Commit Graph

2749 Commits

Author SHA1 Message Date
Brian Anderson
835e963dbd std::rt: Improve the error message when the thread-local ptr is null
Also fix some incorrect comments and variable names.
2013-08-04 15:11:55 -07:00
Erick Tryzelaar
3102b1797e std: replace str::as_c_str with std::c_str 2013-08-04 14:13:17 -07:00
Erick Tryzelaar
bb5bf7c3e0 std: remove str::from_bytes_with_null 2013-08-04 13:32:41 -07:00
Erick Tryzelaar
d5110854f7 std: add test for str::as_c_str 2013-08-04 13:32:41 -07:00
Erick Tryzelaar
cd94e9121b std: cleanup os and str tests 2013-08-04 13:32:41 -07:00
Erick Tryzelaar
dca9ff9a13 std: remove str::NullTerminatedStr 2013-08-04 13:32:41 -07:00
Erick Tryzelaar
fd293dfb0f std: rewrite run::with_{argv,envp,dirp} to copy C strings 2013-08-04 13:32:41 -07:00
Erick Tryzelaar
08b6cb46c6 std: add str.to_c_str() 2013-08-04 13:32:40 -07:00
Erick Tryzelaar
6011f83dec std: minor cleanup 2013-08-04 13:32:40 -07:00
Steven Fackler
147c4fd81b Fixed str::raw::push_byte
It was previously pushing the byte on top of the string's null
terminator. I added a test to make sure it doesn't break in the future.
2013-08-04 16:19:40 -04:00
bors
22f9ce4df6 auto merge of #8243 : stepancheg/rust/ipv, r=brson
multicast functions now take IpAddr (without port), because they dont't
need port.

Uv* types renamed:
* UvIpAddr -> UvSocketAddr
* UvIpv4 -> UvIpv4SocketAddr
* UvIpv6 -> UvIpv6SocketAddr

"Socket address" is a common name for (ip-address, port) pair (e.g. in
sockaddr_in struct).

P. S. Are there any backward compatibility concerns? What is std::rt module, is it a part of public API?
2013-08-04 08:49:55 -07:00
bors
f7c4359a2c auto merge of #8237 : blake2-ppc/rust/faster-utf8, r=brson
Use unchecked vec indexing since the vector bounds are checked by the
loop. Iterators are not easy to use in this case since we skip 1-4 bytes
each lap. This part of the commit speeds up is_utf8 for ASCII input.

Check codepoint ranges by checking the byte ranges manually instead of
computing a full decoding for multibyte encodings. This is easy to read
and corresponds to the UTF-8 syntax in the RFC.

No changes to what we accept. A comment notes that surrogate halves are
accepted.

Before:

	test str::bench::is_utf8_100_ascii ... bench: 165 ns/iter (+/- 3)
	test str::bench::is_utf8_100_multibyte ... bench: 218 ns/iter (+/- 5)

After:
	test str::bench::is_utf8_100_ascii ... bench: 130 ns/iter (+/- 1)
	test str::bench::is_utf8_100_multibyte ... bench: 156 ns/iter (+/- 3)

An improvement upon the previous pull #8133
2013-08-04 07:10:56 -07:00
Brian Anderson
ea84c1fd69 std: Remove gc and stackwalk
These are both obsoleted by the forthcoming new GC.
2013-08-03 23:39:04 -07:00
bors
fbeeeebf47 auto merge of #8264 : thestinger/rust/snapshot, r=Aatch 2013-08-03 23:25:55 -07:00
bors
93432a2c2f auto merge of #8269 : brson/rust/fix-task-cleanup, r=brson
...y/catch

And before collect_failure. These are both running user dtors and need to be handled
in the task try/catch block and before the final task cleanup code.
2013-08-03 21:46:56 -07:00
Daniel Micay
1008945528 remove obsolete foreach keyword
this has been replaced by `for`
2013-08-03 22:48:02 -04:00
Daniel Micay
9f74217d80 register snapshots 2013-08-03 21:09:28 -04:00
Brian Anderson
3c9e393354 std::rt: Run local storage cleanup and the box annihilator inside the try/catch
And before collect_failure. These are both running user dtors and need to be handled
in the task try/catch block and before the final task cleanup code.
2013-08-03 14:43:16 -07:00
bors
18e3db7392 auto merge of #8246 : stepancheg/rust/contains-key, r=thestinger
Map::contains_key can be implemented with Map::find.

Remove several implementations of contains_key.
2013-08-03 13:40:49 -07:00
bors
800dbffa69 auto merge of #8219 : sstewartgallus/rust/fix_dynamic_lib, r=graydon
A test case was also created for this situation to prevent the problem
occuring again.

