auto merge of #13443 : alexcrichton/rust/rollup, r=alexcrichton
Closes #13441 (debuginfo: Fixes and improvements for #12840, #12886, and #13213)
Closes #13433 (Remove references to @Trait from a compiler error message)
Closes #13430 (Fix outdated lint warning about inner attribute)
Closes #13425 (Remove a pile of (mainly) internal `~[]` uses)
Closes #13419 (Stop using transmute_mut in RefCell)
Closes #13417 (Remove an unnecessary file `src/libnative/io/p`.)
Closes #13409 (Closing assorted resolve bugs)
Closes #13406 (Generalized the pretty-print entry points to support `-o <file>`.)
Closes #13403 (test: Add a test for #7663)
Closes #13402 (rustdoc: Prune the paths that do not appear in the index.)
Closes #13396 (rustc: Remove absolute rpaths)
Closes #13371 (Rename ast::Purity and ast::Impure Function. Closes #7287)
Closes #13350 (collections: replace all ~[T] with Vec<T>.)
Alex Crichton [Tue, 8 Apr 2014 22:10:41 +0000 (15:10 -0700)]
rustc: Don't allow priv use to shadow pub use
Previously, a private use statement would shadow a public use statement, all of
a sudden publicly exporting the privately used item. The correct behavior here
is to only shadow the use for the module in question, but for now it just
reverts the entire name to private so the pub use doesn't have much effect.
The behavior isn't exactly what we want, but this no longer has backwards
compatibility hazards.
Alex Crichton [Tue, 8 Apr 2014 22:03:29 +0000 (15:03 -0700)]
rustc: Don't succeed on shadowed nonexistent import
Previously resolve was checking the "import resolution" for whether an import
had succeeded or not, but this was the same structure filled in by a previous
import if a name is shadowed. Instead, this alters resolve to consult the local
resolve state (as opposed to the shared one) to test whether an import succeeded
or not.
Alex Crichton [Tue, 8 Apr 2014 21:31:25 +0000 (14:31 -0700)]
rustc: Disallow importing through use statements
Resolve is currently erroneously allowing imports through private `use`
statements in some circumstances, even across module boundaries. For example,
this code compiles successfully today:
use std::c_str;
mod test {
use c_str::CString;
}
This should not be allowed because it was explicitly decided that private `use`
statements are purely bringing local names into scope, they are not
participating further in name resolution.
As a consequence of this patch, this code, while valid today, is now invalid:
mod test {
use std::c_str;
unsafe fn foo() {
::test::c_str::CString::new(0 as *u8, false);
}
}
While plausibly acceptable, I found it to be more consistent if private imports
were only considered candidates to resolve the first component in a path, and no
others.
Alex Crichton [Tue, 8 Apr 2014 17:15:46 +0000 (10:15 -0700)]
rustc: Use realpath() for sysroot/rpath
When calculating the sysroot, it's more accurate to use realpath() rather than
just one readlink() to account for any intermediate symlinks that the rustc
binary resolves itself to.
For rpath, realpath() is necessary because the rpath must dictate a relative
rpath from the destination back to the originally linked library, which works
more robustly if there are no symlinks involved.
Concretely, any binary generated on OSX into $TMPDIR requires an absolute rpath
because the temporary directory is behind a symlink with one layer of
indirection. This symlink causes all relative rpaths to fail to resolve.
Alex Crichton [Mon, 7 Apr 2014 22:40:58 +0000 (15:40 -0700)]
rustc: Remove absolute rpaths
Concerns have been raised about using absolute rpaths in #11746, and this is the
first step towards not relying on rpaths at all. The only current use case for
an absolute rpath is when a non-installed rust builds an executable that then
moves from is built location. The relative rpath back to libstd and absolute
rpath to the installation directory still remain (CFG_PREFIX).
Alex Crichton [Tue, 8 Apr 2014 18:38:18 +0000 (11:38 -0700)]
test: Add a test for #7663
I think that the test case from this issue has become out of date with resolve
changes in the past 9 months, and it's not entirely clear to me what the
original bug was.
