Commit Graph

1791 Commits

Author SHA1 Message Date
Marijn Haverbeke
9538b00363 Tuple types back, not constructable yet 2011-08-15 11:40:38 +02:00
Michael Sullivan
0340f32748 Eliminate autoderef on binops and unary negation.
Autoderef on binops is basically unused, kind of silly, and
complicates typechecking. There were only three instances of it in the
compiler and the test drivers, two of which were of the form "*foo =
foo + 1", which should be written as "*foo += 1" anyways.
2011-08-12 18:28:03 -07:00
Michael Sullivan
8e92be3b94 Small whitespace cleanups. 2011-08-12 16:36:50 -07:00
Tim Chevalier
c01e487485 Un-XFAIL do-while-body-fails
Made it work both with optimization enabled, and with optimization
disabled. Huzzah!
2011-08-12 16:26:10 -07:00
Brian Anderson
594c7fb0c6 Remove the last remaining vec expressions in rustc 2011-08-12 16:13:13 -07:00
Michael Sullivan
2421312dea Don't rely on binop autoderef in the compiler. 2011-08-12 15:16:01 -07:00
Brian Anderson
7625ed52ee Remove vecs from std::sort 2011-08-12 12:14:06 -07:00
Brian Anderson
119f43e0c7 Rename str::connect_ivec to str::connect 2011-08-12 12:08:27 -07:00
Brian Anderson
740196987e Rename std::str::unsafe_from_bytes_ivec to unsafe_from_bytes 2011-08-12 12:08:27 -07:00
Brian Anderson
a9ce342fa3 Convert all uses of unsafe_from_bytes to unsafe_from_bytes_ivec 2011-08-12 12:08:26 -07:00
Brian Anderson
49b80f9bf7 Remove str::from_bytes
This is exactly the same as str::unsafe_from_bytes
2011-08-12 12:08:26 -07:00
Marijn Haverbeke
de4b383a0f Properly typecheck unary minus
Closes #813
2011-08-12 16:05:56 +02:00
Tim Chevalier
051b401051 Handle _|_ - typed things in the bodies of do-while loops
The resulting code is strange, but perhaps someone else can
fix it. The obvious things, like returning body_res, all resulted
in completely incomprehensible LLVM errors.

Closes #814
2011-08-11 17:23:47 -07:00
Patrick Walton
5079f51386 rustc: Associate type descriptors with allocas as metadata 2011-08-11 14:35:33 -07:00
Brian Anderson
4ff6763b60 rustc: Lowercase "main function not found" error 2011-08-11 10:09:02 -07:00
Brian Anderson
0cddcac016 rustc: Use spans in some main typeck errors 2011-08-11 10:09:02 -07:00
Brian Anderson
612fe3d2d6 Add ast_map::node_span function 2011-08-11 10:09:02 -07:00
Brian Anderson
14b3ad194a Switch a check for main type from an error to an ICE
This code path doesn't look possible, so I think it indicates a bug. Also,
make the message lowercase.
2011-08-11 10:09:02 -07:00
Brian Anderson
0f8e58e3ce Lowercase the "wrong type in main fn" error 2011-08-11 10:09:02 -07:00
Patrick Walton
7b337e3581 rustc: Don't free shared memory when --gc is on 2011-08-10 22:47:18 -07:00
Michael Sullivan
c5c8258937 Rework check_expr substantially.
The bulk of check_expr is now check_expr_with_unifier, which takes an
expected type and a unification function and will perform the
unification on the type it produces. check_expr calls
check_expr_with_unifier with a dummy unifier and a new function,
check_expr_with, takes an expected type and uses the simple unifier.

