Commit graph

46 commits

Author SHA1 Message Date
Niko Matsakis
3f490ca4bd convert region-liveness-drop{-,-no-}may-dangle.rs into ui tests
The "match exact bits of CFG" approach was fragile and uninformative.
2017-12-20 14:38:13 -05:00
Niko Matsakis
80c510e353 when using feature(nll), don't warn about AST-based region errors
Also, keep reporting AST-based region errors that are not occuring in
a fn body.
2017-12-20 14:38:13 -05:00
Santiago Pastorino
0b2db1e616 Add nll feature and make nll imply nll_dump_cause 2017-12-20 14:38:12 -05:00
Niko Matsakis
508a831dca use report_generic_bound_failure when we can in the compiler 2017-12-20 14:38:12 -05:00
Niko Matsakis
6b39781df6 connect NLL machinery to the NiceRegionError code 2017-12-20 14:38:12 -05:00
Niko Matsakis
e28d03f38b only dump causes if we have nothing better 2017-12-20 14:38:11 -05:00
Santiago Pastorino
3a185a510d Add three point error handling to borrowck
Closes #45988
2017-12-20 14:38:10 -05:00
Niko Matsakis
93afb1affc connect NLL type checker to the impl trait code
We now add the suitable `impl Trait` constraints.
2017-12-20 14:04:52 -05:00
Niko Matsakis
4a967c9df7 propagate region_bound_pairs into MIR type-check 2017-12-20 14:04:51 -05:00
Niko Matsakis
fe89f4ba86 get the DefiningTy from the body_owner_kind not type
The type isn't a good idea because we want to treat constants
uniformly, regardless of what type of value they produce.
2017-12-20 14:04:51 -05:00
bors
588f7db8ef Auto merge of #46733 - nikomatsakis:nll-master-to-rust-master-5, r=arielb1
nll part 5

Next round of changes from the nll-master branch.

Extensions:

- we now propagate ty-region-outlives constraints out of closures and into their creator when necessary
- we fix a few ICEs that can occur by doing liveness analysis (and the resulting normalization) during type-checking
- we handle the implicit region bound that assumes that each type `T` outlives the fn body
- we handle normalization of inputs/outputs in fn signatures

Not included in this PR (will come next):

- handling `impl Trait`
- tracking causal information
- extended errors

r? @arielb1
2017-12-20 03:58:15 +00:00
Niko Matsakis
4f43c5b1e7 stop dumping DefPath into "failed type test" errors
The prior messages were not stable across platforms.
2017-12-15 17:04:48 -05:00
Esteban Küber
b562565b09 Same change to point at borrow for mir errors 2017-12-15 13:52:05 -08:00
Niko Matsakis
0d6bd42abb make blame_span deterministic 2017-12-15 10:27:54 -05:00
Niko Matsakis
a118afe7ca add a test regarding relating closure and fn generics
Turns out this works but we had no test targeting it.
2017-12-15 10:27:53 -05:00
Niko Matsakis
e9824c50ed impose inputs/ouputs on MIR after the fact
The input/output types found in `UniversalRegions` are not normalized.
The old code used to assign them directly into the MIR, which would
lead to errors when there was a projection in a argument or return
type. This also led to some special cases in the `renumber` code.