Regardless, it seems like tricky resolve behavior, so tests were added to make
sure things resolved correctly and warnings were correctly reported.
debuginfo: Improve source code position assignment for inlined functions.
This commit makes sure that code inlined from other functions isn't assigned the source position of the call site, since this leads to undesired behavior when setting line breakpoints (issue #12886)
auto merge of #13413 : alexcrichton/rust/once-fn-move, r=brson
This fixes the categorization of the upvars of procs (represented internally
as once fns) to consider usage to require a loan. In doing so, upvars are no
longer allowed to be moved out of repeatedly in loops and such.
Alex Crichton [Tue, 8 Apr 2014 23:59:18 +0000 (16:59 -0700)]
rustc: Prevent repeated moves out of proc upvars
This fixes the categorization of the upvars of procs (represented internally
as once fns) to consider usage to require a loan. In doing so, upvars are no
longer allowed to be moved out of repeatedly in loops and such.
Alex Crichton [Mon, 7 Apr 2014 08:11:31 +0000 (01:11 -0700)]
std: User a smaller stdin buffer on windows
Apparently windows doesn't like reading from stdin with a large buffer size, and
it also apparently is ok with a smaller buffer size. This changes the reader
returned by stdin() to return an 8k buffered reader for stdin rather than a 64k
buffered reader.
Apparently libuv has run into this before, taking a peek at their code, with a
specific comment in their console code saying that "ReadConsole can't handle big
buffers", which I presume is related to invoking ReadFile as if it were a file
descriptor.
Boris Egorov [Mon, 7 Apr 2014 12:01:10 +0000 (19:01 +0700)]
Improve searching for XXX in tidy script (#3303)
Few places where previous version of tidy script cannot find XXX:
* inside one-line comment preceding by a few spaces;
* inside multiline comments (now it finds it if multiline comment starts
on the same line with XXX).
Change occurences of XXX found by new tidy script.
Alex Crichton [Mon, 7 Apr 2014 20:13:21 +0000 (13:13 -0700)]
rustc: Don't read both rlib and dylib metadata
This is an optimization which is quite impactful for compiling small crates.
Reading libstd's metadata takes about 50ms, and a hello world before this change
took about 100ms (this change halves that time).
Recent changes made it such that this optimization wasn't performed, but I think
it's a better idea do to this for now. See #10786 for tracking this issue.
Alex Crichton [Mon, 7 Apr 2014 19:16:43 +0000 (12:16 -0700)]
rustc: Never register syntax crates in CStore
When linking, all crates in the local CStore are used to link the final product.
With #[phase(syntax)], crates want to be omitted from this linkage phase, and
this was achieved by dumping the entire CStore after loading crates. This causes
crates like the standard library to get loaded twice. This loading process is a
fairly expensive operation when dealing with decompressing metadata.
This commit alters the loading process to never register syntax crates in
CStore. Instead, only phase(link) crates ever make their way into the map of
crates. The CrateLoader trait was altered to return everything in one method
instead of having separate methods for finding information.
Alex Crichton [Mon, 7 Apr 2014 19:14:33 +0000 (12:14 -0700)]
rustc: Use CStore, not a separate crate cache
This separate crate cache is one factor which is causing libstd to be loaded
twice during normal compilation. The crates loaded for syntax extensions have a
separate cache than the crates loaded for linking, so all crates are loaded once
per #[phase] they're tagged with.
This removes the cache and instead uses the CStore structure itself as the cache
for loaded crates. This should allow crates loaded during the syntax phase to be
shared with the crates loaded during the link phase.
auto merge of #13393 : alexcrichton/rust/hopefully-fix-bsd, r=sfackler
This appears to be causing the BSD bots to lock up when looking at the core
dumps I've managed to get. Dropping the `FileDesc` structure triggers the `Arc`
it's contained in to get cleaned up, invoking free(). This instead just closes
the file descriptor (the arc itself is never cleaned up).
I'm still not entirely sure why this is a problem because the pthreads runtime
should register hooks for fork() to prevent this sort of deadlock, but perhaps
that's only done on linux?
