mikros/rust
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
3,873 Commits 2 Branches 0 Tags 1.4 GiB
cdda0fd6d6
Commit Graph

1517 Commits

Author SHA1 Message Date
Marijn Haverbeke
86f16b90cd Support type parameters in resources 
Some rather dodgy code was added to trans in the process. I'd love to
discuss it with someone who knows more about types during translation.
 2011-06-30 18:36:12 +02:00
Marijn Haverbeke
31ec26d46a Make calling resource destructors work cross-crate 
Tydescs are currently re-created for each compilation unit (and I
guess for structural types, they have to be, though the duplication
still bothers me). This means a destructor can not be inlined in the
drop glue for a resource type, since other crates don't have access to
the destructor body.

Destructors are now compiled as separate functions with an external
symbol that can be looked up in the crate (under the resource type's
def_id), and called from the drop glue.
 2011-06-30 18:36:12 +02:00
Graydon Hoare
c796a8f24d Re-enable tidy (it was broken) and fix various non-tidy things. 2011-06-29 15:14:55 -07:00
Tim Chevalier
76b755f9ad Remove workaround for a compiler bug that, I guess, got fixed 2011-06-29 14:28:03 -07:00
Tim Chevalier
1ba85932b1 Add a runtime flag to enable/disable claims en masse 
Now, if the environment variable CHECK_CLAIMS is set, then all
claims turn into checks. Otherwise, claims are no-ops.
 2011-06-29 14:28:02 -07:00
Brian Anderson
29afe1a518 Write metadata for sythesized name/vers items to the crate's link attr 
If the crate doesn't specify it's name or version, the defaults need to be
exported in the crate metadata
 2011-06-29 14:21:39 -07:00
Brian Anderson
325f46da06 Move the calculation of the crate name, vers and cmh to a single function 2011-06-29 14:21:39 -07:00
Graydon Hoare
5b2c17fba5 Fix resource encoding bugs. 2011-06-29 12:14:29 -07:00
Patrick Walton
b18cefdfd7 rustc: Fix a leak that resulted from copying a structural interior type containing interior vectors. Add a test case. 2011-06-29 12:16:07 -07:00
Patrick Walton
bbd0d58955 rustc: Unbreak self assignment 2011-06-29 10:22:53 -07:00
Michael Sullivan
1d05718c2d Extend visit to use visit_fn to visit fn expressions. 2011-06-29 09:04:00 -07:00
Brian Anderson
7c500fc0a0 Replace common::new_seq_hash with an adapter around std::smallintmap 
It would be better to either convert ast_map to use smallintmap or make
smallintmap and hashmap follow the same interface, but I don't feel up to
it just now. Closes #585.
 2011-06-28 19:48:33 -07:00
Lindsey Kuper
7b5d34aa9a Comments and cleanup. 2011-06-28 19:06:44 -07:00
Lindsey Kuper
57e5cde3a2 Comments and cleanup. 2011-06-28 18:54:56 -07:00
Brian Anderson
6832f75f0b Support attaching attributes to modules via the crate file. Issue #487 2011-06-28 18:39:34 -07:00
Patrick Walton
74f8eb51ae rustc: Move duplicate_heap_parts to copy glue; add a test case 2011-06-28 17:59:30 -07:00
Tim Chevalier
db478ed4de Remove outdated comments 2011-06-28 17:50:07 -07:00
Michael Sullivan
b944d8d72e Require that both sides of a swap be lvals. 2011-06-28 17:28:06 -07:00
Brian Anderson
83b7215366 Cleanup attribute code. Issue #487 2011-06-28 16:55:07 -07:00
Patrick Walton
dfcfef41df rustc: Rename take glue to copy glue 2011-06-28 16:52:59 -07:00
Tim Chevalier
9f1444c701 Implement "claim" 
Implement "claim" (issue #14), which is a version of "check" that
doesn't really do the check at runtime. It's an unsafe feature.

The new flag --check-claims turns claims into checks automatically --
but it's off by default, so by default, the assertion in a claim
doesn't execute at runtime.
 2011-06-28 16:41:06 -07:00
Eric Holk
866ee6ecb4 Re-enabling some tests. 2011-06-28 16:12:34 -07:00
Eric Holk
64596e6583 Teach the parser and typechecker to understand port[int](). Closes #588 2011-06-28 16:11:04 -07:00
Brian Anderson
8ac60e99ea Extract meta_item sorting from back::link to middle::attr 2011-06-28 15:46:09 -07:00
Brian Anderson
aecc5e6e0f Add comments to middle::attr. Issue #487 2011-06-28 15:34:27 -07:00
Brian Anderson
a83b64d15d Rework how linkage attributes are determined 
The meta items within a crate's link attribute are used in linkage:

 #[link(name = "std",
        vers = "1.0",
        custom = "whatever")];

Name and vers are treated specially, and everything else is hashed together
into the crate meta hash.

