Commit graph

39232 commits

Author SHA1 Message Date
bors 48aeaba934 Auto merge of #22801 - huonw:crateify-lint, r=kmcallister
This pulls out the implementations of most built-in lints into a
separate crate, to reduce edit-compile-test iteration times with
librustc_lint and increase parallelism. This should enable lints to be
refactored, added and deleted much more easily as it slashes the
edit-compile cycle to get a minimal working compiler to test with (`make
rustc-stage1`) from

    librustc -> librustc_typeck -> ... -> librustc_driver ->
        libcore -> ... -> libstd

to

    librustc_lint -> librustc_driver -> libcore -> ... libstd

which is significantly faster, mainly due to avoiding the librustc build
itself.

The intention would be to move as much as possible of the infrastructure
into the crate too, but the plumbing is deeply intertwined with librustc
itself at the moment. Also, there are lints for which diagnostics are
registered directly in the compiler code, not in their own crate
traversal, and their definitions have to remain in librustc.

This is a [breaking-change] for direct users of the compiler APIs:
callers of `rustc::session::build_session` or
`rustc::session::build_session_` need to manually call
`rustc_lint::register_builtins` on their return value.

This should make #22206 easier.
2015-02-28 06:06:18 +00:00
Huon Wilson 39092530a4 Update docs for rustc_lint crateification. 2015-02-28 15:34:00 +11:00
Huon Wilson 532cd5f85a Separate most of rustc::lint::builtin into a separate crate.
This pulls out the implementations of most built-in lints into a
separate crate, to reduce edit-compile-test iteration times with
librustc_lint and increase parallelism. This should enable lints to be
refactored, added and deleted much more easily as it slashes the
edit-compile cycle to get a minimal working compiler to test with (`make
rustc-stage1`) from

    librustc -> librustc_typeck -> ... -> librustc_driver ->
        libcore -> ... -> libstd

to

    librustc_lint -> librustc_driver -> libcore -> ... libstd

which is significantly faster, mainly due to avoiding the librustc build
itself.

The intention would be to move as much as possible of the infrastructure
into the crate too, but the plumbing is deeply intertwined with librustc
itself at the moment. Also, there are lints for which diagnostics are
registered directly in the compiler code, not in their own crate
traversal, and their definitions have to remain in librustc.

This is a [breaking-change] for direct users of the compiler APIs:
callers of `rustc::session::build_session` or
`rustc::session::build_session_` need to manually call
`rustc_lint::register_builtins` on their return value.

This should make #22206 easier.
2015-02-28 15:33:59 +11:00
bors 1c93934fcf Auto merge of #22669 - dotdash:fast_slice_iter, r=huonw
This adds the assume() calls back that got lost when rebasing #21886.
2015-02-28 03:37:20 +00:00
bors e233987ce1 Auto merge of #22860 - Manishearth:rollup, r=alexcrichton
Passes check-stage1, check-stage2
2015-02-27 23:15:17 +00:00
Alex Crichton bde4c1d6fb Test fixes and rebase conflicts 2015-02-27 15:13:35 -08:00
Manish Goregaokar 158d99d3f1 Rollup merge of #22844 - pnkfelix:refactor-signalling-of-E0223, r=Aatch
Refactor signaling of E0223 to avoid multiple span_errs for same code.

(Doing this cuts out one diagnostic during the build.)
2015-02-27 22:07:05 +05:30
Felix S. Klock II 9549350b87 Refactor signaling of E0223 to avoid multiple span_errs for same code. 2015-02-27 22:07:05 +05:30
Manish Goregaokar 08360892a0 Rollup merge of #22848 - tshepang:redundant-mention, r=brson 2015-02-27 22:07:04 +05:30
Manish Goregaokar 2651823614 Rollup merge of #22837 - eddyb:issue-21721, r=dotdash
Closes #21721.
2015-02-27 22:07:04 +05:30
Manish Goregaokar 204adc9ab8 Rollup merge of #22836 - mahkoh:testflush, r=huonw
See 41f8b1e89b/src/libtest/lib.rs (L783-L788)
2015-02-27 22:07:04 +05:30
Manish Goregaokar cf29aa5300 Rollup merge of #22835 - tshepang:remove-some-warnings, r=jakub- 2015-02-27 22:07:04 +05:30
Manish Goregaokar 8c9fe23ff0 Rollup merge of #22833 - laijs:remove-redundant-else-branch, r=dotdash
The branch \"else { continue }\" is the last code inside a loop body,
it is just useless.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
2015-02-27 22:07:03 +05:30
Manish Goregaokar ce5f1b3216 Rollup merge of #22817 - jmesmon:result-or-type, r=huonw
Changes .or() so that it can return a Result with a different E type
than the one it is called on.