A similar problem was also fixed for the symbol method.

There was some minor code cleanup.

I am unsatisfied with using /dev/null as an invalid dynamic library. It is not cross platform.
2013-08-03 10:04:58 -07:00
bors
34101d2320 auto merge of #8213 : kballard/rust/fd-limit, r=brson
Revert the workaround 49b72bd and instead bump the fd limit on OS X.
2013-08-03 07:46:53 -07:00
Huon Wilson
c852a5bb68 std: expose the keyed HashMap constructor, for runtime-less use.
The `new` constructor uses the task-local RNG to retrieve seeds for the
two key values, which requires the runtime. Exposing a constructor that
takes the keys directly allows HashMaps to be used in programs that wish
to avoid the runtime.
2013-08-03 21:00:32 +10:00
bors
39fafd655a auto merge of #8204 : kballard/rust/str-into-owned, r=graydon
The method .into_owned() is meant to be used as an optimization when you
need to get a ~str from a Str, but don't want to unnecessarily copy it
if it's already a ~str.

This is meant to ease functions that look like

  fn foo<S: Str>(strs: &[S])

Previously they could work with the strings as slices using .as_slice(),
but producing ~str required copying the string, even if the vector
turned out be a &[~str] already.

I don't have any concrete uses for this yet, since the one conversion I've done to `&[S]` so far (see PR #8203) didn't actually need owned strings. But having this here may make using `Str` more attractive.

It also may be worth adding an `into_managed()` function, but that one is less obviously useful than `into_owned()`.
2013-08-03 02:10:54 -07:00
Huon Wilson
1992765dd3 std: add benchmark for vec.mut_iter. 2013-08-03 03:13:13 -04:00
Huon Wilson
fbb7cd32c3 std: use ptr.offset where possible in the vec iterator.
Closes #8212.
2013-08-03 03:13:11 -04:00
Kevin Ballard
54e685d4fd option: mutate() and mutate_default() should return bool
Fixes #8047.
2013-08-03 03:11:11 -04:00
toddaaro
1f9c392389 fixed the buffer to make it a more reasonable size 2013-08-03 03:09:23 -04:00
toddaaro
eab749a5f3 modified logging function to truncate output and adjusted error output formatting tests to be compatible with both the new and old runtimes 2013-08-03 03:09:04 -04:00
Daniel Micay
b41d04763e make for parse as foreach does
Closes #6997
2013-08-03 03:00:42 -04:00
Daniel Micay
b3ad685f7f replace all remaining for with foreach or do 2013-08-03 03:00:42 -04:00
Patrick Walton
9457ebee55 librustc: Disallow "unsafe" for external functions 2013-08-02 21:57:59 -07:00
Kevin Ballard
2001cc043b Bump fd limit on macos when running rt tests
OS X defaults the ulimit for open files to 256 for programs launched
from the Terminal (GUI apps get a higher default). Unfortunately this is
too low for the rt tests, which deliberately overcommit and create a lot
of threads (which means a lot of schedulers, and each scheduler needs at
least 2 fds).

By calling sysctl() and setrlimit() we can bump the fd limit up to the
maximum allowed (on stock OS X it's 10240).

Fixes #7772.
2013-08-02 21:20:15 -07:00
Kevin Ballard
70d2be0cec Revert "std::rt: Use a constant 4 threads for multithreaded sched tests"
This workaround was less than ideal. A better solution is to raise the
fd limit.

This reverts commit 49b72bdd77.
2013-08-02 21:18:32 -07:00
Stepan Koltsov
cf9e9b21d5 Add default implementation of Map::contains_key function
Map::contains_key can be implemented with Map::find.

Remove several implementations of contains_key.
2013-08-03 05:54:05 +04:00
Stepan Koltsov
9046516444 Rename IpAddr -> SocketAddr, extract IpAddr from SocketAddr
multicast functions now take IpAddr (without port), because they dont't
need port.