We now renumber uniformly but then pass the input/output types into
the MIR type-checker, which equates them with the types found in MIR.
This allows us to normalize at the same time.
2017-12-15 10:27:53 -05:00
Niko Matsakis
3fcb13ae45 handle projections with regions 2017-12-15 10:27:52 -05:00
Niko Matsakis
85e1d4749e propagate type tests from closure to closure creators
Currently, we only propagate type tests that exclude all regions from
the type.
2017-12-15 10:27:51 -05:00
Niko Matsakis
c7cfa2367b thread through an implicit region body of the fn body 2017-12-15 10:27:50 -05:00
Niko Matsakis
47c1921b9a move some parts of liveness to happen during type checking
This allows us to re-use the `normalize` method on `TypeCheck`, which
is important since normalization may create fresh region
variables. This is not an ideal solution, though, since the current
representation of "liveness constraints" (a vector of (region, point)
pairs) is rather inefficient. Could do somewhat better by converting
to indices, but it'd still be less good than the older code. Unclear
how important this is.
2017-12-15 10:27:49 -05:00
Niko Matsakis
6193c5cc2a translate Verifys into TypeTests and check them 2017-12-15 10:10:56 -05:00
Niko Matsakis
cd564d20ff only propagate ClosureRegionRequirements if non-trivial
Before, we would always have a `Some` ClosureRegionRequirements if we
were inferring values for a closure. Now we only do is it has a
non-empty set of outlives requirements.
2017-12-15 10:10:56 -05:00
Niko Matsakis
a30e2259da fix closure tests now that MIR typeck works properly
These tests had FIXMEs for errors that were not previously being
reported.
2017-12-13 12:20:27 -05:00
David Wood
3dbc11bc63
Updated existing tests with new error messages. 2017-12-11 22:20:15 +00:00
Tommy Ip
2ad5bf056c Update ui tests' line numbers. 2017-12-10 21:14:57 +00:00
Tommy Ip
76d9b309ae Add must-compile-successfully comment to appropriate ui tests. 2017-12-10 21:14:57 +00:00
Niko Matsakis
47037705dc add test comparing free region to bound region
suggested by arielb1
2017-12-07 05:28:04 -05:00
Niko Matsakis
86e7b5c12a use -Znll -Zborrowck=mir -Zverbose consistently 2017-12-07 05:28:04 -05:00
Niko Matsakis
75cf482651 region-ebr-does-not-outlive-static: reuse old test instead 2017-12-07 05:28:04 -05:00
Niko Matsakis
b8e9eaf3ae propagate-approximated-ref: include FIXME comment 2017-12-07 05:28:03 -05:00
Niko Matsakis
2a17b9390b test more failure cases for approximating the "longer" side
as suggested by arielb1
2017-12-07 05:28:03 -05:00
Niko Matsakis
ca60826110 test approximations of the "shorter side" that resort to 'static
This includes the additional case suggested by arielb1.
2017-12-07 05:28:03 -05:00
Niko Matsakis
3c42f18c37 delete stray file 2017-12-07 05:28:03 -05:00
Niko Matsakis
6ee31e2af7 escape-upvar-ref: patch comment per arielb1 2017-12-07 05:28:03 -05:00
Niko Matsakis
8b657d378e escape-upvar-nested: patch "dangling reference" to renamed test 2017-12-07 05:28:02 -05:00
Niko Matsakis
782b9f5c0a escape-argument: improve comment per arielb1's suggestion 2017-12-07 05:28:02 -05:00
Niko Matsakis
dac59ac93a escape-argument-callee: improve comment per arielb1's suggestions 2017-12-07 05:28:02 -05:00
Niko Matsakis
47b8ef51d4 don't print self symbol's internal index unless gensymed
It's just not useful. It also makes it hard to have tests that probe
internal state, since the interning number is very sensitive.

Dumping the number in the case of gensym is not ideal but will do for
now.
2017-12-07 05:28:01 -05:00
Niko Matsakis
05441abd2b add closure requirement tests, improve debugging output
The overall format is now easier to read. Also, There is now graphviz
output, as well as a `#[rustc_regions]` annotation that dumps internal
state.
2017-12-07 05:28:00 -05:00
Paul Daniel Faria
a9cb25b23a inform constraint generation using maybe-init
In particular, if we see a variable is DROP-LIVE, but it is not
MAYBE-INIT, then we can ignore the drop. This leavess attempt to use
more complex refinements of the idea (e.g., for subpaths or subfields)
to future work.
2017-12-04 08:51:11 -05:00
est31
755fa9c23e Update tests for -Zborrowck-mir -> -Zborrowck=mode migration 2017-11-26 16:39:15 +01:00
Oliver Schneider
8937d6a6cf
Merge cfail and ui tests into ui tests 2017-11-24 11:32:35 +01:00
Basile Desloges
094d67ee37 mir-borrowck: Remove parens in the lvalue description of a deref 2017-11-19 20:19:10 +01:00
Niko Matsakis
4b743da596 integrate NLL with MIR type-checker 2017-11-16 05:57:47 -05:00
Niko Matsakis
7b4282e02b emit less info when we warn
Actually, I meant to make this use `delay_span_bug`
2017-11-02 10:34:25 -04:00
Niko Matsakis
fdcd2256f0 trace span info for constraints and report errors 2017-11-02 10:34:25 -04:00