suzanne.soy/typed-racket
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
4,096 Commits 13 Branches 1 Tag 9.2 MiB
c48abf6dff
Commit Graph

227 Commits

Author SHA1 Message Date
Matthew Flatt
1842aaa2bd fixes for some tests 
The `identifier-binding` function is now senstive to the "macro
introduction" scope that's on an indentifier provided to a currently
expanding macro. Unit tests for TR need to use
`syntax-local-intorduce` to remove it, in the same way that `tc-setup`
already does.

Also, recognize a class expansion when it's wrapped with
`#%expression`, since `class` now expands that way sometimes.
 2015-07-16 08:40:35 -06:00
Matthew Flatt
8c006da763 remove syntax-local-make-delta-introducer 2015-07-16 08:40:34 -06:00
Matthew Flatt
2d095c4112 adapt to letrec-syntaxes+values removal and quote-syntax extension 
The `quote-syntax` form now suports a `#:local` option, which is
used for type annotations to preserve binding context.

Since `letrec-syntaxes+values` is now removed by `local-expand`,
some patterns must be adjusted, including the pattern used
for type annotations.
 2015-07-16 08:40:34 -06:00
Matthew Flatt
2f00de99d0 declare empty namespace for submodules 
Using the enclosing module's context for the `#f` part of the
`(module* _ #f ....)` would generally enable a compact encoding of
bindings for the purposes of `module->namespace`. Disabling
`module->namespace` for the generated submodules seems more reliable,
and it seems unlikely that anyone will miss the extra debugging
capability (which is all that is lost for these submodules whose
content is, I think, unspecified).
 2015-07-16 08:40:34 -06:00
AlexKnauth
aec5fad4be make append preserve non-emptyness of first argument 
see https://github.com/racket/scribble/pull/8#discussion_r33423152
 2015-06-28 12:07:42 -04:00
Sam Tobin-Hochstadt
9a2b74dc33 Generate trivial constraints for missing type variables. 
If the type variable isn't mentioned anywhere, then just
pick `Any` as its substitution.

Fixes #152.
 2015-06-23 16:08:23 -04:00
Vincent St-Amour
6bef6e6ca8 Make type of abs and magnitude more precise. 
Fixes a failing test.
 2015-06-16 12:00:07 -05:00
Eric Dobson
0498f548c1 Fix the type of abs on negative zeros. 2015-06-16 10:43:01 -05:00
Asumu Takikawa
0892e23892 Resolve expected type in tc-lambda-unit 
Closes #144
 2015-06-10 12:23:39 -04:00
Asumu Takikawa
bc3443b393 Change Module-Path type to accommodate submodules 
Closes #84
 2015-06-10 12:21:51 -04:00
Sam Tobin-Hochstadt
91e3a1b5f0 Make sure that user-specified clauses come before default clauses in type-case. 
Avoids nasty hash ordering bugs.
 2015-06-02 17:32:33 -04:00
Asumu Takikawa
b63514a2c3 Allow more flexible keywords in ->* types 
Keyword argument types in ->* are now allowed intermixed with
positional arguments instead of at the end.

Closes #145
 2015-06-02 01:05:05 -04:00
Eric Dobson
8014db0f7d Make infer unfold Mu before spliting on Unions. 2015-05-19 20:44:15 -07:00
Eric Dobson
d4eaa47b27 Make subtype use hash of hashs. 
This has a 1% improvement on new-metrics.rkt by my measurements.
 2015-05-19 20:37:41 -07:00
AlexKnauth
e075523e5e change type of partition 2015-05-19 16:29:49 -04:00
Alexis King
a096857a88 Disable the #{} reader macro if a dispatch macro already exists on #\{ 2015-05-14 14:35:30 -07:00
AlexKnauth
b11e48fa93 don't raise error for require/typed in module-begin context 2015-05-14 12:13:14 -04:00
Alex Knauth
c6b1f8fdc0 syntax error for require/typed in untyped code 2015-05-14 12:13:14 -04:00
Asumu Takikawa
92d1dd1c5e Add sealing contracts for row polymorphic types 
This enables contract generation in the negative
direction (untyped->typed) for row polymorphic types
(basically mixin types).

Depends on `class-seal` and `class-unseal` in
the racket/class library.
 2015-05-13 20:52:32 -04:00
Marc Burns
609d6189e9 Improve typecheck for private field initializers 
This patch addresses two issues with `typed/racket/class`:

1. For multiple private fields declared with `define-values`, type
information does not propagate from the values produced by the
initialization expression to the declared fields. This breaks soundness
of private fields: A field can be annotated with a type that does not
contain the field's initial value.

This was resolved by keeping a table of temporary bindings introduced in
the expansion of the initializer along with their types. The field
setter's type is then checked against that of the corresponding
temporary.

2. The class body typechecker assumes that the `expr` of
a `define-values` clause will expand to a bare `(values vs ...)`.