Uv* types renamed:
* UvIpAddr -> UvSocketAddr
* UvIpv4 -> UvIpv4SocketAddr
* UvIpv6 -> UvIpv6SocketAddr

"Socket address" is a common name for (ip-address, port) pair (e.g. in
sockaddr_in struct).
2013-08-03 02:06:24 +04:00
Ben Blum
43fecf3556 Add an assert_may_sleep() check on every context switch. 2013-08-02 17:31:45 -04:00
Ben Blum
92f60f4365 Don't fail from kill signals if already unwinding. 2013-08-02 17:31:45 -04:00
Ben Blum
bd35798773 (cleanup) Use more do...finally in extra::sync. 2013-08-02 17:31:45 -04:00
Ben Blum
d30cca46e6 Fix embarrassing bug where 'unkillable' would unwind improperly when it receives a kill signal. 2013-08-02 17:31:45 -04:00
Ben Blum
cde6ad3992 Fix nasty double-free bug where a newrt chan could get killed after rescheduling but before suppressing_finalize. 2013-08-02 17:31:44 -04:00
Ben Blum
be7738bfa1 Add SendDeferred trait and use it to fix #8214. 2013-08-02 17:31:44 -04:00
blake2-ppc
0504d7e57b std: Speed up str::is_utf8
Use unchecked vec indexing since the vector bounds are checked by the
loop. Iterators are not easy to use in this case since we skip 1-4 bytes
each lap. This part of the commit speeds up is_utf8 for ASCII input.

Check codepoint ranges by checking the byte ranges manually instead of
computing a full decoding for multibyte encodings. This is easy to read
and corresponds to the UTF-8 syntax in the RFC.

No changes to what we accept. A comment notes that surrogate halves are
accepted.

Before:

	test str::bench::is_utf8_100_ascii ... bench: 165 ns/iter (+/- 3)
	test str::bench::is_utf8_100_multibyte ... bench: 218 ns/iter (+/- 5)

After:
	test str::bench::is_utf8_100_ascii ... bench: 130 ns/iter (+/- 1)
	test str::bench::is_utf8_100_multibyte ... bench: 156 ns/iter (+/- 3)
2013-08-02 23:20:57 +02:00
bors
f1c1f92d0c auto merge of #8221 : brson/rust/single-threaded, r=graydon
This is the last major runtime feature needed for the transition to the new scheduler.
2013-08-02 10:52:50 -07:00
bors
986df44753 auto merge of #8195 : bblum/rust/task-cleanup, r=brson
In the first commit it is obvious why some of the barriers can be changed to ```Relaxed```, but it is not as obvious for the once I changed in ```kill.rs```. The rationale for those is documented as part of the documenting commit.

Also the last commit is a temporary hack to prevent kill signals from being received in taskgroup cleanup code, which could be fixed in a more principled way once the old runtime is gone.
2013-08-02 07:31:52 -07:00
bors
af97339771 auto merge of #8193 : cmr/rust/linux-errno, r=graydon 2013-08-02 05:49:53 -07:00
bors
dbde42e59e auto merge of #8175 : brson/rust/nodbg, r=graydon
This stuff is ancient, unused, and tied to oldsched
2013-08-02 01:13:51 -07:00
Brian Anderson
c3fa411459 std: Implement SingleThreaded spawn mode for newsched 2013-08-02 00:26:45 -07:00
Steven Stewart-Gallus
2047026fef Fix calling destructor on uninitialized dynamic library crash.
A test case was also created for this situation to prevent the problem
occuring again.

A similar problem was also fixed for the symbol method.

There was some minor code cleanup.
2013-08-01 23:41:33 -07:00
Daniel Micay
234acad404 replace range with an external iterator 2013-08-02 00:51:14 -04:00
Kevin Ballard
aa94dfa625 str: Add method .into_owned(self) -> ~str to Str
The method .into_owned() is meant to be used as an optimization when you
need to get a ~str from a Str, but don't want to unnecessarily copy it
if it's already a ~str.

This is meant to ease functions that look like

  fn foo<S: Str>(strs: &[S])

Previously they could work with the strings as slices using .as_slice(),
but producing ~str required copying the string, even if the vector
turned out be a &[~str] already.
2013-08-01 15:54:58 -07:00
toddaaro
8e98eabce5 modified local to include an implementation for try_unsafe_borrow::<Task> so that the log methods will work 2013-08-01 15:38:48 -07:00
toddaaro
1d82fe5aea fixed incorrect handling of returned scheduler option and restructed scheduler functions slightly 2013-08-01 15:18:29 -07:00
toddaaro
a5f55b3ead minor tweaks - unboxed the coroutine so that it is no longer a ~ pointer inside the task struct, and also added an assert to verify that send is never called inside scheduler context as it is undefined (BROKEN) if that happens 2013-08-01 15:17:19 -07:00
toddaaro
997719c13d Fixed a race where a scheduler configured to only run tasks pinned to it would "bounch" a regular task in and out of the work queue without allowing a different scheduler to run it. 2013-08-01 15:16:01 -07:00
Ben Blum
37f385e44b Have linked failure tests run on the new scheduler instead of requiring RUST_NEWRT to test. 2013-08-01 15:16:01 -07:00
toddaaro
f7eed22387 A major refactoring that changes the way the runtime uses TLS. In the
old design the TLS held the scheduler struct, and the scheduler struct
held the active task. This posed all sorts of weird problems due to
how we wanted to use the contents of TLS. The cleaner approach is to
leave the active task in TLS and have the task hold the scheduler. To
make this work out the scheduler has to run inside a regular task, and
then once that is the case the context switching code is massively
simplified, as instead of three possible paths there is only one. The
logical flow is also easier to follow, as the scheduler struct acts
somewhat like a "token" indicating what is active.

