Commit Graph

23311 Commits

Author SHA1 Message Date
Kevin Ballard 550bc9bd1c path2: Reimplement PosixPath in terms of ~[u8] 2013-10-15 20:10:10 -07:00
Kevin Ballard b413913066 path2: Start reimplementing in terms of ~[u8] instead of ~str
As documented in #7225, we cannot rely on paths being representable in
utf-8. Specifically, Linux allows anything (besides NUL) in a path.
Redesign GenericPath in light of this.

PosixPath hasn't been reimplemented yet for ~[u8].
2013-10-15 20:10:10 -07:00
Kevin Ballard e97d61672b path2: Implement PosixPath
Fixes #5389 (new conventions for Path constructor)
2013-10-15 20:10:10 -07:00
Kevin Ballard d202834092 Initial sketching out of the new path module
Define the basic types, and the GenericPath trait.

This module is currently called path2. It will be renamed later.
2013-10-15 20:10:10 -07:00
Steven Fackler 194302493c Remove extra::flatpipes
Closes #9884
2013-10-15 18:54:35 -07:00
Daniel Micay 1e128d7931 remove executable flag from source file 2013-10-15 19:34:45 -04:00
Daniel Micay e1a26ad271 use element count in slices, not size in bytes
This allows the indexing bounds check or other comparisons against an
element length to avoid a multiplication by the size.
2013-10-15 16:23:28 -04:00
Daniel Micay aa93381e14 fix overflow on bounds checks
Closes #9020
2013-10-15 16:23:28 -04:00
Daniel Micay 420b4260b4 fix bounds checking failure message
casting the `uint` to an `int` can result in printing high values as
negative intege
2013-10-15 16:23:26 -04:00
bors 1252ad4092 auto merge of #9860 : thestinger/rust/abort, r=Aatch
This should be preferred to calling the libc `abort` function.
2013-10-14 19:51:34 -07:00
Daniel Micay 45e2cf39a2 add an `abort` intrinsic
This should be preferred to the libc `abort` function.
2013-10-14 22:45:43 -04:00
Alex Crichton 4d2a402555 rustdoc: Don't abort if there's nothing to document
Instead there's no index.html file emitted because there wasn't anything to
document.

Closes #9828
2013-10-14 18:00:54 -07:00
Alex Crichton 3d693d74b8 rustdoc: Use privacy visibility for pruning
This commit ends rustdoc's approximation of privacy and instead uses the result
of the various compiler passes instead.

Closes #9827
2013-10-14 18:00:54 -07:00
Brian Anderson 818ebf2ed6 std::rt: Fix the set up of the main thread so that it doesn't try to steal work
This is causing really awful scheduler behavior where the main thread scheduler is
continually waking up, stealing work, discovering it can't actually run the work,
and sending it off to another scheduler.
2013-10-14 16:08:18 -07:00
Alex Crichton 082cc96090 Refine privacy error messages to be more accurate
This stops labeling everything as "is private" when in fact the destination may
be public. Instead, the clause "is inaccessible" is used and the private part of
the flag is called out with a "is private" message.

Closes #9793
2013-10-14 15:12:40 -07:00
bors a7e8957c59 auto merge of #9823 : pnkfelix/rust/issue7655-bench-papercuts, r=alexcrichton
r? anyone.

You can see a bit more discussion on #7655.

This does not close any ticket; I am just scratching an itch.  Note in particular that I picked the value `{:>9} ns/iter` pretty much out of a hat.  :)
2013-10-14 15:06:36 -07:00
Felix S. Klock II 602b3cd56c Only use padded test names to calculate the target padding size. 2013-10-14 17:19:15 -04:00
Felix S. Klock II c56bf67150 Issue 7655: align the bench printouts so that the numbers tend to be aligned.
(scratching an itch.)

Rebased and updated.
Fixed bug: omitted field init from embedded struct literal in a test.
Fixed bug: int underflow during padding length calculation.
2013-10-14 17:18:34 -04:00
bors c8e77d5586 auto merge of #9606 : steveklabnik/rust/abi_removal, r=alexcrichton
They've been replaced by putting the name on the extern block.