Essentially:

    fn or(self, res: Result<T, E>) -> Result<T, E>

becomes

    fn or<F>(self, res: Result<T, F>) -> Result<T, F>

This brings `or` in line with the existing `and` & `or_else`

This is a
[breaking-change]
Due to some code needing additional type annotations.
2015-02-27 22:07:03 +05:30
Manish Goregaokar ef80ade406 Rollup merge of #22809 - nikomatsakis:issue-22655-infinite-recursion, r=pnkfelix
Change MarkerTrait to be invariant. This is a (small) loss of expressiveness, but is necessary for now to work around #22806. Fixes #22655.

r? @pnkfelix
2015-02-27 22:07:03 +05:30
Manish Goregaokar 91569a95f1 Rollup merge of #22795 - alexcrichton:issue-22617, r=huonw
Keeps the method consistent with `Iterator::any`.

Closes #22617
[breaking-change]
2015-02-27 22:07:02 +05:30
Manish Goregaokar fb92e312d9 Rollup merge of #22788 - vadimcn:fix-precise_time_ns, r=pnkfelix
which starts happening after ~2 hours of machine uptime.
Closes #17845
2015-02-27 22:07:02 +05:30
Manish Goregaokar 08c25f37df Rollup merge of #22748 - jxcl:string-backslash, r=steveklabnik
r? @steveklabnik

Closes #22698

I wasn't sure that this was appropriate for the book, but I've added this to the reference. I also noticed that one of the U+ symbols in the character literals section was missing the graves.
2015-02-27 20:37:39 +05:30
Manish Goregaokar c46fe78c32 Rollup merge of #22732 - jxcl:cargo-new, r=steveklabnik
r? @steveklabnik
2015-02-27 20:37:39 +05:30
Manish Goregaokar 487ee79e3f Rollup merge of #22846 - dhuseby:bitrig-cleanup, r=alexcrichton
This patch contains a couple time fixes to make Rust compile on Bitrig again.  This does not affect OpenBSD.
2015-02-27 20:37:39 +05:30
Manish Goregaokar d9704bdfdb Rollup merge of #22843 - vhbit:no-jemalloc-fix, r=alexcrichton
"body": null,
2015-02-27 20:37:38 +05:30
Manish Goregaokar 19ab0b9a3c Rollup merge of #22823 - kmcallister:fix-21370, r=huonw
Fixes #21370.

`unused-macro-with-follow-violation.rs` was already handled correctly. That test is just for good measure. :)

I have a more involved plan to clean this up, but it ran into difficulties such as #22814.
2015-02-27 20:37:37 +05:30
Manish Goregaokar 6a97fbab4b Rollup merge of #22818 - pnkfelix:fsk-issue-22536, r=nikomatsakis
Ensure we do not zero when \"moving\" types that are Copy.

Uses more precise `type_needs_drop` for deciding about emitting cleanup code.  Added notes about the weaknesses regarding `ty::type_contents` here.

Fix #22536
2015-02-27 20:37:37 +05:30
Manish Goregaokar 5d4e01766b Rollup merge of #22803 - huonw:field-stability, r=alexcrichton
We were recording stability attributes applied to fields in the
compiler, and even annotating it in the libs, but the compiler didn't
actually do the checks to give errors/warnings in user crates.

Details in the commit messages.
2015-02-27 20:37:36 +05:30
bors bd0d8e47e5 Auto merge of #22573 - nwin:impl-debug-rwlock-weak, r=Manishearth
Implements `Debug`  for `RwLock` and `arc::Weak` in the same way it is implemented for `rc::Weak` (basically copy & paste).

The lack of this implementation prevents the automatic implementation of `Debug` for structs containing members of these types.
2015-02-27 10:35:51 +00:00
bors bd27985457 Auto merge of #22853 - FlaPer87:snap, r=alexcrichton
r? @alexcrichton
 
cc @nikomatsakis @pnkfelix
2015-02-27 07:46:15 +00:00
bors dd077d5dec Auto merge of #22857 - alexcrichton:net-flaky, r=alexcrichton
Instead of allocating the same ports for ipv4 and ipv6 tests, instead draw all
ports from the same pool. Some tests connect to just "localhost" on a particular
port which may accidentally be interacting with other tests as the ipv-what-ness
isn't specified with the string "localhost"

Relevant logs:

* [Deadlock of the `net::tcp::tests::listen_localhost` test][mac]
* [Failure of the `fast_rebind` test][win1]
* [Failure of `multiple_connect_interleaved_lazy_schedule_ip4`][win2]

[mac]: https://gist.github.com/alexcrichton/349c7ce7c620c1adb2f2
[win1]: https://gist.github.com/alexcrichton/7e3611faae2e1edaee6f
[win2]: https://gist.github.com/alexcrichton/4f5f87749af3ad0f9851
2015-02-27 05:21:05 +00:00
Alex Crichton c2400bb269 std: Draw from the same port pool during tests
Instead of allocating the same ports for ipv4 and ipv6 tests, instead draw all
ports from the same pool. Some tests connect to just "localhost" on a particular
port which may accidentally be interacting with other tests as the ipv-what-ness
isn't specified with the string "localhost"

