Alex Crichton [Fri, 28 Feb 2014 20:55:30 +0000 (12:55 -0800)]
std: Switch stdout/stderr to buffered by default
Similarly to #12422 which made stdin buffered by default, this commit makes the
output streams also buffered by default. Now that buffered writers will flush
their contents when they are dropped, I don't believe that there's no reason why
the output shouldn't be buffered by default, which is what you want in 90% of
cases.
As with stdin, there are new stdout_raw() and stderr_raw() functions to get
unbuffered streams to stdout/stderr.
Alex Crichton [Fri, 28 Feb 2014 20:46:46 +0000 (12:46 -0800)]
std: Flush when buffered writers are dropped
It's still not entirely clear what should happen if there was an error when
flushing, but I'm deferring that decision to #12628. I believe that it's crucial
for the usefulness of buffered writers to be able to flush on drop. It's just
too easy to forget to flush them in small one-off use cases.
bors [Sat, 1 Mar 2014 09:51:35 +0000 (01:51 -0800)]
auto merge of #12638 : luqmana/rust/op-no-ref, r=alexcrichton
From my comment on #11450:
The reason for the ICE is because for operators `rustc` does a little bit of magic. Notice that while you implement the `Mul` trait for some type `&T` (i.e a reference to some T), you can simply do `Vec2 {..} * 2.0f32`. That is, `2.0f32` is `f32` and not `&f32`. This works because `rustc` will automatically take a reference. So what's happening is that with `foo * T`, the compiler is expecting the `mul` method to take some `&U` and then it can compare to make sure `T == U` (or more specifically that `T` coerces to `U`). But in this case, the argument of the `mul` method is not a reference and hence the "no ref" error.
I don't think we should ICE in this case since we do catch the mismatched trait/impl method and hence provide a better error message that way.
bors [Sat, 1 Mar 2014 07:21:37 +0000 (23:21 -0800)]
auto merge of #12626 : alexcrichton/rust/assert-eq, r=thestinger
Formatting via reflection has been a little questionable for some time now, and
it's a little unfortunate that one of the standard macros will silently use
reflection when you weren't expecting it. This adds small bits of code bloat to
libraries, as well as not always being necessary. In light of this information,
this commit switches assert_eq!() to using {} in the error message instead of
{:?}.
In updating existing code, there were a few error cases that I encountered:
* It's impossible to define Show for [T, ..N]. I think DST will alleviate this
because we can define Show for [T].
* A few types here and there just needed a #[deriving(Show)]
* Type parameters needed a Show bound, I often moved this to `assert!(a == b)`
* `Path` doesn't implement `Show`, so assert_eq!() cannot be used on two paths.
I don't think this is much of a regression though because {:?} on paths looks
awful (it's a byte array).
Concretely speaking, this shaved 10K off a 656K binary. Not a lot, but sometime
significant for smaller binaries.
Alex Crichton [Fri, 28 Feb 2014 09:23:06 +0000 (01:23 -0800)]
std: Change assert_eq!() to use {} instead of {:?}
Formatting via reflection has been a little questionable for some time now, and
it's a little unfortunate that one of the standard macros will silently use
reflection when you weren't expecting it. This adds small bits of code bloat to
libraries, as well as not always being necessary. In light of this information,
this commit switches assert_eq!() to using {} in the error message instead of
{:?}.
In updating existing code, there were a few error cases that I encountered:
* It's impossible to define Show for [T, ..N]. I think DST will alleviate this
because we can define Show for [T].
* A few types here and there just needed a #[deriving(Show)]
* Type parameters needed a Show bound, I often moved this to `assert!(a == b)`
* `Path` doesn't implement `Show`, so assert_eq!() cannot be used on two paths.
I don't think this is much of a regression though because {:?} on paths looks
awful (it's a byte array).
Concretely speaking, this shaved 10K off a 656K binary. Not a lot, but sometime
significant for smaller binaries.
Alex Crichton [Fri, 28 Feb 2014 20:33:49 +0000 (12:33 -0800)]
rustdoc: Capture all output from rustc by default
This helps prevent interleaving of error messages when running rustdoc tests.
