bors [Sat, 4 Jun 2016 17:47:55 +0000 (10:47 -0700)]
Auto merge of #33816 - nikomatsakis:projection-cache-2, r=arielb1
Projection cache and better warnings for #32330
This PR does three things:
- it lays the groundwork for the more precise subtyping rules discussed in #32330, but does not enable them;
- it issues warnings when the result of a leak-check or subtyping check relies on a late-bound region which will late become early-bound when #32330 is fixed;
- it introduces a cache for projection in the inference context.
I'm not 100% happy with the approach taken by the cache here, but it seems like a step in the right direction. It results in big wins on some test cases, but not as big as previous versions -- I think because it is caching the `Vec<Obligation>` (whereas before I just returned the normalized type with an empty vector). However, that change was needed to fix an ICE in @alexcrichton's future-rs module (I haven't fully tracked the cause of that ICE yet). Also, because trans/the collector use a fresh inference context for every call to `fulfill_obligation`, they don't profit nearly as much from this cache as they ought to.
Still, here are the results from the future-rs `retry.rs`:
~~Another example is the example from #31849. For that, I get 34s to run item-bodies without any cache. The version of the cache included here takes 2s to run item-bodies type-checking. An alternative version which doesn't track nested obligations takes 0.2s, but that version ICEs on @alexcrichton's future-rs (and may well be incorrect, I've not fully convinced myself of that). So, a definite win, but I think there's definitely room for further progress.~~
Pushed a modified version which improves performance of the case from #31849:
```
lunch-box. time rustc --stage0 ~/tmp/issue-31849.rs -Z no-trans
real 0m33.539s
user 0m32.932s
sys 0m0.570s
lunch-box. time rustc --stage2 ~/tmp/issue-31849.rs -Z no-trans
real 0m0.195s
user 0m0.154s
sys 0m0.042s
```
Some sort of cache is also needed for unblocking further work on lazy normalization, since that will lean even more heavily on the cache, and will also require cycle detection.
bors [Sat, 4 Jun 2016 08:23:02 +0000 (01:23 -0700)]
Auto merge of #33991 - alexcrichton:rustbuild-more-clean, r=aturon
rustbuild: Clean more on `make clean`
Be sure to not use the old build cache for the bootstrap build system nor the
old caches of the compiler/cargo extractions (in case something went wrong).
bors [Sat, 4 Jun 2016 05:32:15 +0000 (22:32 -0700)]
Auto merge of #33460 - shepmaster:16-bit-pointers, r=Aatch
Support 16-bit pointers as well as i/usize
I'm opening this pull request to get some feedback from the community.
Although Rust doesn't support any platforms with a native 16-bit pointer at the moment, the [AVR-Rust][ar] fork is working towards that goal. Keeping this forked logic up-to-date with the changes in master has been onerous so I'd like to merge these changes so that they get carried along when refactoring happens. I do not believe this should increase the maintenance burden.
This is based on the original work of Dylan McKay (@dylanmckay).
bors [Fri, 3 Jun 2016 07:13:38 +0000 (00:13 -0700)]
Auto merge of #34016 - sanxiyn:travis-docker, r=brson
Use Docker for Travis
The primary motivtion is to use system LLVM from ubuntu.com, instead of llvm.org.
Travis provides two environments: Ubuntu 12.04 LTS aka precise by default, and Ubuntu 14.04 LTS aka trusty if you specify dist: trusty. According to travis-ci/travis-ci#5821, Ubuntu 16.04 LTS aka xenial is unlikely to be available this year, and Travis recommends to use Docker.
LLVM 3.7 binary for 12.04 and 14.04 is not available from ubuntu.com, that's why we used llvm.org. But LLVM 3.7 binary for 16.04 is available from ubuntu.com, and we can use Docker to run on 16.04.
Hopefully this pacifies the 32bit windows. Apparently there’s an ABI out there that not only allows
non-64 bit variadic arguments, but also has differing ABI for them!
Amanieu d'Antras [Wed, 25 May 2016 04:44:28 +0000 (05:44 +0100)]
Fix undefined behavior when re-locking a mutex from the same thread
The only applies to pthread mutexes. We solve this by creating the
mutex with the PTHREAD_MUTEX_NORMAL type, which guarantees that
re-locking from the same thread will deadlock.
