bors [Wed, 22 Jan 2014 19:06:24 +0000 (11:06 -0800)]
auto merge of #10943 : fhahn/rust/issue-7313-replace-c-types, r=alexcrichton
I've started working on a patch for #7313 . So far I tried to replace C types in `src/libstd/unstable/*` and related files.
So far, I have two questions. Is there a convention for passing pointers around in `std` as Rust types? Sometimes pointers are passed around as `*c_char` (which seems to be an `*i8`), `*c_void` or `*u8`, which leads to a lot of casts. E.g: [`exchange_malloc`](https://github.com/fhahn/rust/compare/issue-7313-replace-c-types?expand=1#diff-39f44b8c3f4496abab854b3425ac1617R60) used to return a `*c_char` but the function in turn only calls `malloc_raw` which returns a `*c_void`.
Is there a specific reason for this?
The second question is about `CString` and related functions like `with_c_str`. At the moment these functions use `*c_char*`. Should I replace it with `*u8` or keep it, because it's an wrapper around classical C strings?
bors [Tue, 21 Jan 2014 23:56:16 +0000 (15:56 -0800)]
auto merge of #11129 : SimonSapin/rust/foo-vs-foo_opt, r=alexcrichton
[On 2013-12-06, I wrote to the rust-dev mailing list](https://mail.mozilla.org/pipermail/rust-dev/2013-December/007263.html):
> Subject: Let’s avoid having both foo() and foo_opt()
>
> We have some functions and methods such as [std::str::from_utf8](http://static.rust-lang.org/doc/master/std/str/fn.from_utf8.html) that may succeed and give a result, or fail when the input is invalid.
>
> 1. Sometimes we assume the input is valid and don’t want to deal with the error case. Task failure works nicely.
>
> 2. Sometimes we do want to do something different on invalid input, so returning an `Option<T>` works best.
>
> And so we end up with both `from_utf8` and `from_utf8`. This particular case is worse because we also have `from_utf8_owned` and `from_utf8_owned_opt`, to cover everything.
>
> Multiplying names like this is just not good design. I’d like to reduce this pattern.
>
> Getting behavior 1. when you have 2. is easy: just call `.unwrap()` on the Option. I think we should rename every `foo_opt()` function or method to just `foo`, remove the old `foo()` behavior, and tell people (through documentation) to use `foo().unwrap()` if they want it back?
>
> The downsides are that unwrap is more verbose and gives less helpful error messages on task failure. But I think it’s worth it.
The email discussion has gone around long enough. Let’s discuss a concrete proposal. For the following functions or methods, I removed `foo` (that caused task failure) and renamed `foo_opt` (that returns `Option`) to just `foo`.
Vector methods:
* `get_opt` (rename only, `get` did not exist as it would have been just `[]`)
* `head_opt`
* `last_opt`
* `pop_opt`
* `shift_opt`
* `remove_opt`
`std::path::BytesContainer` method:
* `container_as_str_opt`
`std::str` functions:
* `from_utf8_opt`
* `from_utf8_owned_opt` (also remove the now unused `not_utf8` condition)
Is there something else that should recieve the same treatement?
I did not rename `recv_opt` on channels based on @brson’s [feedback](https://mail.mozilla.org/pipermail/rust-dev/2013-December/007270.html).
bors [Tue, 21 Jan 2014 19:26:13 +0000 (11:26 -0800)]
auto merge of #11700 : bharrisau/rust/thumb, r=alexcrichton
To build for the cortex-M series ARM processors LLC needs to be told to build for the thumb instruction set. There are two ways to do this, either with the triple "thumb\*-\*-\*" or with -march=thumb (which just overrides the triple anyway). I chose the first way.
The following will fail because the local cc doesn't know what to do with -mthumb.
````
rustc test.rs --lib --target thumb-linux-eab
error: linking with `cc` failed: exit code: 1
note: cc: error: unrecognized command line option ‘-mthumb’
````
Changing the linker works as expected.