This has an interesting bit of shuffling with I/O handles, but other than that
this is just using the APIs laid out in the previous commit.
bors [Fri, 28 Feb 2014 21:26:30 +0000 (13:26 -0800)]
auto merge of #12616 : alexcrichton/rust/size, r=huonw
I've been playing around with code size when linking to libstd recently, and these were some findings I found that really helped code size. I started out by eliminating all I/O implementations from libnative and instead just return an unimplemented error.
In doing so, a `fn main() {}` executable was ~378K before this patch, and about 170K after the patch. These size wins are all pretty minor, but they all seemed pretty reasonable to me. With native I/O not stubbed out, this takes the size of an LTO executable from 675K to 400K.
Alex Crichton [Fri, 28 Feb 2014 05:11:50 +0000 (21:11 -0800)]
std: Flag run_fmt() as #[inline(always)]
This function is a tiny wrapper that LLVM doesn't want to inline, and it ends up
causing more bloat than necessary. The bloat is pretty small, but it's a win of
at least 7k for small executables, and I imagine that the number goes up as
there are more calls to fail!().
Alex Crichton [Fri, 28 Feb 2014 04:37:40 +0000 (20:37 -0800)]
std: Avoid using "{:?}" in format strings
This removes all usage of Poly in format strings from libstd. This doesn't
prevent more future strings from coming in, but it at least removes the ones for
now.
Alex Crichton [Fri, 28 Feb 2014 04:25:13 +0000 (20:25 -0800)]
std: Remove lots of allocations from log settings
Most of these are unnecessary because we're only looking at static strings. This
also moves to Vec in a few places instead of ~[T].
This didn't end up getting much of a code size win (update_log_settings is the
third largest function in the executables I'm looking at), but this seems like a
generally nice improvement regardless.
bors [Fri, 28 Feb 2014 20:06:30 +0000 (12:06 -0800)]
auto merge of #12607 : alexcrichton/rust/io++, r=brson
This lowers the #[allow(missing_doc)] directive into some of the lower modules
which are less mature. Most I/O modules now require comprehensive documentation.
Alex Crichton [Fri, 28 Feb 2014 19:37:04 +0000 (11:37 -0800)]
syntax: Refactor diagnostics to focus on Writers
This commit alters the diagnostic emission machinery to be focused around a
Writer for emitting errors. This allows it to not hard-code emission of errors
to stderr (useful for other applications).
bors [Fri, 28 Feb 2014 18:51:34 +0000 (10:51 -0800)]
auto merge of #12533 : alexcrichton/rust/svh, r=brson
These hashes are used to detect changes to upstream crates and generate errors which mention that crates possibly need recompilation.
More details can be found in the respective commit messages below. This change is also accompanied with a much needed refactoring of some of the crate loading code to focus more on crate ids instead of name/version pairs.
Alex Crichton [Mon, 24 Feb 2014 08:31:08 +0000 (00:31 -0800)]
std: Improve some I/O documentation
This lowers the #[allow(missing_doc)] directive into some of the lower modules
which are less mature. Most I/O modules now require comprehensive documentation.
Alex Crichton [Fri, 28 Feb 2014 01:07:27 +0000 (17:07 -0800)]
syntax: Expand format!() deterministically
Previously, format!("{a}{b}", a=foo(), b=bar()) has foo() and bar() run in a
nondeterminisc order. This is clearly a non-desirable property, so this commit
uses iteration over a list instead of iteration over a hash map to provide
deterministic code generation of these format arguments.
Alex Crichton [Tue, 25 Feb 2014 03:45:20 +0000 (19:45 -0800)]
rustc: Add the concept of a Strict Version Hash
This new SVH is used to uniquely identify all crates as a snapshot in time of
their ABI/API/publicly reachable state. This current calculation is just a hash
of the entire crate's AST. This is obviously incorrect, but it is currently the
reality for today.
This change threads through the new Svh structure which originates from crate
dependencies. The concept of crate id hash is preserved to provide efficient
matching on filenames for crate loading. The inspected hash once crate metadata
is opened has been changed to use the new Svh.
The goal of this hash is to identify when upstream crates have changed but
downstream crates have not been recompiled. This will prevent the def-id drift
problem where upstream crates were recompiled, thereby changing their metadata,
but downstream crates were not recompiled.
In the future this hash can be expanded to exclude contents of the AST like doc
comments, but limitations in the compiler prevent this change from being made at
this time.
