- 04 12月, 2014 11 次提交
-
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
This closes #19168. It's possible that if the downloading of `rustup.sh` is interrupted, bad things could happen, such as running a naked "rm -rf /" instead of "rm -rf /path/to/tmpdir". This wraps rustup.sh's functionality in a function that gets called at the last time that should protect us from these truncation errors.
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
由 Erick Tryzelaar 提交于
-
- 01 12月, 2014 5 次提交
-
-
由 bors 提交于
We only build LLVM for the host architecture, not the target architecture, so this was just a minor typo in the parameters uses. Closes #19383
-
由 bors 提交于
Fixes #19398.
-
由 bors 提交于
This is the style followed by most other error messages.
-
由 bors 提交于
I.e. we should not prematurely build operand constraints at the expansion time. Otherwise `--pretty expanded` diverges: ``` $ cat t.rs #![feature(asm)] pub fn main() { unsafe { asm!("" : : : "hello", "world") }; } $ rustc t.rs --pretty #![feature(asm)] pub fn main() { unsafe { asm!("" : : : "hello" , "world") }; } $ rustc t.rs --pretty expanded #![feature(asm)] #![feature(phase)] #![no_std] #![feature(globs)] #[phase(plugin, link)] extern crate "std" as std; #[prelude_import] use std::prelude::*; pub fn main() { unsafe { asm!("": : : "~{hello},~{world}") }; } ``` (The last code *does* compile, but won't do the expected thing.)
-
- 30 11月, 2014 11 次提交
-
-
由 Corey Farwell 提交于
-
由 bors 提交于
This implementation existed on MaybeOwned, but has been lost in the transition to Cows. Let's put it back. @aturon r?
-
由 Kang Seonghoon 提交于
-
由 P1start 提交于
This breaks code that looks like this: trait Foo { extern "C" unsafe fn foo(); } impl Foo for Bar { extern "C" unsafe fn foo() { ... } } Change such code to look like this: trait Foo { unsafe extern "C" fn foo(); } impl Foo for Bar { unsafe extern "C" fn foo() { ... } } Fixes #19398. [breaking-change]
-
由 Alex Crichton 提交于
We only build LLVM for the host architecture, not the target architecture, so this was just a minor typo in the parameters uses. Closes #19383
-
由 P1start 提交于
-
由 bors 提交于
* Remove public reexports, as a part of #19253 * Rename getopts::Fail_ to getopts::Fail * Didn't see a reason for the suffixed '_' * Removed getopts::FailType * Looked like it was only beings used for tests; refactored the tests to stop requiring it * A few other non-breaking trivial refactoring changes [breaking-change]
-
由 Kang Seonghoon 提交于
Otherwise `--pretty expanded` diverges.
-
由 bors 提交于
Closes #19371
-
由 bors 提交于
Fixes a small omission of `are` in the sentence: `There also a few things you can do with a tuple as a whole, without...` r @steveklabnik?
-
由 Murarth 提交于
-
- 29 11月, 2014 1 次提交
-
-
由 Matej Lach 提交于
-
- 28 11月, 2014 12 次提交
-
-
由 bors 提交于
Fixes #19344
-
由 Steve Klabnik 提交于
Fixes #19344
-
由 bors 提交于
Since #19076 was merged, I believe mipsel + linux maybe need add to the list here, too.
-
由 bors 提交于
This PR lets `rustc` generate debuginfo for variables captured by unboxed closures. Fixes #19356 @nikomatsakis This PR will probably conflict with #19338. If this gets merged before, you should be able to just leave the test case as it is (maybe remove the `#![feature(unboxed_closures)]` directive).
-
由 bors 提交于
The previous code was giving an incorrect result (not x/3). Also, this function does not work with signed integers. It now accepts `u32` instead of `i32`.
-
由 Sean McArthur 提交于
This implementation existed on MaybeOwned, but has been lost in the transition to Cows. Let's put it back.
-
由 bors 提交于
-
由 Liigo Zhuang 提交于
-
由 bors 提交于
- `int_to_str_bytes_common()` doesn't have a return value. - `float_to_str_bytes_common()` has an old-style doc comment.
-
由 bors 提交于
Teach rustdoc that struct variant fields have inherited visibility. Fix #19048
-
由 Corey Farwell 提交于
* Remove public reexports, as a part of #19253 * Rename getopts::Fail_ to getopts::Fail * Didn't see a reason for the suffixed '_' * Removed getopts::FailType * Looked like it was only beings used for tests; refactored the tests to stop requiring it * A few other non-breaking trivial refactoring changes [breaking-change]
-
由 bors 提交于
-