Issue #487
 2011-06-28 15:25:47 -07:00
Brian Anderson
0608e277b6 Add some helper functions for attributes. Issue #487 2011-06-28 15:25:47 -07:00
Tim Chevalier
28459ca0eb Handle lazy binops properly in typestate 
The typestate analysis now reflects that the second operand of a
logical and or or may not be evaluated.
 2011-06-28 13:07:52 -07:00
Patrick Walton
1ca0db5421 rustc: Add a "type-owns-heap-mem" cache. 2x translation speedup. 2011-06-28 11:36:11 -07:00
Patrick Walton
5f5b7e3348 rustc: Don't call cmp glue if the type is a simple scalar 2011-06-28 11:24:34 -07:00
Marijn Haverbeke
fb14afd5eb Add test cases for resources 2011-06-28 18:00:38 +02:00
Marijn Haverbeke
07125aa38b Fix bug that prevented boxes-in-resources from being freed 2011-06-28 18:00:05 +02:00
Marijn Haverbeke
e2cb97a2fc Add some missing cases for ty_res in ty.rs 2011-06-28 18:00:05 +02:00
Marijn Haverbeke
d730bb730a Make it possible to use * to dereference a resource 2011-06-28 18:00:05 +02:00
Paul Stansifer
c3901cdf8e Add simple syntax extension (#simplext) 2011-06-28 15:02:12 +00:00
Marijn Haverbeke
be748fc215 Properly guard resource values against double-drop 
Ths involved adding an extra field to their representation, so that there is
something to check for zero.
 2011-06-28 16:14:15 +02:00
Patrick Walton
663aa76635 rustc: Call cmp glue directly if we statically know it 2011-06-27 18:35:01 -07:00
Patrick Walton
dfccd5a138 rustc: Fix stack explosion on -O0 from Alloca() instead of alloca() in cmp glue 2011-06-27 18:27:26 -07:00
Tim Chevalier
85b5b2a8e4 Tests for while loops that may invalidate constraints 
Wrote some small test cases that use while loops and moves, to
make sure the poststate for the loop body gets propagated into the
new prestate and deinitialization gets reflected.

Along with that, rewrite the code for intersecting states. I still
find it dodgy, but I guess I'll continue trying to add more tests.
Also, I'll probably feel better about it once I start formalizing
the algorithm.
 2011-06-27 18:14:23 -07:00
Tim Chevalier
6d1050b1c7 pure_exp should set the state, not extend it 
This fixes a bug where de-initializations were getting masked
(and programs that used a variable that had been de-initialized
snuck through).
 2011-06-27 18:14:23 -07:00
Brian Anderson
df2cfac07f Move what's left of metadata::cwriter into middle::trans 2011-06-27 16:30:27 -07:00
Brian Anderson
119193497e Extract metadata::decoder from metadata::creader 2011-06-27 16:30:27 -07:00
Brian Anderson
5601a6f534 Move metadata::cwriter::encode to metadata::tyencode 2011-06-27 16:30:27 -07:00
Brian Anderson
ba5c7a570d Rename middle::metadata to metadata::cwriter. Move creader to metadata 
Preparation for a lot more metadata refactoring
 2011-06-27 16:30:27 -07:00
Brian Anderson
d4a5487e1c Limit exports from middle::metadata 2011-06-27 16:30:27 -07:00
Patrick Walton
e4b5d910dc rustc: Directly emit calls to glue if possible 2011-06-27 15:38:21 -07:00
Brian Anderson
36fb754476 Index obj and res ctor metadata by ctor id, not item id 
Both the type and the ctor were trying to use the same id and this was making
creader unable to find the ctor when running 'rustc --ls'
 2011-06-27 12:34:44 -07:00
Tim Chevalier
a6f36ba3f5 removing log statement that snuck in there 2011-06-27 11:38:07 -07:00
Tim Chevalier
1656fa19ef Implement move for scalar values 
Before, something like:

let int y = 42;
let int x;
x <- y;

