The fix is straight-forward, but there are several changes
while fixing the issue.
1) disallow `mut` keyword when making a new struct
In code base, there are following code,
```rust
struct Foo { mut a: int };
let a = Foo { mut a: 1 };
```
This is because of structural record, which is
deprecated corrently (see issue #3089) In structural
record, `mut` keyword should be allowd to control
mutability. But without structural record, we don't
need to allow `mut` keyword while constructing struct.
2) disallow structural records in parser level
This is related to 1). With structural records, there
is an ambiguity between empty block and empty struct
To solve the problem, I change parser to stop parsing
structural records. I think this is not a problem,
because structural records are not compiled already.
Misc. issues
There is an ambiguity between empty struct vs. empty match stmt.
with following code,
```rust
match x{} {}
```
Two interpretation is possible, which is listed blow
```rust
match (x{}) {} // matching with newly-constructed empty struct
(match x{}) {} // matching with empty enum(or struct) x
// and then empty block
```
It seems that there is no such code in rust code base, but
there is one test which uses empty match statement:
https://github.com/mozilla/rust/blob/incoming/src/test/run-pass/issue-3037.rs
All other cases could be distinguished with look-ahead,
but this can't be. One possible solution is wrapping with
parentheses when matching with an uninhabited type.
```rust
enum what { }
fn match_with_empty(x: what) -> ~str {
match (x) { //use parentheses to remove the ambiguity
}
}
```
- Removed space between struct name and parentheses
- Fixed indentation of the rest of the file (missing end)
- Don't print parentheses for structs with no fields
- Added test
I have seen a few people confused on how to explicitly instantiate generic functions, since the syntax differs from C++'s and C#'s, which is probably where most people asking questions about generic functions are coming from. The only use of the `::<T>` syntax in the reference right now is in the section on paths, which is possibly not where someone trying to find out about generic functions is going to start looking. The tutorial doesn't mention it at all, but I think it's all right to make the reference a tiny bit more redundant and avoid stuffing the tutorial with syntax details.
----
The "Generic functions" subsection mentions that generic functions are instantiated based on context, so let's also mention right away (with a link to the #paths section) that an explicit form is available.
This also adds an example that explicitly instantiates a generic function to the function call expression section.
Before:
````
test.rs:3:21: 3:30 error: expected constant integer for repeat count but found variable
test.rs:3 let a = ~[0, ..n]; //~ ERROR expected constant integer for repeat count but found variable
^~~~~~~~~
````
After:
````
test.rs:3:27: 3:28 error: expected constant integer for repeat count but found variable
test.rs:3 let a = ~[0, ..n]; //~ ERROR expected constant integer for repeat count but found variable
^
````
This adds a few words about unit-like struct types (`struct Foo;`) in the sections for `struct` items, structure expressions and structure types (and fixes an adjacent typo or two). The added text is at the same time triply redundant because of how the sections are split and rather brief because I don't think there's that much to say about field-less structs without digressing into `impl`s and generic functions and whatnot, but it's probably better than nothing for a start.
The added arm for the grammar of struct expressions is really awkward. It's just
| expr_path
which is clearly not unambiguously a struct expression, but it didn't feel right not to add anything to the grammar chunk (and I can't tell whether the arm for enum-like structs is somehow unambiguous with regular enum expressions, either). Is this okay?
This changes various type_names to TypeNames and fixes the example for `tcp::accept` that was still using the old `match` syntax and `{|args| ...}` closures.
The `accept` example was fairly outdated. I was just going to stay away from all the IO things until the scheduler revamp lands, but `accept` is probably one of the obvious starting points for networking stuff for a learner, and I don't want to get in the way of anyone's enthusiasm.
Doesn't touch non-comment lines, so I hope I will get away without learning about unit tests. It doesn't seem like the test system is set up to extract tests from doc comments right now.
- Removed space between struct name and parentheses
- Fixed indentation of the rest of the file (missing end)
- Don't print parentheses for structs with no fields
- Added test
The "Generic functions" subsection mentions that generic functions are
instantiated based on context, so let's also mention right away (with a
link to the #paths section) that an explicit form is available.
This also adds an example to the function call expression section that
explicitly instantiates a generic function.
Doesn't touch non-comment lines. This changes various type_names to TypeNames
and fixes the example for `tcp::accept` that was still using the old
`match` syntax and `{|args| ...}` closures.
r?
This fixes the current [random failures](http://buildbot.rust-lang.org/builders/auto-linux/builds/291/steps/test/logs/stdio) on the bots and closes#4436 by removing `unwrap_shared_mutable_state` and the code that depends on it. The result is that ARC-like things will not be unwrappable. This feature is complex and is not used outside of test cases.
Note that there is not consensus to remove it.
(second commit)
The first two commits are the actual fix going into place, and the last is just dealing with the fallout in the rest of the compiler.
The test added in the first two commits had 0 failures before this patch, and if the glob imports were changed to explicit imports then the errors showed up. Due to this behavior, I figured that the desired behavior was for glob imports to not accidentally leak a lot of non-public imports/functions/types into other modules.
There was quite a lot of fallout, and it all passes `make check` locally, but I doubt that it will pass on all platforms because there's probably some guarded off thing I missed.
I plan on making another patch soon which changes the default level of `unused_imports` to `warn` which will hopefully reduce a lot of the `use` noise throughout. In conjunction with #5104, and a few minor fixes, I think that the default level of `warn` is actually really usable.
The basic idea is that we add a new kind of adjustment, AutoAddEnv, that pads
an extern fn into a closure by adding the extra NULL word. Then there are a few
misc changes in trans to get the LLVM types to match up.
Fixes#4808.