````
rustc test.rs --lib --target thumb-linux-eabi --linker arm-none-eabi-gcc
````
Ideally I'd have the triple thumb-none-eabi, but adding a new OS looks like much more work (and I'm not familiar enough with what it does to know if it is needed).
bors [Tue, 21 Jan 2014 18:06:18 +0000 (10:06 -0800)]
auto merge of #11665 : alexcrichton/rust/zed-cleanup, r=brson
* Stop using hardcoded numbers that have to all get updated when something changes (inevitable errors and rebase conflicts) as well as removes some unneeded -Z options (obsoleted over time).
* Remove `std::rt::borrowck`
bors [Tue, 21 Jan 2014 04:06:23 +0000 (20:06 -0800)]
auto merge of #11662 : alexcrichton/rust/faster-parens, r=huonw
The included test case would essentially never finish compiling without this
patch. It recursies twice at every ExprParen meaning that the branching factor
is 2^n
The included test case will take so long to parse on the old compiler that it'll
surely never let this crop up again.
bors [Tue, 21 Jan 2014 02:31:38 +0000 (18:31 -0800)]
auto merge of #11486 : Matthias247/rust/doc, r=cmr
I wrote a chapter for the FFI tutorial that describes how to perform callbacks from C code to Rust and gives some hints about what to consider and synchronization.
I just needed that for my own wrapper and thought it would be helpful for others.
Alex Crichton [Sun, 19 Jan 2014 08:00:39 +0000 (00:00 -0800)]
Fix a pathological const checking case
The included test case would essentially never finish compiling without this
patch. It recursies twice at every ExprParen meaning that the branching factor
is 2^n
The included test case will take so long to parse on the old compiler that it'll
surely never let this crop up again.
bors [Mon, 20 Jan 2014 18:16:30 +0000 (10:16 -0800)]
auto merge of #11664 : bjz/rust/identities, r=alexcrichton
`Zero` and `One` have precise definitions in mathematics as the identities of the `Add` and `Mul` operations respectively. As such, types that implement these identities are now also required to implement their respective operator traits. This should reduce their misuse whilst still enabling them to be used in generalized algebraic structures (not just numbers). Existing usages of `#[deriving(Zero)]` in client code could break under these new rules, but this is probably a sign that they should have been using something like `#[deriving(Default)]` in the first place.
For more information regarding the mathematical definitions of the additive and multiplicative identities, see the following Wikipedia articles:
Note that for floating point numbers the laws specified in the doc comments of `Zero::zero` and `One::one` may not always hold. This is true however for many other traits currently implemented by floating point numbers. What traits floating point numbers should and should not implement is an open question that is beyond the scope of this pull request.
The implementation of `std::num::pow` has been made more succinct and no longer requires `Clone`. The coverage of the associated unit test has also been increased to test for more combinations of bases, exponents, and expected results.
bors [Mon, 20 Jan 2014 14:16:29 +0000 (06:16 -0800)]
auto merge of #11661 : huonw/rust/fixed-length-instantiation, r=thestinger
Previously, they were treated like ~[] and &[] (which can have length
0), but fixed length vectors are fixed length, i.e. we know at compile
time if it's possible to have length zero (which is only for [T, .. 0]).
The implementation has been made more succinct and no longer requires Clone. The coverage of the associated unit test has also been increased to check more combinations of bases, exponents, and expected results.
Add operator trait constraints to std::num::{Zero, One} and document their appropriate use
Zero and One have precise definitions in mathematics. Documentation has been added to describe the appropriate uses for these traits and the laws that they should satisfy.
For more information regarding these identities, see the following wikipedia pages:
bors [Mon, 20 Jan 2014 03:46:35 +0000 (19:46 -0800)]
auto merge of #11649 : FlaPer87/rust/pow, r=cmr
There was an old and barely used implementation of pow, which expected
both parameters to be uint and required more traits to be implemented.
Since a new implementation for `pow` landed, I'm proposing to remove
this old impl in favor of the new one.