These changes also necessitated changing a large number of runtime
tests, and rewriting most of the runtime testing helpers.

Polish level is "low", as I will very soon start on more scheduler
changes that will require wiping the polish off. That being said there
should be sufficient comments around anything complex to make this
entirely respectable as a standalone commit.
2013-08-01 15:14:00 -07:00
Ben Blum
963d37e821 Temporary workaround to prevent taskgroup cleanup code from failing without an exception handler. 2013-08-01 17:07:31 -04:00
Ben Blum
aeaed77301 Make a forgotten assert in comm be cfg(test)-dependent 2013-08-01 16:52:37 -04:00
Ben Blum
036a6d2f00 Document task killing design and relaxed barrier rationale. 2013-08-01 16:52:37 -04:00
Ben Blum
880246618b Relax some atomic barriers. Loosen up all that tension. There, doesn't that feel good? 2013-08-01 16:52:37 -04:00
bors
82b24559e6 auto merge of #8190 : thestinger/rust/for, r=thestinger 2013-08-01 12:52:29 -07:00
Corey Richardson
2b026718c3 Add a boatload of Linux x86/x86-64/arm errnos 2013-08-01 15:37:01 -04:00
Daniel Micay
94f1a5d6f8 migrate to foreach 2013-08-01 15:28:45 -04:00
blake2-ppc
78cde5b9fb std: Change Times trait to use do instead of for
Change the former repetition::

    for 5.times { }

to::

    do 5.times { }

.times() cannot be broken with `break` or `return` anymore; for those
cases, use a numerical range loop instead.
2013-08-01 16:54:22 +02:00
blake2-ppc
7e210a8129 std: Replace for with do { .. } expr in std::gc
Change all users of old-style for with internal iterators to using
`do`-loops.

The code in stackwalk.rs does not actually implement the
looping protocol (no break on return false).

The code in gc.rs does not use loop breaks, nor does any code using it.

We remove the capacity to break from the loops in std::gc and implement
the walks using `do { .. }` expressions.

No behavior change.
2013-08-01 16:54:22 +02:00
blake2-ppc
e5a64f2add std: Remove the internal iterator methods from trait Set
.intersection(), .union() etc methods in trait std::container::Set use
internal iters. Remove these methods from the trait.

I reported issue #8154 for the reinstatement of iterator-based set algebra
methods to the Set trait.

For bitv and treemap, that lack Iterator implementations of set
operations, preserve them as methods directly on the types themselves.