Alex Crichton [Tue, 25 Feb 2014 02:13:51 +0000 (18:13 -0800)]
rustc: Simplify crate loading constraints
The previous code passed around a {name,version} pair everywhere, but this is
better expressed as a CrateId. This patch changes these paths to store and pass
around crate ids instead of these pairs of name/version. This also prepares the
code to change the type of hash that is stored in crates.
Alex Crichton [Fri, 28 Feb 2014 03:53:03 +0000 (19:53 -0800)]
std: Add cfg(test) to UnsafeArc assertions
This is a ubiquitous type in concurrent code, and the assertions are causing
significant code bloat for simple operations such as reading the pointer
(injecting a failure point, etc).
I am testing executable sizes with no I/O implementations (everything stubbed
out to return nothing), and this took the size of a libnative executable from
328K to 207K (37% reduction in size), so I think that this is one assertion
that's well worth configuring off for now.
Huon Wilson [Thu, 27 Feb 2014 07:48:21 +0000 (18:48 +1100)]
Publicise types/add #[allow(visible_private_types)] to a variety of places.
There's a lot of these types in the compiler libraries, and a few of the
older or private stdlib ones. Some types are obviously meant to be
public, others not so much.
bors [Fri, 28 Feb 2014 08:26:34 +0000 (00:26 -0800)]
auto merge of #12544 : erickt/rust/hash, r=acrichto
This PR allows `HashMap`s to work with custom hashers. Also with this patch are:
* a couple generic implementations of `Hash` for a variety of types.
* added `Default`, `Clone` impls to the hashers.
* added a `HashMap::with_hasher()` constructor.
bors [Fri, 28 Feb 2014 07:01:55 +0000 (23:01 -0800)]
auto merge of #12614 : alexcrichton/rust/rollup, r=alexcrichton
Closes #12546 (Add new target 'make dist-osx' to create a .pkg installer for OS X) r=brson
Closes #12575 (rustc: Move local native libs back in link-args) r=brson
Closes #12587 (Provide a more helpful error for tests that fail due to noexec) r=brson
Closes #12589 (rustc: Remove codemap and reachable from metadata encoder) r=alexcrichton
Closes #12591 (Fix syntax::ext::deriving{,::*} docs formatting.) r=huonw
Closes #12592 (Miscellaneous Vim improvements) r=alexcrichton
Closes #12596 (path: Implement windows::make_non_verbatim()) r=alexcrichton
Closes #12598 (Improve the ctags function regular expression) r=alexcrichton
Closes #12599 (Tutorial improvement (new variant of PR #12472).) r=pnkfelix
Closes #12603 (std: Export the select! macro) r=pcwalton
Closes #12605 (Fix typo in doc of Binary trait in std::fmt) r=alexcrichton
Closes #12613 (Fix bytepos_to_file_charpos) r=brson
bors [Fri, 28 Feb 2014 05:46:53 +0000 (21:46 -0800)]
auto merge of #12348 : brunoabinader/rust/libcollections-list-refactory, r=alexcrichton
This PR includes:
- Create an iterator for ```List<T>``` called ```Items<T>```;
- Move all list operations inside ```List<T>``` impl;
- Removed functions that are already provided by ```Iterator``` trait;
- Refactor on ```len()``` and ```is_empty``` using ```Container``` trait;
- Bunch of minor fixes;
A replacement for using @ is intended, but still in discussion.
Nick Cameron [Thu, 27 Feb 2014 23:53:36 +0000 (12:53 +1300)]
Fix bytepos_to_file_charpos.
Make bytepos_to_charpos relative to the start of the filemap rather than its previous behaviour which was to be realtive to the start of the codemap, but ignoring multi-byte chars in earlier filemaps. Rename to bytepos_to_file_charpos. Add tests for multi-byte chars.
Alex Crichton [Mon, 24 Feb 2014 05:30:18 +0000 (21:30 -0800)]
std: Export the select! macro
Mark it as #[experimental] for now. In theory this attribute will be read in the
future. I believe that the implementation is solid enough for general use,
although I would not be surprised if there were bugs in it still. I think that
it's at the point now where public usage of it will start to uncover hopefully
the last few remaining bugs.