This was resolved by generalizing the template for matching an expanded
`define-values` initializer and extracting the type information from the
`expr` instead of each element in `(vs ...)`.
 2015-05-12 18:42:05 -04:00
Asumu Takikawa
5751a2e1cf Refactor get-field checking to avoid bad names 2015-05-12 15:08:53 -04:00
Eric Dobson
90061c2b96 Add filters for trivial cases in arithemetic procedures. 2015-05-02 08:26:50 -07:00
Eric Dobson
22342d6a35 Simplify flsqrt type. 2015-05-02 08:26:50 -07:00
Eric Dobson
fc4e216fc2 Simplify flexp type. 2015-05-02 08:26:50 -07:00
Eric Dobson
142b52918c Fix types for flmin and flmax. 
flmin and flmax only take two arguments.
 2015-05-02 08:26:50 -07:00
Eric Dobson
848b69b73d Simplify flonum comparison types. 2015-05-02 08:26:50 -07:00
Eric Dobson
d288b14244 Expand type of fl/. 2015-05-02 08:26:50 -07:00
Eric Dobson
c50b6291e2 Cleanup fl* type. 2015-05-02 08:26:50 -07:00
Eric Dobson
811cd211e9 Tighten fl+ and fl- types. 2015-05-02 08:26:49 -07:00
Eric Dobson
c3c134780e Cleanup fx shifts. 2015-05-02 08:26:49 -07:00
Eric Dobson
3c978d82cb Tighten fxior and fxxor. 2015-05-02 08:26:49 -07:00
Eric Dobson
424658f6ab Tighten fxmin and fxmax. 
Replace Zero case with all NonNegative or NonPositive values.
 2015-05-02 08:26:49 -07:00
Eric Dobson
93b538b694 Cleanup types of fixnum comparisons. 
Unify Zero and One cases, and remove extraneous ones.
 2015-05-02 08:26:49 -07:00
Eric Dobson
114b9376f1 Tighten type of fxabs. 2015-05-02 08:26:49 -07:00
Eric Dobson
ad09e941a8 Tighten type of fxquotient 2015-05-02 08:26:49 -07:00
Eric Dobson
c923663e55 Tighten type of fx* 
Make multiplication by Zero return Zero and by One return the other arg.
 2015-05-02 08:26:49 -07:00
Eric Dobson
6bc24fdf83 Fixup type of fx- 
fx- is not commutative.
Output is always a fixnum.
 2015-05-02 08:26:48 -07:00
Eric Dobson
b6a53b8916 Cleanup types of fx+ 
Fix case for two Bytes which should be Index not PosIndex.
 2015-05-02 08:26:48 -07:00
Eric Dobson
86d5845bc4 Replace exclude-zero with commutative-equality/strict-filter. 
Make it more obvious what the generated type is.
Also removes NaN from the specific type for flonums because that was
incorrect.
 2015-05-02 08:26:48 -07:00
Eric Dobson
69e613c831 Simplify types for nan? and infinite? 2015-05-02 08:26:48 -07:00
Eric Dobson
e91cdad6b0 Tighten types on sgn. 2015-05-02 08:26:48 -07:00
Eric Dobson
ddd503a203 Fix type on integer-sqrt/remainder. 
It doesn't return exact integers when given inexact input.
 2015-05-02 08:26:48 -07:00
Eric Dobson
eeef1a7c55 Simplify types of integer-sqrt. 2015-05-02 08:26:40 -07:00
Eric Dobson
70fdc3dd13 Make flonum and ext-flonum types use the same definitions. 
This avoids problems where one could be changed to fix a bug and the
other would be missed.
 2015-04-28 20:27:44 -07:00
Vincent St-Amour
257d4ec9a7 Fix use of values in type of integer-sqrt/remainder. 
Closes PR 15034.
 2015-04-28 17:49:53 -04:00
Asumu Takikawa
a80f55946d Adjust error messages for : annotation form 
Print better error messages for expression like
(: 2 Zero). Also fix and improve error messages for
multiple/missing types.

Closes PR 15044
 2015-04-21 16:07:13 -04:00
Andrew Kent
cbbf5536d8 no objects for mutated private class fields 
tidy up ->acc
 2015-04-08 17:53:00 -04:00
Asumu Takikawa
7bf4314af4 Simplify Name type representation 
Avoids contract errors by using a simpler representation
and sticking with it throughout.
 2015-04-08 11:23:48 -04:00
Asumu Takikawa
d22bf8cc11 Fix type-name-env contracts and uses 2015-04-08 11:23:48 -04:00
Eric Dobson
f0c48cfb87 Make abs return its argument on -NonNegReal. 2015-04-07 09:50:33 -07:00
Eric Dobson
a522768b7e Make tc-subst restrict returned types based on argument types. 2015-04-07 09:50:32 -07:00
Eric Dobson
7a29e8e369 Simplify cases in comparisons with -Zero or -RealZero 2015-04-07 08:23:32 -07:00
Eric Dobson
1ad0a2b181 Remove extra cases in comparisons for -One 2015-04-07 08:23:32 -07:00
Eric Dobson
b77bd862a0 Stop trying to represent rest objects because they are currently broken. 
Closes #83.
 2015-04-07 08:13:38 -07:00
Eric Dobson
b8921ff4df Fix contract on subst-type. 
subst-type now can be called with Values? and such.
 2015-04-06 09:58:06 -07:00
Eric Dobson
a33b27a3c9 Remove many cases of = that are unnecessary. 2015-04-05 19:10:51 -07:00
Eric Dobson
7b762933c6 Simplify the type of zero?. 2015-04-05 19:07:12 -07:00
Eric Dobson
be225976d3 Make subtype work with restricted filters. 
Closes PR 15025.
 2015-04-05 19:07:12 -07:00
Eric Dobson
db3826c474 Add comments for changes to substitution. 2015-04-05 18:01:11 -07:00
Sam Tobin-Hochstadt
6a262c19fa Add prefab structs to HISTORY. 2015-04-03 21:33:39 -04:00
Sam Tobin-Hochstadt
0793021514 History update for Typed Racket. 2015-04-03 18:45:38 -04:00
Sam Tobin-Hochstadt
583ca906b3 Significantly reduce dependencies. 
In conjunction with a small change to syntax/parse, this means
that `typed/racket/base` no longer depends on `racket/set`,
`racket/contract/base`, or `racket/generic`.

