overhead is negligible. Speed-wise, this is a clear win on a [benchmark] comparing the fast path to a naive, hand-optimized, inaccurate algorithm. Specifically, this change narrows the gap from a roughly 5x difference to a roughly 3.4x difference.
Andrea Canciani [Tue, 12 Jan 2016 10:58:52 +0000 (11:58 +0100)]
Remove dead `InternalBufWriter` implementation
In 8d90d3f36871a00023cc1f313f91e351c287ca15 `BufStream`, the only
consumer of `InternalBufWriter`, was removed. As implied by the name,
this type is private, hence it is currently dead code.
Issue 30530: initialize allocas for `Datum::to_lvalue_datum_in_scope`.
In particular, bring back the `zero` flag for `lvalue_scratch_datum`,
which controls whether the alloca's created immediately at function
start are uninitialized at that point or have their embedded
drop-flags initialized to "dropped".
Then made `to_lvalue_datum_in_scope` pass "dropped" as `zero` flag.
bors [Tue, 12 Jan 2016 08:30:56 +0000 (08:30 +0000)]
Auto merge of #30678 - Amanieu:no_elf_tls, r=alexcrichton
I also re-enabled the use of `#[thread_local]` on AArch64. It was originally disabled in the PR that introduced AArch64 (#19790), but the reasons for this were not explained. `#[thread_local]` seems to work fine in my tests on AArch64, so I don't think this should be an issue.