would cause an LLVM assertion failure. Fixed it.
 2011-06-27 11:32:06 -07:00
Brian Anderson
29a8219c4c Add a very simple map implementation for sequential integer keys 
Use it for the ast_map. Cuts 40% off the time spent prior to LLVM.
 2011-06-26 20:37:08 -07:00
Brian Anderson
08b49a5d8f Reformulate an assert in ty::tag_variants 
This was doing a redundant hashmap lookup. Removing the redundancy trims 5%
(2.8s) off rustc's compile time
 2011-06-26 18:02:47 -07:00
Brian Anderson
cfaa0f4b91 Fail typechecking for bad binop/type combinations 
Includes assignment operations. Add regression tests for lots of less useful,
less used or unexpected combinations, as well as a selection of compile-fail
tests. Closes #500 (again!)
 2011-06-25 19:42:59 -07:00
Marijn Haverbeke
afa632124f Use single-bar or to make tstate/states.rs prettier 
Sorry. This is the kind of thing I do when I'm on a plane and too
tired to manage anything that requires thinking.
 2011-06-25 21:16:03 +02:00
Marijn Haverbeke
7432017d5e Primitive support for non-copyable values 2011-06-25 21:15:11 +02:00
Marijn Haverbeke
f6753be655 Allow moving out of temporary values 
This will probably need more work, as moving doesn't appear to do
quite the right thing yet in general, and we should also check
somewhere that we're not, for example, moving out the content out of
an immutable field (probably moving out of fields is not okay in
general).
 2011-06-25 21:15:11 +02:00
Marijn Haverbeke
61fc12d0d0 Partial implementation of resources 
Non-copyability is not enforced yet, and something is still flaky with
dropping of the internal value, so don't actually use them yet. I'm
merging this in so that I don't have to keep merging against new
patches.
 2011-06-25 21:15:11 +02:00
Marijn Haverbeke
781a265b88 Remove variable name 'res' from test suite 2011-06-25 21:15:04 +02:00
Tim Chevalier
582e1f13f0 Invalidate constraints correctly after an assignment expression 
Modified typestate to throw away any constraints mentioning a
variable on the LHS of an assignment, recv, assign_op, or on
either side of a swap.

Some code cleanup as well.
 2011-06-24 22:36:53 -07:00
Michael Sullivan
818d7c9398 Use visit_fn_{pre,post} to avoid going past lambdas during writeback instead of explicitly casing. 2011-06-24 15:55:08 -07:00
Michael Sullivan
b732ec6f82 Have walk call visit_fn_pre and visit_fn_post on fn expressions. 
As a side effect, this fixes a bug where typestate would not properly be checked
inside of fn expressions.
 2011-06-24 15:46:20 -07:00
Michael Sullivan
f3717da1b1 Modify the fn vistors in walk so that they can handle functions without names. Update the typestate code to understand this. 2011-06-24 15:46:20 -07:00
Patrick Walton
baf474443a rustc: Increment by the unit size in iter_structural_ty when iterating over interior vectors if the unit size is dynamic 2011-06-24 17:36:30 -04:00
Marijn Haverbeke
2ebd19477b Sanitize use of ids for obj constructors. 
Typeck and trans used to, by historical coincidence, use the item_obj
node id, which was used to identify the obj type by the rest of the
system, for the constructor function. This is now identified by the
ctor id stored in the tag throughout.
 2011-06-24 21:22:52 +02:00
Marijn Haverbeke
9643aedb04 Remove uses of variable name 'res' from rustc 
This in preparation of making 'res' a keyword for defining resources.
Please don't introduce too many new ones in the meantime...
 2011-06-24 21:22:23 +02:00
Marijn Haverbeke
3d7fdb509a Remove def_obj 
Since obj constructors and types have different def_ids now,
their def can simply be a def_fn and a def_ty.
 2011-06-24 21:13:25 +02:00
Patrick Walton
b4c0893a39 rustc: If needed, duplicate types' heap parts recursively through structural types 2011-06-24 14:32:50 -04:00
Brian Anderson
b9fc4dfc54 rustc: Cleanup trans_if, trans_alt 2011-06-23 22:16:16 -07:00
Brian Anderson
05c0216654 rustc: Add ternary operator. Closes #565 
The implementation is so simple it might be considered cheating: at almost
every step the expr_ternary is just converted to expr_if.
 2011-06-23 22:16:11 -07:00
Michael Sullivan
5495ad17d1 Fix type inference inside of anonymous functions. 2011-06-23 21:36:10 -07:00
Patrick Walton
f3798fd213 rustc: Load immediate types when iterating over interior vectors 2011-06-23 17:43:03 -04:00
Tim Chevalier
3cf00c1577 Correct error message for argument mode mismatch 
If you use a function expecting an alias argument in a context that
expects a function expecting a value argument, or vice versa, the
previous error message complained that the number of arguments was
wrong. Fixed the error message to be accurate.
 2011-06-23 11:54:10 -07:00
Tim Chevalier
9a48bd2f21 Compute typestate properly for move 
typestate now drops constraints correctly in the post-state of
a move expression or a declaration whose op is a move. It doesn't
yet drop constraints mentioning variables that get updated.