The benchmark shows that the new implementation is faster than the one being removed:
```
test num::bench::bench_pow_function ..bench: 9429 ns/iter (+/- 2055)
test num::bench::bench_pow_with_uint_function ...bench: 28476 ns/iter (+/- 2202)
```
bors [Mon, 20 Jan 2014 02:21:39 +0000 (18:21 -0800)]
auto merge of #10801 : musitdev/rust/jsondoc2, r=cmr
I update the example of json use to the last update of the json.rs code. I delete the old branch.
From my last request, I remove the example3 because it doesn't compile. I don't understand why and I don't have the time now to investigate.
bors [Mon, 20 Jan 2014 00:56:40 +0000 (16:56 -0800)]
auto merge of #11644 : huonw/rust/less-fatality, r=cmr
This means that compilation continues for longer, and so we can see more
errors per compile. This is mildly more user-friendly because it stops
users having to run rustc n times to see n macro errors: just run it
once to see all of them.
bors [Sun, 19 Jan 2014 13:56:35 +0000 (05:56 -0800)]
auto merge of #11639 : sfackler/rust/macro-crate-path, r=alexcrichton
If the library is in the working directory, its path won't have a "/"
which will cause dlopen to search /usr/lib etc. It turns out that Path
auto-normalizes during joins so Path::new(".").join(path) is actually a
no-op.
Huon Wilson [Sun, 19 Jan 2014 07:29:47 +0000 (18:29 +1100)]
rustc: check instantiability of fixed length vectors properly.
Previously, they were treated like ~[] and &[] (which can have length
0), but fixed length vectors are fixed length, i.e. we know at compile
time if it's possible to have length zero (which is only for [T, .. 0]).
bors [Sun, 19 Jan 2014 07:16:33 +0000 (23:16 -0800)]
auto merge of #11616 : huonw/rust/ast_map, r=pnkfelix
NodeIds are sequential integers starting at zero, so we can achieve some
memory savings by just storing the items all in a line in a vector.
The occupancy for typical crates seems to be 75-80%, so we're already
more efficient than a HashMap (maximum occupancy 75%), not even counting
the extra book-keeping that HashMap does.
bors [Sun, 19 Jan 2014 05:56:34 +0000 (21:56 -0800)]
auto merge of #11615 : adwhit/rust/master, r=cmr
This is my first patch so feedback appreciated!
Bug when initialising `bitv:Bitv::new(int,bool)` when `bool=true`. It created a `Bitv` with underlying representation `!0u` rather than the actual desired bit layout ( e.g. `11111111` instead of `00001111`). This works OK because a size attribute is included which keeps access to legal bounds. However when using `BitvSet::from_bitv(Bitv)`, we then find that `bitvset.contains(i)` can return true when `i` should not in fact be in the set.
```
let bs = BitvSet::from_bitv(Bitv::new(100, true));
assert!(!bs.contains(&127)) //fails
```
The fix is to create the correct representation by treating various cases separately and using a bitshift `(1<<nbits) - 1` to generate correct number of `1`s where necessary.
bors [Sun, 19 Jan 2014 04:31:47 +0000 (20:31 -0800)]
auto merge of #11567 : divtxt/rust/master, r=cmr
I found the boxes diagram in the tutorial misleading about how the enum worked.
The current diagram makes it seem that there is a separate Cons struct when there is only one type of struct for the List type, and Nil is drawn almost as if it's not consuming memory.
I'm aware of the optimization that happens for this enum which takes advantage of the fact that pointer cannot be null but this is an implementation detail and not the only one that applies here. I can add a note below the diagram mentioning this if you like.
Brian Anderson [Sun, 19 Jan 2014 02:23:46 +0000 (18:23 -0800)]
Pass the correct --target flag when type checking pretty-printed code in tests
This makes pretty print tests that have aux crates work correctly on Android.
Without they generate errors ICEs about incorrect node ids. Not sure why.