Kevin Ballard [Thu, 27 Feb 2014 08:19:27 +0000 (00:19 -0800)]
path: clean up some lint warnings and an obsolete comment
Get rid of the unnecessary parenthesies that crept into some macros.
Remove a FIXME that was already fixed.
Fix a comment that wasn't rendering correctly in rustdoc.
Chris Morgan [Thu, 27 Feb 2014 05:54:54 +0000 (16:54 +1100)]
Fix Vim section movements for standard Rust style.
(Expressed another way: make `[[` et al. work with the curly brace at
the end of a line as is standard Rust style, not just at the start is it
is by default in Vim, from K&R style.)
This came out of #11492, where a simpler but less effective technique
was initially proposed; some discussion of the techniques, ways and
means can be found there.
There are still a few caveats:
- Operator-pending mode behaves differently to the standard behaviour:
if inside curly braces, it should delete up to and including the
closing of the outermost curly brace (that doesn't seem to me
consistent with documented behaviour, but it's what it does). Actual
behaviour (the more logical and consistent, in my opinion): up to the
start of the next outermost curly brace.
- With folding enabled (`set fdm=syntax`), `[[` and `]]` do not behave
as they should: the default behaviour treats an entire closed fold as
one line for these purposes while this code does not (I explicitly
`set nofoldenable` in the function—the side-effects are worse with
folds enabled), leading to unexpected behaviour, the worst of which is
`[[` and/or `]]` not working in visual mode on a closed fold (visual
mode keeps it at the extreme end of the region line of the folded
region, so it's always going back to the opening line of that fold and
immediately being shoved back to the end by visual mode).
- `[[` and `]]` are operating inside comments, whereas the standard
behaviour skips comments.
- The viewport position is sometimes changed when it should not be
necessary.
Chris Morgan [Thu, 27 Feb 2014 05:45:48 +0000 (16:45 +1100)]
Highlight the `mod` in `extern mod x;` as Error.
Just like the bare keyword `crate` is highlighted as Error (a little
dubious, actually, given macros), `mod` is invalid after `extern`: it's
obsolete syntax.
Chris Morgan [Thu, 27 Feb 2014 05:30:28 +0000 (16:30 +1100)]
Fix syntax::ext::deriving{,::*} docs formatting.
The most significant fix is for `syntax::ext::deriving::encodable`,
where one of the blocks of code, auspiciously containing `<S>` (recall
that Markdown allows arbitrary HTML to be contained inside it), was not
formatted as a code block, with a fun but messy effect.
Felix Crux [Thu, 27 Feb 2014 01:27:30 +0000 (20:27 -0500)]
Provide a more helpful error for tests that fail due to noexec
The rustdoc tests create and execute a file in a temporary directory. By
default on UNIX-like platforms this is in `/tmp`, which some users mount
with the `noexec` option. In those cases, the tests fail in a mysterious
way. This change adds a note that suggests what the problem might be, if
the error looks like it could have been caused by the `noexec` setup.
Alex Crichton [Wed, 26 Feb 2014 16:52:08 +0000 (08:52 -0800)]
rustc: Move local native libs back in link-args
With linkers on unix systems, libraries on the right of the command line are
used to resolve symbols in those on the left of the command line. This means
that arguments must have a right-to-left dependency chain (things on the left
depend on things on the right).
This is currently done by ordering the linker arguments as
1. Local object
2. Local native libraries
3. Upstream rust libraries
4. Upstream native libraries
This commit swaps the order of 2 and 3 so upstream rust libraries have access to
local native libraries. It has been seen that some upstream crates don't specify
the library that they link to because the name varies per platform (e.g.
lua/glfw/etc).
This commit enables building these libraries by allowing the upstream rust crate
to have access to local native libraries. I believe that the failure mode for
this scheme is when an upstream rust crate depends on a symbol in an upstream
library which is then redefined in a local library. This failure mode is
incredibly uncommon, and the failure mode also varies per platform (OSX behaves
differently), so I believe that a change like this is fine to make.
bors [Fri, 28 Feb 2014 02:51:53 +0000 (18:51 -0800)]
auto merge of #11979 : FlaPer87/rust/static, r=nikomatsakis
This pull request partially addresses the 2 issues listed before. As part of the work required for this PR, `NonCopyable` was completely removed.