To do this, I had to change typestate to use trit-vectors instead
of bit-vectors, because for every constraint, there are three
possible values: known-to-be-false (e.g. after x <- y, init(y) is
known-to-be-false), known-to-be-true, and unknown. Before, we
conflated known-to-be-false with unknown. But move requires them
to be treated differently. Consider:

(program a)
(a1) x = 1;
(a2) y <- x;
(a3) log x;

(program b)
(b1) x = 1;
(b2) y <- z;
(b3) log x;

With only two values, the postcondition of statement a2 for
constraint init(x) is the same as that of b2: 0. But in (a2)'s
postcondition, init(x) *must* be false, but in (b2)'s condition,
it's just whatever it was in the postcondition of the preceding statement.
 2011-06-22 22:13:42 -07:00
Brian Anderson
54566e9037 rustc: Convert field access on invalid types from an ICE to a fatal error 
Closes #367
 2011-06-22 21:31:32 -07:00
Lindsey Kuper
d9f452a2a8 Some work on supporting forwarding slots in vtables (issue #539). 2011-06-22 21:07:20 -07:00
Brian Anderson
523a088451 rustc: Use the rhs span when unifying binops 
This makes the error message correctly identify the term that wasn't the
expected type. Issue #516.
 2011-06-22 20:50:53 -07:00
Brian Anderson
ae234d61ea rustc: Remove some unneeded type annotations 2011-06-22 20:45:02 -07:00
Brian Anderson
d2b7ea877c rustc: Don't commit unification changes until unify succeeds 
This is so that subsequent reports about type mismatches get the types
correct.

Issue #516
 2011-06-22 20:43:50 -07:00
Brian Anderson
ab98f16373 rustc: Downcase metadata::Encode 2011-06-22 18:09:53 -07:00
Paul Stansifer
b4c3b83f26 Fix bug: globbed imports were importing everything visible from the other 
module, not just everything exported.
 2011-06-22 15:41:39 -07:00
Brian Anderson
adc18bb24a rustc: Introduce and parse additional meta_item forms 
Examples: #[test], #[link(name = "vers")]

Issue #487
 2011-06-22 13:55:58 -07:00
Michael Sullivan
7a4fb084f1 Pull environment loading out into a function. 2011-06-22 07:03:16 -07:00
Michael Sullivan
6b40cedcb1 Pull building the environment for for_each into a seperate function. 2011-06-22 07:03:16 -07:00
Michael Sullivan
c07443e6eb Removing out-of-date comment. 2011-06-22 07:03:15 -07:00
Tim Chevalier
1a6d029b07 Emit a better error message for unbound type parameters in nested functions 
This code was causing a bounds check failure:

fn hd[U](&vec[U] v) -> U {
  fn hd1(&vec[U] w) -> U {
    ret w.(0);
  }
  ret hd1(v);
}

because in hd1, U was being treated as if it referred to a type
parameter of hd1, rather than referring to the lexically enclosing binding
for U that's part of hd.

I'm actually not sure whether this is a legit program or not. But I wanted
to get rid of the bounds check error, so I assumed that program shouldn't
compile and made it a proper error message.
 2011-06-21 17:58:49 -07:00
Tim Chevalier
7fb35ecf84 Serialize constraints in types (literal arguments still not supported) 
This involved, in part, changing the ast::def type so that a def_fn
has a "purity" field. This lets the typechecker determine whether
functions defined in other crates are pure.

It also required updating some error messages in tests. As a test
for cross-crate constrained functions, I added a safe_slice function
to std::str (slice(), with one of the asserts replaced with a
function precondition) and some test cases (various versions of
fn-constraint.rs) that call it. Also, I changed "fn" to "pred" for
some of the boolean functions in std::uint.
 2011-06-21 14:37:33 -07:00
Marijn Haverbeke
3b6d94d489 Move names and ids of native items into their recs, rather than their tags 2011-06-21 23:11:00 +02:00
Lindsey Kuper
26d8eaefa7 Puts out burning tinderbox (oops, AST nodes don't have def_ids/anns). 2011-06-21 14:07:28 -07:00
Lindsey Kuper
84005fadbf Some progress on support for extending objects with new fields (issue 
into four separate issues (#538, #539, #540, #543) with corresponding
tests.
 2011-06-21 13:31:27 -07:00
Marijn Haverbeke
c34e9b33d9 Move expr ids into the expr record type 
This simplifies the tag variants a bit and makes expr_node_id
obsolete.
 2011-06-21 22:25:32 +02:00
Patrick Walton
c109c9dd73 rustc: Output an unimplemented message when alias encounters an unknown sequence type 2011-06-21 11:10:14 -04:00
Patrick Walton
062c6b946b rustc: Implement "for" over interior vectors 2011-06-21 11:10:14 -04:00
Marijn Haverbeke
56eb986584 Fix some uses of span_warn in alias.rs that should be span_fatal 2011-06-20 23:53:29 +02:00
Marijn Haverbeke
588dc3897b Make trans use the new ast_map instead of building its own 2011-06-20 23:53:28 +02:00
Marijn Haverbeke
854b3a9b73 Use ast_map in typeck, instead of building another index 2011-06-20 23:53:28 +02:00
Marijn Haverbeke
40db3aa6fb Add a separate AST mapping phase 
This will replace the various node_id-to-node mappings done in several
other passes. This commit already uses the new map in resolve, dropping
the ast_map that was built there before.
 2011-06-20 23:53:28 +02:00
Marijn Haverbeke
77af54bf6f Reduce confusing variable naming in resolve 
'id' is now used for node_ids, 'name' for idents
 2011-06-20 23:53:28 +02:00
Marijn Haverbeke
75681f9ad7 Get rid of def_ids and anns in AST nodes, use single node_id 
This reduces some redundancy in the AST data structures and cruft in
the code that works with them. To get a def_id from a node_id, apply
ast::local_def, which adds the local crate_num to the given node_id.
Most code only deals with crate-local node_ids, and won't have to
create def_ids at all.
 2011-06-20 23:53:28 +02:00
Eric Holk
a2dcd08cc2 Added string duplication to deep_copy. Closes #520. 2011-06-20 13:47:02 -07:00
Eric Holk
992153672b Removed duplicated ret_ty_of_fn. This also means the native function branch is working now, so this commit closes #506. 2011-06-20 12:04:33 -07:00
Eric Holk
5640bead58 Teach ret_ty_of_fn_ty to understand native functions. Closes #506. 2011-06-20 12:04:33 -07:00
Graydon Hoare
97cf57adc1 Fix pp blank-lines insertion at beginning of file and before 'let' decls. 2011-06-20 10:45:33 -04:00
Brian Anderson
200dbe4c4f Remove various rustboot workarounds 2011-06-19 20:48:20 -07:00
Patrick Walton
3f7380ccec rustc: Change smallintmap to use an ivec and use it for the node type table. 3x typechecking speedup. 2011-06-19 18:03:19 -07:00
Brian Anderson
6cf9b17086 rustc: Make name resolution errors less fatal 
Failure happens at the end of name resolution