Guillaume Gomez [Thu, 2 Jun 2016 11:47:08 +0000 (13:47 +0200)]
Rollup merge of #34019 - kennytm:fix-33958, r=steveklabnik
Restore original meaning of std::fs::read_dir's example changed in #33958
`DirEntry.file_type().is_dir()` will not follow symlinks, but the original example (`fs::metadata(&path).is_dir()`) does. Therefore the change in #33958 introduced a subtle difference that now it won't enter linked folders. To preserve the same behavior, we use `Path::is_dir()` instead, which does follow symlink.
bors [Thu, 2 Jun 2016 04:48:32 +0000 (21:48 -0700)]
Auto merge of #33947 - xosmig:btree_split_off, r=gereeter
Implement split_off for BTreeMap and BTreeSet (RFC 509)
Fixes #19986 and refactors common with append methods.
It splits the tree with O(log n) operations and then calculates sizes by traversing the lower one.
bors [Wed, 1 Jun 2016 17:21:55 +0000 (10:21 -0700)]
Auto merge of #33814 - lambda:rtabort-use-platform-abort, r=alexcrichton
Open code the __fastfail intrinsic for rtabort! on windows
As described https://msdn.microsoft.com/en-us/library/dn774154.aspx
This is a Windows 8+ mechanism for terminating the process quickly,
which degrades to either an access violation or bugcheck in older versions.
I'm not sure this is better the the current mechanism of terminating
with an illegal instruction, but we recently converted unix to
terminate more correctly with SIGABORT, and this *seems* more correct
for windows.
kennytm [Wed, 1 Jun 2016 16:01:53 +0000 (00:01 +0800)]
Restore original meaning of std::fs::read_dir's example changed in #33958.
DirEntry.file_type().is_dir() will not follow symlinks, but the original
example (fs::metadata(&path).is_dir()) does. Therefore the change in
#33958 introduced a subtle difference that now it won't enter linked
folders. To preserve the same behavior, we use Path::is_dir() instead,
which does follow symlink.
bors [Wed, 1 Jun 2016 13:21:53 +0000 (06:21 -0700)]
Auto merge of #33794 - petrochenkov:sanity, r=nrc
Add AST validation pass and move some checks to it
The purpose of this pass is to catch constructions that fit into AST data structures, but not permitted by the language. As an example, `impl`s don't have visibilities, but for convenience and uniformity with other items they are represented with a structure `Item` which has `Visibility` field.
This pass is intended to run after expansion of macros and syntax extensions (and before lowering to HIR), so it can catch erroneous constructions that were generated by them. This pass allows to remove ad hoc semantic checks from the parser, which can be overruled by syntax extensions and occasionally macros.
The checks can be put here if they are simple, local, don't require results of any complex analysis like name resolution or type checking and maybe don't logically fall into other passes. I expect most of errors generated by this pass to be non-fatal and allowing the compilation to proceed.
I intend to move some more checks to this pass later and maybe extend it with new checks, like, for example, identifier validity. Given that syntax extensions are going to be stabilized in the measurable future, it's important that they would not be able to subvert usual language rules.
In this patch I've added two new checks - a check for labels named `'static` and a check for lifetimes and labels named `'_`. The first one gives a hard error, the second one - a future compatibility warning.
Fixes https://github.com/rust-lang/rust/issues/33059 ([breaking-change])
cc https://github.com/rust-lang/rfcs/pull/1177
Rollup merge of #33973 - zackmdavis:stable_features_warning_notes_version_stabilized, r=brson
stable features lint warning mentions version stabilized
To accomplish this, we alter the checks in `rustc::middle::stability` to
use the `StabilityLevel` defined in `syntax::attr` (which includes the
version in which the feature was stabilized) rather than the local
`StabilityLevel` in the same module, and make the
`declared_stable_lang_features` field of
`syntax::feature_gate::Features` hold a Vec of feature-name, span
tuples (in analogy to the `declared_lib_features` field) rather than
just spans.
Rollup merge of #33967 - dsprenkels:enum_pattern_resolve_ice, r=petrochenkov
resolve: record pattern def when `resolve_pattern` returns `Err(true)`
I propose a fix for issue #33293.
In 1a374b8, (pr #33046) fixed the error reporting of a specific case, but the change that was introduced did not make sure that `record_def` was called in all cases, which lead to an ICE in [1].
This change restores the original `else` case, but keeps the changes that were committed in 1a374b8.
Rollup merge of #33921 - jameysharp:patch-1, r=alexcrichton
Inline simple Cursor write calls
Implementing the Write trait for Cursors over slices is so light-weight that under some circumstances multiple writes can be fused into a single instruction. In general I think inlining these functions is a good idea because most of the code can be constant-folded and copy-propagated away.