Huon Wilson [Fri, 17 Jan 2014 12:23:09 +0000 (23:23 +1100)]
syntax: convert ast_map to use a SmallIntMap.
NodeIds are sequential integers starting at zero, so we can achieve some
memory savings by just storing the items all in a line in a vector.
The occupancy for typical crates seems to be 75-80%, so we're already
more efficient than a HashMap (maximum occupancy 75%), not even counting
the extra book-keeping that HashMap does.
bors [Sat, 18 Jan 2014 22:36:41 +0000 (14:36 -0800)]
auto merge of #11620 : alexcrichton/rust/rustc-silent, r=brson
This commit re-works how the monitor() function works and how it both receives
and transmits errors. There are a few cases in which the compiler can abort:
1. A normal compiler error. In this case, the compiler raises a FatalError as
the failure value of the task. If this happens, then the monitor task does
nothing. It ignores all stderr output of the child task and it also
suppresses the failure message of the main task itself. This means that on a
normal compiler error just the error message itself is printed.
2. A normal internal compiler error. These are invoked from sess.span_bug() and
friends. In these cases, they follow the same path (raising a FatalError),
but they will also print an ICE message which has a URL to go report a bug.
3. An actual compiler bug. This happens whenever anything calls fail!() instead
of going through the session itself. In this case, we print out stuff about
RUST_LOG=2 and we by default capture all stderr and print via warn!() so it's
only printed out with the RUST_LOG var set.
bors [Sat, 18 Jan 2014 21:01:47 +0000 (13:01 -0800)]
auto merge of #11607 : alexcrichton/rust/issue-9957, r=cmr
For `use` statements, this means disallowing qualifiers when in functions and
disallowing `priv` outside of functions.
For `extern mod` statements, this means disallowing everything everywhere. It
may have been envisioned for `pub extern mod foo` to be a thing, but it
currently doesn't do anything (resolve doesn't pick it up), so better to err on
the side of forwards-compatibility and forbid it entirely for now.
Flavio Percoco [Sat, 18 Jan 2014 17:10:17 +0000 (18:10 +0100)]
Replace old pow_with_uint with the new pow func
There was an old and barely used implementation of pow, which expected
both parameters to be uint and required more traits to be implemented.
Since a new implementation for `pow` landed, I'm proposing to remove
this old impl in favor of the new one.
The benchmark shows that the new implementation is faster than the one
being removed:
test num::bench::bench_pow_function ..bench: 9429 ns/iter (+/- 2055)
test num::bench::bench_pow_with_uint_function ...bench: 28476 ns/iter (+/- 2202)
Alex Crichton [Fri, 17 Jan 2014 20:12:46 +0000 (12:12 -0800)]
rustc: Clean up error reporting
This commit re-works how the monitor() function works and how it both receives
and transmits errors. There are a few cases in which the compiler can abort:
1. A normal compiler error. In this case, the compiler raises a FatalError as
the failure value of the task. If this happens, then the monitor task does
nothing. It ignores all stderr output of the child task and it also
suppresses the failure message of the main task itself. This means that on a
normal compiler error just the error message itself is printed.
2. A normal internal compiler error. These are invoked from sess.span_bug() and
friends. In these cases, they follow the same path (raising a FatalError),
but they will also print an ICE message which has a URL to go report a bug.
3. An actual compiler bug. This happens whenever anything calls fail!() instead
of going through the session itself. In this case, we print out stuff about
RUST_LOG=2 and we by default capture all stderr and print via warn!() so it's
only printed out with the RUST_LOG var set.
Alex Crichton [Thu, 16 Jan 2014 21:27:27 +0000 (13:27 -0800)]
Forbid unnecessary visibility on view items
For `use` statements, this means disallowing qualifiers when in functions and
disallowing `priv` outside of functions.
For `extern mod` statements, this means disallowing everything everywhere. It
may have been envisioned for `pub extern mod foo` to be a thing, but it
currently doesn't do anything (resolve doesn't pick it up), so better to err on
the side of forwards-compatibility and forbid it entirely for now.