  #[abi = "foo"]

goes to

  extern "foo" { }

Closes #9483.
2013-10-14 07:26:47 -07:00
bors b571039021 auto merge of #9848 : huonw/rust/move-trait-doc, r=alexcrichton
This means the text is visible in rustdoc.
2013-10-14 06:16:32 -07:00
Steve Klabnik 309ab958e6 Removing ccdecl
as per https://github.com/mozilla/rust/pull/9606#discussion_r6930872
2013-10-14 14:33:05 +02:00
Huon Wilson a6d7fe6209 std::vec: move documentation from impls to traits.
This means the text is visible in rustdoc.
2013-10-14 22:21:47 +11:00
Steve Klabnik 16fc6a694c Remove unused abi attributes.
They've been replaced by putting the name on the extern block.

  #[abi = "foo"]

goes to

  extern "foo" { }

Closes #9483.
2013-10-14 13:10:36 +02:00
bors 5b10781c7b auto merge of #9844 : bytbox/rust/master, r=alexcrichton 2013-10-13 23:41:30 -07:00
Corey Richardson 023466ccca extra: implement Encodable/Decodable for Uuid 2013-10-14 01:47:55 -04:00
bors 6aa556db3b auto merge of #9843 : sfackler/rust/rustdoc-strip-private, r=alexcrichton
In addition, the renderer will add comments to structs and enums saying
that fields or variants have been stripped.

The comments are currently
```rust
// some fields stripped
// some variants stripped
```

I was thinking of changing that to "some private..." but passes other than
strip-private may strip fields or variants as well.

cc @alexcrichton
2013-10-13 22:31:29 -07:00
Steven Fackler d6d31d788d Make Rustdoc strip private fields
In addition, the renderer will add comments to structs and enums saying
that fields or variants have been stripped.
2013-10-13 22:28:45 -07:00
Scott Lawrence 85f19a845d std::libc: rustdoc indicates reexports now 2013-10-14 00:48:34 -04:00
bors 009c3d8bae auto merge of #9836 : eriklyon/rust/doc-typos, r=alexcrichton 2013-10-13 11:31:26 -07:00
Erik Lyon 8b65a45879 fix typos in doc/tutorial.md 2013-10-13 10:49:44 -07:00
bors 91abfd425d auto merge of #9832 : luqmana/rust/sps, r=alexcrichton
Fixes #9830.
2013-10-12 21:26:20 -07:00
Luqman Aden e88064d18f librustc: Combine C_struct and C_packed_struct. 2013-10-12 23:19:22 -04:00
bors 0bad7e1a37 auto merge of #9608 : hmarr/rust/vec-get-opt, r=huonw
This adds `get_opt` to `std::vec`, which looks up an item by index and returns an `Option`. If the given index is out of range, `None` will be returned, otherwise a `Some`-wrapped item will be returned.

Example use case:

```rust
use std::os;

fn say_hello(name: &str) {
  println(fmt!("Hello, %s", name));
}

fn main(){
  // Try to get the first cmd line arg, but default to "World"
  let args = os::args();
  let default = ~"World";
  say_hello(*args.get_opt(1).unwrap_or(&default));
}
```

If there's an existing way of implementing this pattern that's cleaner, I'll happily close this. I'm also open to naming suggestions (`index_opt`?)
2013-10-12 20:16:19 -07:00
Luqman Aden 95609699e3 Update test for packed structs to also test being placed in statics. 2013-10-12 21:52:14 -04:00
Luqman Aden 4b4b2136e0 librustc: Don't ICE on packed structs in statics. 2013-10-12 21:42:24 -04:00
bors c6e3501da1 auto merge of #9825 : thestinger/rust/rc, r=thestinger 2013-10-12 07:26:20 -07:00
Daniel Micay ea0fc0e32b rc: fix docstring 2013-10-12 10:19:56 -04:00
bors 35a944a488 auto merge of #9815 : thestinger/rust/type, r=huonw
Example:

    void ({ i64, %tydesc*, i8*, i8*, i8 }*, i64*, %"struct.std::fmt::Formatter[#1]"*)*