Issue #440
 2011-06-19 15:23:11 -07:00
Brian Anderson
c5238d57eb rustc: Make import resolution errors less fatal 
Failure happens at the end of import resolution

Issue #440
 2011-06-19 14:35:50 -07:00
Brian Anderson
76a752e7a4 rustc: Export only what's needed from middle::ty 
The list is formidable.
 2011-06-19 12:27:43 -07:00
Brian Anderson
80a8af7b47 rustc: Export only what's used from middle::resolve 2011-06-19 12:27:43 -07:00
Brian Anderson
43427dae0c rustc: Rename session.span_err -> span_fatal, err -> fatal 
Issue #440
 2011-06-19 12:19:53 -07:00
Brian Anderson
cf9ed08a50 Revert previous 6 commits. Hopefully put out Windows fire. 
Revert "rustc: Export only what's needed from middle::ty"

This reverts commit 4255d58aa5.

Revert "rustc: Make name resolution errors less fatal"

This reverts commit b8ab9ea89c.

Revert "rustc: Make import resolution errors less fatal"

This reverts commit 92a8ae94b9.

Revert "rustc: Export only what's used from middle::resolve"

This reverts commit 4539a2cf7a.

Revert "rustc: Re-introduce session.span_err, session.err"

This reverts commit 7fe9a88e31.

Revert "rustc: Rename session.span_err -> span_fatal, err -> fatal"

This reverts commit c394a7f49a.
 2011-06-19 03:29:19 -07:00
Brian Anderson
4255d58aa5 rustc: Export only what's needed from middle::ty 
The list is formidable.
 2011-06-19 02:42:05 -07:00
Brian Anderson
b8ab9ea89c rustc: Make name resolution errors less fatal 
Failure happens at the end of name resolution

Issue #440
 2011-06-19 02:42:05 -07:00
Brian Anderson
92a8ae94b9 rustc: Make import resolution errors less fatal 
Failure happens at the end of import resolution

Issue #440
 2011-06-19 02:42:01 -07:00
Brian Anderson
4539a2cf7a rustc: Export only what's used from middle::resolve 2011-06-19 02:41:03 -07:00
Brian Anderson
c394a7f49a rustc: Rename session.span_err -> span_fatal, err -> fatal 
Issue #440
 2011-06-19 02:40:59 -07:00
Brian Anderson
514813148b rustc: Store the lhs and rhs of receive exprs in left to right order 
With the changing of receive semantics the parser has been putting the rhs
expression in the first argument of expr_recv and the lhs in the second, and
all subsequent passes have been referring to them backwords (but still doing
the right thing because they were assuming that lhs was the port and rhs was
the receiver).