Let me know if you would prefer this copyright notice to only mention the year of original publication (please make sure that it is really 2011 as stated in the current version of the documentation, and not 2010 like Rust's code) and I'll make the simplification.
bors [Wed, 1 Jun 2016 01:49:48 +0000 (18:49 -0700)]
Auto merge of #33979 - retep998:why-the-long-face, r=alexcrichton
Attempt to diagnose #33844
https://github.com/rust-lang/rust/issues/33844 is a spurious failure that causes builds to fail due to the linker command sometimes failing with error 206, which means that the command is too long. This PR makes rustc print out the linker arguments in that case so the reason for it being so long can be diagnosed and hopefully fixed.
Niko Matsakis [Thu, 26 May 2016 10:11:16 +0000 (06:11 -0400)]
expand `DepNode::TraitSelect` to include type ids
To handle the general case, we include a vector of def-ids, so that we
can account for things like `(Foo, Bar)` which references both `Foo` and
`Bar`. This means it is not Copy, so re-jigger some APIs to use
borrowing more intelligently.
Niko Matsakis [Sat, 21 May 2016 12:16:07 +0000 (08:16 -0400)]
add a higher-ranked match routine
Currently, when projecting out of a higher-ranked where-clause, we
instantiate all higher-ranked regions with lifetime variables. This is
unnecessary since the language rules ought to guarantee (modulo #32330)
that each of those higher-ranked regions is equated with some regions
from the input types. This routine figures out what those regions are
and just uses them. Also, since #32330 is not fully fixed, it detects
when we may have unconstrained variables and indicates that in its
return value.
Niko Matsakis [Sat, 21 May 2016 12:11:54 +0000 (08:11 -0400)]
remove VerifyRegSubReg and cleanup region infer
We used to make region->region edges part of the verify set, but this
change stores them like other edges, as a full-fledged constraint.
Besides making the code somewhat cleaner, this allows them to be more
easily dropped as part of `pop_skolemized`. This change also refactors
the code a bit to remove some intermediate data structures that are no
longer particular useful (e.g., VarValue).
Niko Matsakis [Thu, 21 Apr 2016 09:15:53 +0000 (05:15 -0400)]
warn if leak-check relies on LBRs that will change
When we do a "HR subtype" check, we replace all late-bound regions (LBR)
in the subtype with fresh variables, and skolemize the late-bound
regions in the supertype. If those skolemized regions from the supertype
wind up being super-regions (directly or indirectly) of either
- another skolemized region; or,
- some region that pre-exists the HR subtype check
- e.g., a region variable that is not one of those created
to represent bound regions in the subtype
then the subtype check fails.
What will change when we fix #32330 is that some of the LBR in the
subtype may become early-bound. In that case, they would no longer be in
the "permitted set" of variables that can be related to a skolemized
type.
So the foundation for this warning is to collect variables that we found
to be related to a skolemized type. For each of them, we have a
`BoundRegion` which carries a `Issue32330` flag. We check whether any of
those flags indicate that this variable was created from a lifetime
that will change from late- to early-bound. If so, we issue a warning
indicating that the results of compilation may change.
This is imperfect, since there are other kinds of code that will not
compile once #32330 is fixed. However, it fixes the errors observed in
practice on crater runs.
Niko Matsakis [Thu, 21 Apr 2016 09:10:10 +0000 (05:10 -0400)]
add `Issue32330` warning marker to bound regions
This indicates whether this `BoundRegion` will change from late to early
bound when issue 32330 is fixed. It also indicates the function on
which the lifetime is declared.
Niko Matsakis [Wed, 20 Apr 2016 23:51:56 +0000 (19:51 -0400)]
make HR algorithms account for region subtyping
Currently, we consider region subtyping a failure
if a skolemized lifetime is relatable to any
other lifetime in any way at all. But a more precise
formulation is to say that a skolemized lifetime:
- must not have any *incoming* edges in the region graph
- only has *outgoing* edges to nodes that are `'static`
To enforce the latter requirement, we add edges from `'static -> 'x` for
each lifetime '`x' reachable from a skolemized region.
We now have to add a new `pop_skolemized` routine to do cleanup.
Whereas before if there were *any* edges relating to a skolemized
region, we would return `Err` and hence rollback the transaction, we now
tolerate some edges and return `Ok`. Therefore, the `pop_skolemized`
routine runs and cleans up those edges.
Alex Crichton [Tue, 31 May 2016 20:24:28 +0000 (13:24 -0700)]
rustbuild: Clean more on `make clean`
Be sure to not use the old build cache for the bootstrap build system nor the
old caches of the compiler/cargo extractions (in case something went wrong).
projects / rust.git / log