Relevant logs:

* [Deadlock of the `net::tcp::tests::listen_localhost` test][mac]
* [Failure of the `fast_rebind` test][win1]
* [Failure of `multiple_connect_interleaved_lazy_schedule_ip4`][win2]

[mac]: https://gist.github.com/alexcrichton/349c7ce7c620c1adb2f2
[win1]: https://gist.github.com/alexcrichton/7e3611faae2e1edaee6f
[win2]: https://gist.github.com/alexcrichton/4f5f87749af3ad0f9851
2015-02-26 19:04:42 -08:00
bors e5cd6534c1 Auto merge of #22765 - sanxiyn:dedup-rustdoc, r=alexcrichton
rustdoc impl item did not include default methods for local crates, but did include them for external crates. This resulted in duplicate methods. Fix so that impl item does not include default methods for external crates.

Fix #22595.
2015-02-27 02:58:15 +00:00
Huon Wilson 060661d2b4 Add some missing stability attributes on struct fields. 2015-02-27 12:52:43 +11:00
Julian Orth c9168cca72 libtest: flush output after every write
Useful for debugging tests that hang forever.
2015-02-27 02:36:40 +01:00
Flavio Percoco 9d0d72345d register snapshot 880fb89 2015-02-27 01:48:49 +01:00
bors b47aebe3fc Auto merge of #22825 - semarie:openbsd-pathbuf, r=alexcrichton
unbreak openbsd/bitrig build after 2d200c9c (the complete build/check on openbsd is in progress)

@dhuseby maybe this one is in your list ?
2015-02-26 23:04:03 +00:00
Tshepang Lekhonkhobe 6b01f7ac05 doc: the last mention of the word 'iterator' is redundant 2015-02-26 23:23:02 +02:00
Dave Huseby 804c071d8b fixing a few bitrig build breakers 2015-02-26 13:03:06 -08:00
Valerii Hiora d0bb57cfc3 Fixed build with jemalloc disabled 2015-02-26 20:27:53 +02:00
bors a5214e4330 Auto merge of #22715 - edwardw:send-sync-audit, r=alexcrichton
In the process, also replaces two raw mutable pointers with `Unique` to
spell out the ownership semantics.

cc #22709
2015-02-26 15:43:38 +00:00
Seo Sanghyeon 695846331b Add a rustdoc test for default methods in external crates 2015-02-27 00:28:57 +09:00
Seo Sanghyeon 5e1d4fffff Add a way to assert the number of occurrences to htmldocck 2015-02-27 00:27:57 +09:00
Eduard Burtescu 704ce1d735 Revert hacks and add test for LLVM aborts due to empty aggregates.
Closes #21721.
2015-02-26 16:44:07 +02:00
Felix S. Klock II 9a6e3b900f Add doc to core::intrinsics pointing out conservativeness of needs_drop. 2015-02-26 15:12:27 +01:00
Felix S. Klock II 4e334d6141 Regression test for #22536. 2015-02-26 15:12:27 +01:00
Felix S. Klock II 5e4867eef1 Use more precise type_needs_drop for deciding about emitting cleanup code.
Fix #22536
2015-02-26 15:12:26 +01:00
Felix S. Klock II 2d0e6caeec Add push_ctxt instrumentation to all functions that might call base::zero_mem.
(Instrumentation work for Issue 22536.)
2015-02-26 15:12:26 +01:00
Lai Jiangshan 9cdb2568ca remove the redundant else branch
The branch "else { continue }" is the last code inside a loop body,
it is just useless.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
2015-02-26 17:38:21 +08:00
bors 3a96d6a981 Auto merge of #22827 - sfackler:should_panic_pt1, r=alexcrichton
This needs to make it into the snapshot before #21824 can land.

r? @alexcrichton
2015-02-26 09:23:36 +00:00
nwin 36ba96ea3c Implement Debug for RwLock, arc::Weak and Mutex 2015-02-26 10:18:13 +01:00
Edward Wang 68490068e0 Send/Sync audit for libcollections
In the process, also replaces a raw mutable pointers with Unique to
spell out the ownership semantics.

cc #22709
2015-02-26 15:43:40 +08:00
bors 41f8b1e89b Auto merge of #22810 - japaric:cow-path, r=alexcrichton
The Path/PathBuf pair already implements the required `Borrow`/`ToOwned` traits and can be used in a `Cow` pointer, so why not?

r? @alexcrichton
2015-02-26 07:01:18 +00:00
Steven Fackler c3b77ab41c Whitelist #[should_panic] 2015-02-25 22:04:00 -08:00