This makes all code agree on what lhs and rhs mean for receive expressions.
 2011-06-18 20:23:05 -07:00
Brian Anderson
3852f1eee3 Typecheck block tail expressions that are fn return values 2011-06-18 20:23:05 -07:00
Brian Anderson
ddbb31eb8e rustc: Remove the meta keyword 
Issue #487
 2011-06-18 20:23:05 -07:00
Patrick Walton
4f13879218 rustc: Allocate derived tydescs before dynamically sized allocas 2011-06-18 18:43:07 -07:00
Patrick Walton
877f001512 rustc: Don't bother to supply an alignment argument to the memmove intrinsic; LLVM is very fussy about what it considers a constant. 2011-06-18 16:59:44 -07:00
Patrick Walton
2ba7729ae7 rustc: Cache results of type_has_pointers. 70% translation speedup. 2011-06-18 00:41:25 -07:00
Patrick Walton
5894e400b6 rustc: Add missing case for interior vectors in alias.rs 2011-06-17 19:16:26 -07:00
Tim Chevalier
7fc89f573d Restructure the "checking" pass in typestate 
I noticed that typestate was being lazier than it should be,
because it was only checking typestate for statements and
top-level expression (that is, the expression in a stmt_expr, but
not any subexpressions). So I rewrote the checks in tstate/ck.rs
to use walk, which exposed a few bugs in typestate that I fixed.

Also added some more test cases for if-check.
 2011-06-17 19:09:18 -07:00
Patrick Walton
a4eb6630f3 rustc: Increment by the dynamically-computed size in ivec::trans_append. Uncomment init_fn test in lib-ivec. 2011-06-17 19:01:14 -07:00
Patrick Walton
5fb9325848 rustc: Fix a bunch of memory management bugs relating to generic interior vectors. Uncomment all tests in lib-ivec. 2011-06-17 17:31:24 -07:00
Patrick Walton
68122edac3 Revert "rustc: Fix a bunch of memory management bugs relating to generic interior vectors" 
This reverts commit 2b5e40311d.
 2011-06-17 16:38:47 -07:00
Patrick Walton
2b5e40311d rustc: Fix a bunch of memory management bugs relating to generic interior vectors 2011-06-17 16:23:06 -07:00
Eric Holk
2c1b269960 Implemented enough of deep_copy that we can pass scalars and channels to tasks. Closes #507. 2011-06-17 15:17:46 -07:00
Eric Holk
81fc51a8d7 Started using deep_copy for spawn arguments. 2011-06-17 15:17:46 -07:00
Patrick Walton
e43a3755dd rustc: Add a type_owns_heap_mem() predicate 2011-06-17 12:32:18 -07:00
Patrick Walton
dfb806b367 rustc: Cast to opaque interior vector types in trans_vec_append if necessary. Puts out burning tinderbox. 2011-06-17 12:23:04 -07:00
Patrick Walton
4918c99f69 rustc: Copy the contents of generic interior vectors 2011-06-17 12:22:53 -07:00
Patrick Walton
4bb7597542 Revert "rustc: Copy the contents of generic interior vectors" 
This reverts commit 5bdbe1dfc2.
 2011-06-17 12:03:21 -07:00
Patrick Walton
5bdbe1dfc2 rustc: Copy the contents of generic interior vectors 2011-06-17 11:45:51 -07:00
Brian Anderson
e8228e1260 rustc: Demand that if conditions have bool type 
Closes #513
 2011-06-17 11:35:00 -07:00
Brian Anderson
a4d88dacc9 rustc: Fail to unify if two ty_vars don't unify 
Closes #500
 2011-06-17 11:35:00 -07:00
Brian Anderson
753da64fd6 rustc: Demand unification of both sides of a binop 
Issue #500
 2011-06-17 11:35:00 -07:00
Paul Stansifer
391348ec86 Consistify ast::local. 2011-06-16 17:56:24 -07:00
Patrick Walton
ab7ebb608a rustc: Fix dominance issue when translating generic interior vectors 2011-06-16 17:48:32 -07:00
Patrick Walton
e50c918e6b rustc: Implement type walking and comparison glue for unsafe pointers 2011-06-16 17:07:02 -07:00
Graydon Hoare
b84fffaa4e Reformat a bunch of recent churn. 2011-06-16 16:55:46 -07:00
Patrick Walton
d90ad442b4 rustc: Work around leak when translating interior vectors. Un-XFAIL interior-vec.rs. 2011-06-16 15:07:21 -07:00
Patrick Walton
e7575b9f3a rustc: Add cases for istr and ivec to visit 2011-06-16 14:26:11 -07:00
Lindsey Kuper
a5563e173e Clean up whitespace oddities perpetrated by pretty-printer. 2011-06-16 12:21:49 -07:00
Tim Chevalier
219924e669 Bring back if-check 
Add "if check" (expr_if_check), a variation on check that executes
an "else" clause rather than failing if the check doesn't hold.
 2011-06-16 12:13:20 -07:00
Michael Sullivan
272c7e2e87 Translate the swap operator. 2011-06-16 11:47:54 -07:00
Michael Sullivan
bc33a5fe50 Fix bug in move on structure types. 2011-06-16 11:47:54 -07:00
Michael Sullivan
cfd1f74ebb Bring swap through typechecking and typestate. 2011-06-16 11:47:54 -07:00
Patrick Walton
4673eb6839 rustc: Fix LLVM type error when generating cmp glue for interior vectors 2011-06-16 11:16:03 -07:00
Marijn Haverbeke
15f71b3600 Refactor ast::item representation 
Most of the fields in an AST item were present in all variants. Things
could be simplified considerably by putting them in the rec rather
than in the variant tags.
 2011-06-16 13:32:35 +02:00
Patrick Walton
1fa9133b76 rustc: Implement interior vector concatenation 2011-06-15 18:03:25 -07:00
Eric Holk
260aa408f3 Merge branch 'master' of github.com:graydon/rust into fastcall 2011-06-15 17:45:24 -07:00
Brian Anderson
94d971c0e8 rustc: Remove unused variable from check_assignment 2011-06-15 16:48:59 -07:00
Tim Chevalier
a5bca7df88 Refactor data structures representing constraints (again...) 
I added a "resolved" version of the ast::constr type -- ty::constr_def
-- that has a def_id field instead of an ann_field. This is more
consistent with other types and eliminates some checking.