Alex Crichton [Mon, 7 Apr 2014 22:52:51 +0000 (15:52 -0700)]
native: Try hard to not malloc on a forked child
This appears to be causing the BSD bots to lock up when looking at the core
dumps I've managed to get. Dropping the `FileDesc` structure triggers the `Arc`
it's contained in to get cleaned up, invoking free(). This instead just closes
the file descriptor (the arc itself is never cleaned up).
I'm still not entirely sure why this is a problem because the pthreads runtime
should register hooks for fork() to prevent this sort of deadlock, but perhaps
that's only done on linux?
auto merge of #13358 : tbu-/rust/pr_doc_equivrel, r=cmr
Add requirements of TotalEq and TotalOrd
Clarify that TotalEq needs an underlying equivalence relation and that TotalOrd
needs a total ordering and specifically named the required (and sufficient)
attributes.
auto merge of #13356 : alexcrichton/rust/ignore-flaky, r=huonw
This test relies on the parent to be descheduled before the child sends its
data. This has proved to be unreliable on libnative on the bots. It's a fairly
trivial test regardless, so ignoring it for now won't lose much.
auto merge of #13165 : sfackler/rust/io-vec, r=alexcrichton
`Reader`, `Writer`, `MemReader`, `MemWriter`, and `MultiWriter` now work with `Vec<u8>` instead of `~[u8]`. This does introduce some extra copies since `from_utf8_owned` isn't usable anymore, but I think that can't be helped until `~str`'s representation changes.
auto merge of #13346 : ben0x539/rust/priv-field-in, r=alexcrichton
In the error message for when a private field is used, include the name of the struct, or if it's a struct-like enum variant, the names of the variant and the enum.
auto merge of #13340 : FlaPer87/rust/code-model, r=cmr
Rust currently defaults to `RelocPIC` regardless. This patch adds a new
codegen option that allows choosing different relocation-model. The
available models are:
Clarify that TotalEq needs an underlying equivalence relation and that TotalOrd
needs a total ordering and specifically named the required (and sufficient)
attributes.
Rust currently defaults to `RelocPIC` regardless. This patch adds a new
codegen option that allows choosing different relocation-model. The
available models are:
auto merge of #13268 : alexcrichton/rust/parse-closure, r=cmr
In summary these are some example transitions this change makes:
'a || => ||: 'a
proc:Send() => proc():Send
The intended syntax for closures is to put the lifetime bound not at the front
but rather in the list of bounds. Currently there is no official support in the
AST for bounds that are not 'static, so this case is currently specially handled
in the parser to desugar to what the AST is expecting. Additionally, this moves
the bounds on procedures to the correct position, which is after the argument
list.
The current grammar for closures and procedures is:
This does not currently handle the << ambiguity in `Option<<'a>||>`, I am
deferring that to a later patch. Additionally, this removes the support for the
obsolete syntaxes of ~fn and &fn.
Alex Crichton [Sun, 6 Apr 2014 07:41:25 +0000 (00:41 -0700)]
std: Ignore a flaky std::comm test
This test relies on the parent to be descheduled before the child sends its
data. This has proved to be unreliable on libnative on the bots. It's a fairly
trivial test regardless, so ignoring it for now won't lose much.
Alex Crichton [Wed, 2 Apr 2014 16:47:11 +0000 (09:47 -0700)]
syntax: Tweak parsing lifetime bounds on closures
In summary these are some example transitions this change makes:
'a || => ||: 'a
proc:Send() => proc():Send
The intended syntax for closures is to put the lifetime bound not at the front
but rather in the list of bounds. Currently there is no official support in the
AST for bounds that are not 'static, so this case is currently specially handled
in the parser to desugar to what the AST is expecting. Additionally, this moves
the bounds on procedures to the correct position, which is after the argument
list.
The current grammar for closures and procedures is:
This does not currently handle the << ambiguity in `Option<<'a>||>`, I am
deferring that to a later patch. Additionally, this removes the support for the
obsolete syntaxes of ~fn and &fn.