This PR also replaces the content of `type_is_pod` with `TypeContents::is_pod`, although `type_is_content` is currently not being used anywhere. I kept it for consistency with the other functions that exist in this module.
Taken from [this](https://github.com/mozilla/rust/pull/11979#issuecomment-35768249) comment.
I expect some code that, at a high-level, works like this:
- For each *mutable* static item, check that the **type**:
- cannot own any value whose type has a dtor
- cannot own any values whose type is an owned pointer
- For each *immutable* static item, check that the **value**:
- does not contain any ~ or box expressions (including ~[1, 2, 3] sort of things, for now)
- does not contain a struct literal or call to an enum variant / struct constructor where
- the type of the struct/enum is freeze
- the type of the struct/enum has a dtor
bors [Thu, 27 Feb 2014 22:56:56 +0000 (14:56 -0800)]
auto merge of #12584 : alexcrichton/rust/windows-files, r=brson
These commits fix handling of binary files on windows by using the raw `CreateFile` apis directly, also splitting out the windows/unix implementations to their own files because everything was configured between the two platforms.
With this fix in place, this also switches `rustc` to using libnative instead of libgreen. I have confirmed that this PR passes through try on all bots.
Alex Crichton [Wed, 26 Feb 2014 21:03:40 +0000 (13:03 -0800)]
rustc: Use libnative for the compiler
The compiler itself doesn't necessarily need any features of green threading
such as spawning tasks and lots of I/O, so libnative is slightly more
appropriate for rustc to use itself.
This should also help the rusti bot which is currently incompatible with libuv.
Alex Crichton [Wed, 26 Feb 2014 20:57:00 +0000 (12:57 -0800)]
native: Improve windows file handling
This commit splits the file implementation into file_unix and file_win32. The
two implementations have diverged to the point that they share almost 0 code at
this point, so it's easier to maintain as separate files.
The other major change accompanied with this commit is that file::open is no
longer based on libc's open function on windows, but rather windows's CreateFile
function. This fixes dealing with binary files on windows (test added in
previous commit).
This also changes the read/write functions to use ReadFile and WriteFile instead
of libc's read/write.
Alex Crichton [Wed, 26 Feb 2014 20:55:23 +0000 (12:55 -0800)]
std: Small cleanup and test improvement
This weeds out a bunch of warnings building stdtest on windows, and it also adds
a check! macro to the io::fs tests to help diagnose errors that are cropping up
on windows platforms as well.
Flavio Percoco [Wed, 26 Feb 2014 18:22:41 +0000 (19:22 +0100)]
Forbid certain types for static items
- For each *mutable* static item, check that the **type**:
- cannot own any value whose type has a dtor
- cannot own any values whose type is an owned pointer
- For each *immutable* static item, check that the **value**:
- does not contain any ~ or box expressions
(including ~[1, 2, 3] sort of things)
- does not contain a struct literal or call to an enum
variant / struct constructor where
- the type of the struct/enum has a dtor
bors [Thu, 27 Feb 2014 04:01:42 +0000 (20:01 -0800)]
auto merge of #12486 : MicahChalmer/rust/emacs-fixes-round-3, r=brson
I've added details in the description of each comment as to what it does, which I won't redundantly repeat here in the PR. They all relate to indentation in the emacs rust-mode.
What I will note here is that this closes #8787. It addresses the last remaining case (not in the original issue description but in a comment), of indenting `match` statements. With the changes here, I believe every problem described in the issue description or comments of #8787 is addressed.
bors [Thu, 27 Feb 2014 01:26:37 +0000 (17:26 -0800)]
auto merge of #12586 : chris-morgan/rust/fix-pretty-print-slash-star-star-star-crash, r=alexcrichton
The pretty printer was treating block comments with more than two
asterisks after the first slash (e.g. `/***`) as doc comments (which are
attributes), whereas in actual fact they are just regular comments.
Chris Morgan [Thu, 27 Feb 2014 01:16:18 +0000 (12:16 +1100)]
Fix a pretty printer crash on `/***`.
The pretty printer was treating block comments with more than two
asterisks after the first slash (e.g. `/***`) as doc comments (which are
attributes), whereas in actual fact they are just regular comments.