Incidentally, I removed the def_map argument to the top-level function
in middle::alias, since the ty::ctxt already has a def_map field.
 2011-06-15 15:38:12 -07:00
Patrick Walton
52d4f48144 rustc: Move ivec stuff to an ivec module in trans 2011-06-15 14:43:00 -07:00
Marijn Haverbeke
9b5ebe726c Implement checking of alias safety in tail calls. 2011-06-15 21:39:27 +02:00
Marijn Haverbeke
b2b8002fd3 Fix unsafe uses of aliases in tail calls 2011-06-15 21:39:27 +02:00
Lindsey Kuper
fa5a7e7453 Simple anonymous objects get through translation. 2011-06-15 12:18:52 -07:00
Eric Holk
01ea27b205 Step 1 of moving task startup to always be cdecl. 2011-06-15 12:10:15 -07:00
Graydon Hoare
ce72993488 Reformat source tree (minus a couple tests that are still grumpy). 2011-06-15 11:19:50 -07:00
Brian Anderson
1c9277e0d3 rustc: Make room in remaining AST item nodes for attributes 
Issue #487
 2011-06-15 09:48:55 -07:00
Brian Anderson
31116dcdd5 rustc: Make room in item_mod for metadata 
Issue #487
 2011-06-15 09:46:37 -07:00
Marijn Haverbeke
c7ab88cc35 Implement checking against assignments to immutable obj fields 2011-06-15 14:59:51 +02:00
Marijn Haverbeke
0ddade3d7a Extend alias analysis to check assignments 
This is a somewhat odd place to put these checks, but the data tracked
by that pass, and the available functions, make it trivial to do such
a check there.
 2011-06-15 14:10:07 +02:00
Marijn Haverbeke
5fb518abc9 Fix assignments to immutable fields throughout the code 2011-06-15 14:10:07 +02:00
Marijn Haverbeke
180db08470 Refactor expr_root in alias.rs 
It was too clunky and specific before. Now returns a vec of dereference specs
instead of a bunch of special-case information. Further accessors extract
the information they need from this vec.
 2011-06-15 11:03:23 +02:00
Patrick Walton
ac743cfcb0 Merge pull request #490 from msullivan/fix_vec_append 
Move the implementation of vec_append from llvm assembly to a regular upcall
 2011-06-14 18:32:04 -07:00
Patrick Walton
c6be352f73 rustc: Implement on-heap resizing for interior vectors 2011-06-14 18:23:13 -07:00
Michael Sullivan
ccd1bcf7a7 Make tydesc glue functions use cdecl. 2011-06-14 16:09:21 -07:00
Michael Sullivan
5378f48a03 Remove vec_append_glue from the compiler. 2011-06-14 16:09:17 -07:00
Patrick Walton
14258b87a7 rustc: Free heapified interior vectors 2011-06-14 16:00:30 -07:00
Patrick Walton
bd18860be7 rustc: Fix LLVM type error that occurred when translating large unique vector literals 2011-06-14 16:00:02 -07:00
Michael Sullivan
95def699f1 Implement upcall_vec_append and use it for vector append. 2011-06-14 15:35:09 -07:00
Tim Chevalier
05b7eb6dff Make the typechecker check that dynamically sized arguments are passed by alias 
Before, this would cause an assertion failure in trans.
 2011-06-14 15:11:48 -07:00
Tim Chevalier
1513539b67 Improve error message for a failed assert or check 2011-06-14 14:17:28 -07:00
Patrick Walton
bf5840d530 rustc: Do a dynamic alloca for generic interior vectors; fix data pointer calculation when spilling vectors 2011-06-14 14:03:38 -07:00
Marijn Haverbeke
f24408d89a Don't error when trans_lval is called on a non-lval 
This allows directly calling a function expression, or the
result of an alt/if/block expression.
 2011-06-14 21:38:56 +02:00
Marijn Haverbeke
471436a299 Implement function expressions/anonymous functions 
Looks like 'fn(..arg..) -> ret {body;}. They don't support type
parameters or upvars yet.
 2011-06-14 18:22:46 +02:00
Patrick Walton
c3bf7d07e9 rustc: Implement simple interior vector append translation 2011-06-13 18:57:25 -07:00
Tim Chevalier
a1bb4a4ded Refactor some typestate-related data structures 2011-06-13 18:21:22 -07:00
Eric Holk
d1857d30fc This is the mega-ucontext commit. It replaces the task switching mechanism with a new one inspired by ucontext. It works under Linux, OS X and Windows, and is Valgrind clean on Linux and OS X (provided the runtime is built with gcc). 
This commit also moves yield and join to the standard library, as requested in #42. Join is currently a no-op though.
 2011-06-13 18:14:13 -07:00
Brian Anderson
1595c9d767 rustc: Assignment expressions have nil type. Closes #483 2011-06-13 17:57:43 -07:00
Tim Chevalier
d65ad8c31c Change decl to local in expr_for and expr_for_each 
Since the decl in a for or for-each loop must always be a local
decl, I changed the AST to express this. Fewer potential match
failures and "the impossible happened" error messages = yay!
 2011-06-13 17:33:04 -07:00
Patrick Walton
913383d775 rustc: Preserve more type information for interior vectors at the LLVM level, allowing SROA to take place. log ~[ 2 ].(0) now optimizes to just log 2. 2011-06-12 00:44:43 -07:00
Patrick Walton
770bf24d72 rustc: Use in-bounds GEPs for fetching length and data of interior vectors 2011-06-11 23:32:21 -07:00
Patrick Walton
4375329031 rustc: Translate interior vector index expressions 2011-06-11 23:15:16 -07:00
Patrick Walton
8b318be4ce rustc: Typecheck vector index expressions for interior vectors 2011-06-11 19:52:49 -07:00
Patrick Walton
00eb3f5798 rustc: Sketch out translation of interior vector literals and take/drop glue 2011-06-10 19:36:33 -07:00
Tim Chevalier
698c6406ba Reject programs with unsatisfied predicate constraints 
Generate appropriate constraints for calls to functions with
preconditions, and reject calls where those constraints don't
hold true in the prestate.