For HashSet, these methods are replaced by the present .union_iter()
etc.
2013-08-01 16:54:22 +02:00
blake2-ppc
b18bd785ec std: Replace for with do { .. } expr where internal iterators are used 2013-08-01 16:54:22 +02:00
blake2-ppc
6d7a0c8cbc std: Use do blocks instead of for with .iter_bytes() 2013-08-01 16:54:22 +02:00
bors
fe28ac6891 auto merge of #8164 : brson/rust/noportset, r=pcwalton
...haredChan.
2013-08-01 07:46:38 -07:00
bors
18d124b9b5 auto merge of #8158 : bblum/rust/task-cleanup, r=brson
r? @brson
2013-08-01 06:01:39 -07:00
bors
9824d90ffd auto merge of #8155 : stepancheg/rust/unit-zero, r=alexcrichton 2013-08-01 04:19:35 -07:00
Daniel Micay
1fc4db2d08 migrate many for loops to foreach 2013-08-01 05:34:55 -04:00
bors
8ec70ae5de auto merge of #8162 : thestinger/rust/no-copy, r=brson 2013-07-31 22:52:31 -07:00
Daniel Micay
dabd476203 make in and foreach get treated as keywords 2013-08-01 00:21:13 -04:00
Brian Anderson
05eff5f731 extra: Remove dbg module and rt support code
This stuff is ancient, unused, and tied to oldsched
2013-07-31 18:51:57 -07:00
Brian Anderson
ebd14c92f8 std: Remove PortSet. Not supported by new scheduler. Replace uses with SharedChan. 2013-07-31 14:59:49 -07:00
bors
ef87e2cc96 auto merge of #8151 : sanxiyn/rust/atomicrmw, r=cmr
#8039 broke ARM build, and nothing uses these yet.
2013-07-31 14:58:20 -07:00
Ben Blum
bc7cee7bbf Move atomically to unstable::sync, and document what it actually does. Close #7872. 2013-07-31 14:37:22 -04:00
Ben Blum
2e6dc161b6 Give tasks useful names. #2891 2013-07-31 14:37:22 -04:00
Stepan Koltsov
a5912736be Implement Zero for unit 2013-07-31 18:19:26 +04:00
Seo Sanghyeon
428ea7d7ce Revert atomicrmw {max, min, umax, umin} 2013-07-31 19:44:20 +09:00
bors
8b7e241e02 auto merge of #8139 : brson/rust/rm-old-task-apis, r=pcwalton
This removes a bunch of options from the task builder interface that are irrelevant to the new scheduler and were generally unused anyway. It also bumps the stack size of new scheduler tasks so that there's enough room to run rustc and changes the interface to `Thread` to not implicitly join threads on destruction, but instead require an explicit, and mandatory, call to `join`.
2013-07-31 02:10:24 -07:00
bors
8a737b5020 auto merge of #8138 : Dretch/rust/posix-path-push, r=pcwalton
\ is allowed inside file names on linux, for example my system has a file at:

`/run/udev/firmware-missing/intel-ucode\x2f06-3a-09`
2013-07-31 00:22:21 -07:00
bors
5633a5363b auto merge of #8008 : bblum/rust/select, r=brson
Main logic in ```Implement select() for new runtime pipes.```. The guts of the ```PortOne::try_recv()``` implementation are now split up across several functions, ```optimistic_check```, ```block_on```, and ```recv_ready```.

There is one weird FIXME I left open here, in the "implement select" commit -- an assertion I couldn't get to work in the receive path, on an invariant that for some reason doesn't hold with ```SharedPort```. Still investigating this.
2013-07-30 18:58:17 -07:00
Brian Anderson
33df9fc1d0 std: Remove foreign_stack_size spawn option. Irrelevant to future FFI changes 2013-07-30 17:10:23 -07:00
Brian Anderson
8f835d42d7 std: Remove get_task function. Unused 2013-07-30 17:10:21 -07:00
bors
6534b4d4ce auto merge of #8115 : bjz/rust/num-traits, r=brson
Continues #4819
2013-07-30 17:10:19 -07:00
Gareth Smith
de0092c48e No longer treat \ as a path separator on posix systems. 2013-07-30 23:43:39 +01:00
Brian Anderson
81dda9d9cb std: Remove CurrentScheduler spawn mode. Unused 2013-07-30 14:23:46 -07:00
Brian Anderson
08480e4fda std: Remove ExistingScheduler spawn mode. Unused 2013-07-30 14:23:46 -07:00
Brian Anderson
5d2b8d4372 std: Remove PlatformThread spawn mode. Obsolete 2013-07-30 14:23:45 -07:00
Brian Anderson
85fd75ac47 std: Remove ThreadPerTask spawn mode. Unimplemented 2013-07-30 14:23:45 -07:00
Brian Anderson
cb9ee7f5be std: Remove ManualThreads spawn mode 2013-07-30 14:23:45 -07:00
Brian Anderson
0144c83213 std::rt: Change Thread interface to require an explicit join
Makes it more obvious what's going on
2013-07-30 14:23:44 -07:00
Brian Anderson
7265cc6530 std::rt: Use 2MB stacks
Seems to be around the minimum needed by rustc without split stacks
2013-07-30 14:17:56 -07:00
blake2-ppc
8f9014c159 std: Mark the static constants in str.rs as private
static variables are pub by default, which is not reflected in our code
(we need to use priv).
2013-07-30 19:34:54 +02:00
Ben Blum
6b75e92afe UnsafeArc methods return unsafe pointers, so are not themselves unsafe. 2013-07-30 13:19:26 -04:00
Ben Blum
fa8102ab4a Unkillable is not unsafe. Close #7832. 2013-07-30 13:19:25 -04:00
Ben Blum
9675cd311a (cleanup) Fix unimplemented message for kill_all in newsched. 2013-07-30 13:19:25 -04:00
Ben Blum
3f6b4c24ec Add a better-for-testing optimistic_check() for pipes with cfg(test). 2013-07-30 13:19:25 -04:00