Before, we would print 20 levels deep due to recursion in the type
definition.
2013-10-11 19:21:16 -07:00
bors 399a42575a auto merge of #9799 : catamorphism/rust/rustpkg-exitcodes, r=catamorphism,metajack
r? @metajack When I started writing the rustpkg tests, task failure didn't set the
exit code properly. But bblum's work from July fixed that. Hooray! I
just didn't know about it till now.

So, now rustpkg uses exit codes in a more conventional way, and some of
the tests are simpler.

The bigger issue will be to make task failure propagate the error message.
Right now, rustpkg does most of the work in separate tasks, which means if
a task fails, rustpkg can't distinguish between different types of failure
(see #3408)
2013-10-11 18:11:17 -07:00
Daniel Micay 7bad416765 have LLVM print type strings for us
Example:

    void ({ i64, %tydesc*, i8*, i8*, i8 }*, i64*, %"struct.std::fmt::Formatter[#1]"*)*

Before, we would print 20 levels deep due to recursion in the type
definition.
2013-10-11 20:26:08 -04:00
Tim Chevalier 7472bae5e7 rustpkg: Set exit codes properly and make tests take advantage of that
When I started writing the rustpkg tests, task failure didn't set the
exit code properly. But bblum's work from July fixed that. Hooray! I
just didn't know about it till now.

So, now rustpkg uses exit codes in a more conventional way, and some of
the tests are simpler.

The bigger issue will be to make task failure propagate the error message.
Right now, rustpkg does most of the work in separate tasks, which means if
a task fails, rustpkg can't distinguish between different types of failure
(see #3408)
2013-10-11 17:15:52 -07:00
bors 80878ff369 auto merge of #9809 : fhahn/rust/remove-old-cratemap-code, r=alexcrichton
This patch removes the code responsible for handling older CrateMap versions (as discussed during #9593). Only the new (safer) layout is supported now.
2013-10-11 12:21:20 -07:00
bors c147ec6d06 auto merge of #9794 : thestinger/rust/rc, r=alexcrichton
I've left out a way to construct from a `Send` type until #9509 is resolved. I am confident that this interface can remain backwards compatible though, assuming we name the `Pointer` trait method `borrow`.

When there is a way to convert from `Send` (`from_send`), a future RAII-based `Mut` type can be used with this to implemented a mutable reference-counted pointer. For now, I've left around the `RcMut` type but it may drastically change or be removed.
2013-10-11 11:11:21 -07:00
Daniel Micay 18be986c99 clean up the `Rc`/`RcMut` types and move to libstd 2013-10-11 13:28:36 -04:00
Florian Hahn f3b1f79716 Remove support for older CrateMap versions 2013-10-11 19:16:20 +02:00
Daniel Micay e192b6d7c5 correct names for `#[no_send]`/`#[no_freeze]` tests 2013-10-11 12:21:19 -04:00
bors ed37b00b06 auto merge of #9803 : alexcrichton/rust/less-pub2, r=brson
This change was waiting for privacy to get sorted out, which should be true now
that #8215 has landed.

Closes #4427
2013-10-11 08:56:19 -07:00
bors 93a08f49fa auto merge of #9802 : ben0x539/rust/mkdtemp-raii, r=alexcrichton
this incidentally stops `make check` from leaving directories in `/tmp` (Closes #9764)
2013-10-11 07:11:23 -07:00
Benjamin Herr 63e9e496f6 extra::tempfile: replace mkdtemp with an RAII wrapper
this incidentally stops `make check` from leaving directories in `/tmp`
2013-10-11 15:55:37 +02:00
Alex Crichton 8b4423b04f De-pub some private runtime components
This change was waiting for privacy to get sorted out, which should be true now
that #8215 has landed.

Closes #4427
2013-10-11 06:49:18 -07:00