I think this generally makes thing cleaner, but the purpose for doing
this is to enable type inferred lambda-blocks to be useful by allowing
the argument types to be unified before the body of the lambda is
checked.
2011-08-10 20:18:07 -07:00
Michael Sullivan
67e361a940 Introduce a ty_infer ast node and use it instead of option::t[ty].
This actually basically makes things worse, since we get less nice
type system guarentees but it will make doing type inferred blocks a
fair deal less painful. I'm not /really/ happy about this...
2011-08-10 20:18:07 -07:00
Patrick Walton
1e6074ca73 rustc: Mark functions as "rust" GC 2011-08-10 20:12:30 -07:00
Michael Sullivan
43c9fe65bd Eliminate the last vestiges of init_recv. 2011-08-10 19:21:29 -07:00
Michael Sullivan
00b781e495 Some cleanup in check_expr. 2011-08-10 19:20:00 -07:00
Patrick Walton
ba7c8f18d4 rustc: Remove unused "trace" upcalls 2011-08-10 18:09:52 -07:00
Patrick Walton
d1494901d0 rustc: Declare GC-related intrinsics 2011-08-10 17:59:33 -07:00
Patrick Walton
19424dfab6 rustc: Add unique pointers to the set of types we support 2011-08-10 17:24:22 -07:00
Tim Chevalier
a8a4d4ec05 Use actual type, not declared type, when zeroing move arguments
trans was failing with a bounds check error because the caller
was using the declared type (an out-of-scope ty param) and not
the actual type in a list of argument types to zero.

Closes #811
2011-08-10 17:22:57 -07:00
Lindsey Kuper
bf81a05223 Factor out creation of object body types. 2011-08-10 16:53:32 -07:00
Michael Sullivan
bcc30cb4f6 Rename unify::simple to unify::unify. 2011-08-10 16:43:03 -07:00
Michael Sullivan
7d5092ee7b Some trivial cleanup. 2011-08-10 16:28:34 -07:00
Lindsey Kuper
007af36bb3 Comments, cleanup, whitespace, refactoring. 2011-08-10 15:11:32 -07:00
Patrick Walton
adce35acd4 rustc: Use polymorphic logging 2011-08-10 14:35:12 -07:00
Patrick Walton
8e6e6f5f89 rustc: Make iter_structural_ty_full and friends take one value, not two. Shaves a second off codegen. 2011-08-09 18:31:14 -07:00
Patrick Walton
044b16ad68 rustc: Remove the code to generate cmp glue 2011-08-09 18:31:13 -07:00
Graydon Hoare
a5997f2eb2 Actually perform handoff from caller to callee on move-mode args. 2011-08-09 17:56:26 -07:00
Patrick Walton
865ca749dc rustc: Use shapes instead of codegen'd cmp glue 2011-08-09 17:02:17 -07:00
Lindsey Kuper
5d9680fc7e Move object-system-related trans stuff to its own file. 2011-08-09 16:42:55 -07:00
Brian Anderson
61425c8d13 Convert remaning ivec decls to new syntax
These snuck in during rebasing
2011-08-09 15:53:26 -07:00
Erick Tryzelaar
00ccd6ba42 Remove support for the ivec T[] syntax. 2011-08-09 15:53:26 -07:00
Erick Tryzelaar
8b15045224 Port the compiler to the ivec type [T] syntax. 2011-08-09 15:53:26 -07:00
Lindsey Kuper
d5c8bb9d86 A little backwarding cleanup/refactoring. 2011-08-09 11:19:24 -07:00
Lindsey Kuper
a334adaab9 Thread "self" through the stack. Backwarding! Closes #702. 2011-08-09 10:59:13 -07:00
Graydon Hoare
b2cac5afa3 Implement typestate checking for move-mode args. Un-XFAIL compile-fail/move-arg.rs. 2011-08-08 19:41:54 -07:00
Patrick Walton
72fdb1a3bf rustc: Emit the shape of "float" as f64, not f32 2011-08-08 16:52:10 -07:00
Michael Sullivan
059c66b9f7 Don't bother zeroing out slots in cleanups. 2011-08-08 16:17:09 -07:00
Graydon Hoare
b54e7e4506 Add new arg-passing mode 'move' denoted with '-T'. Translate as pass-by-value, doesn't deinit source yet nor get proper analysis in typestate, alias passes. 2011-08-08 15:53:41 -07:00
Paul Stansifer
a9471d8296 Revert "rustc: bzero in zero_alloca. Shaves off a second or three."
This causes trouble in Valgrind in drop glue in parsing.

This reverts commit 4d180793f0.
2011-08-05 16:38:49 -07:00
Eric Holk
b62e80c1f0 Atomic ref counting for chans. 2011-08-05 15:27:28 -07:00