...by which I mean that it works for one test case :-)
 2011-06-10 19:17:15 -07:00
Lindsey Kuper
f6d7d75699 More work on typechecking anon objs. 2011-06-10 17:18:08 -07:00
Graydon Hoare
8876e2c29e Oops, fix logging string specs. Nobody likes writing RUST_LOG=_ZN5rustc5front7creaderE, do they? 2011-06-10 16:32:21 -07:00
Tim Chevalier
774c6d0198 Friendlier error message for break outside a loop 
The typechecker should really check for this, but at least now
there's an error message rather than an inexhaustive match failure
in trans.
 2011-06-10 14:34:01 -07:00
Tim Chevalier
e5eacf8ea9 Use RHS type when translating assignments 
In code like "auto foo = fail", a type gets inferred for foo
depending on how it's used. However, fail still has type _|_ and
still should be treated that way: particularly, its value shouldn't
be copied. Fixed trans to reflect that.
 2011-06-10 14:05:06 -07:00
Patrick Walton
f1431adb40 rustc: Add a generic struct_elt() function to trans 2011-06-10 13:49:22 -07:00
Patrick Walton
451fc176d3 rustc: Derive flags for unsafe pointers. Puts out burning tinderbox. 2011-06-10 13:13:46 -07:00
Graydon Hoare
52c4c9014f Encode meta tags in the crate and start sketching enhanced logic for resolving crate "use" directives. 
Now with extra "works on OSX" kludging.
 2011-06-10 12:57:29 -07:00
Patrick Walton
4634f236a9 rustc: Add some missing cases to ty.rs for interior vectors, and modify the pattern match so this is less likely to happen again. Add the LLVM type mapping as well. 2011-06-10 12:07:38 -07:00
Patrick Walton
cc3d484d9d rustc: Add missing cases for istr and ivec to equal_type_structures() 2011-06-10 11:00:56 -07:00
Patrick Walton
7b7c746c1e Merge pull request #447 from paulstansifer/quick_error_message_fix 
Error message, instead of segfault, when recursive types are used.
 2011-06-10 10:53:42 -07:00
Patrick Walton
dfdd6dbc54 rustc: Remove the bitwise not operator 2011-06-10 10:47:33 -07:00
Lindsey Kuper
7a066ba547 Dead code elimination. 2011-06-10 10:23:48 -07:00