Timings on my machine go from ~200ms for `#lang typed/racket/base`
as the whole file, to ~100ms. For comparison, `racket/base` is 30ms
and `#lang racket` is 150ms. `#lang typed/racket` is ~200ms with
this change.

Changes include:
 - not using `in-syntax`
 - switching to `syntax/parse/pre`
 - avoiding `template` from `syntax/parse`
 2015-04-03 12:50:55 -04:00
Asumu Takikawa
7b5478e0bc Make Place-Channel <: (Evtof Any) 2015-04-02 11:00:25 -04:00
Asumu Takikawa
c70910aaa0 Fix type restriction in values->tc-results 
Previously the restriction didn't account for traversing
the object type with the given path. This also relies on
the previous commit that adds subtyping on filters.

(because this change seems to introduce filters which
 only differ by an unrolling of a recursive type, but
 the old subtyping only worked for identical types)
 2015-04-02 10:58:45 -04:00
Asumu Takikawa
8acc86bb9b Add subtyping for types inside filters 
This makes a type like

  (-> Any Boolean : #:+ Integer)

a subtype of a type like

  (-> Any Boolean : #:+ Number)

For not filters, the direction is reversed.
 2015-04-01 18:54:00 -04:00
Eric Dobson
889d96ca78 Strengthen infinity checks in comparison operations. 
This adds checks for infinity when comparing against reals and accounts
for NaNs.
 2015-03-30 09:25:48 -07:00
Eric Dobson
3889036b12 Simplify types for positive? and negative?. 2015-03-30 09:12:51 -07:00
Eric Dobson
5de22c80f7 Make values->tc-results restrict to the original type of the object. 2015-03-30 09:12:25 -07:00
Eric Dobson
e5ea1f4bb6 Make zero? sound with regards to NaN. 2015-03-28 14:08:20 -07:00
Sam Tobin-Hochstadt
718bc82b5d Avoid providing def-redirect. 2015-03-26 12:23:48 -04:00
Alexis King
b54d8dc04a Strengthen the types for member, memv, memq, and memf 
Fixes #61.
 2015-03-25 16:44:42 -07:00
Vincent St-Amour
b3bdb40ab7 Optimize pair operations, even in the absence of syntactic pair types. 2015-03-24 16:47:06 -04:00
Alexis King
2a4ce892ae Merge branch 'integral-to-inexact' 2015-03-20 19:43:04 -07:00
Alexis King
aab2d9efcf Include more specific types for integers converted to floats 2015-03-19 18:54:11 -07:00
Sam Tobin-Hochstadt
ee7bed777b Add dependency on 6.2.0.2. 
This is definitely needed for the new `begin-encourage-inline`
submodule, but also needed for various other changes (such as
`dynamic-object/c`).
 2015-03-19 21:45:15 -04:00
Sam Tobin-Hochstadt
0dce19e4bf Reduce more dependencies. 
Use `syntax/parse/pre` (checked with @rmculpepper).
Use new `begin-encourage-inline` submodule.
Remove use of `match` in TR runtime-loaded code.
 2015-03-19 21:41:58 -04:00
Sam Tobin-Hochstadt
4709536653 Split prims.rkt to reduce dependencies. 
There are several new files: one for structure prims, one for
annotation-related prims, one for contract related ones. The prims-contract
file uses trickery with submodules to avoid a direct dependency
on `racket/contract`.

Additionally, the rewritten versions of `for/set` et al now use
a submodule to avoid a direct dependency on `racket/set`.
 2015-03-18 18:47:49 -04:00
Sam Tobin-Hochstadt
333a8b9bd7 Remove long lines, and inline vector-copy to remove dependency. 2015-03-18 18:23:56 -04:00
Asumu Takikawa
7a09bac1e3 Handle more cases for opt-arg function contracts 
Allow more cases that are allowed for ordinary function
contracts and explicitly error instead of internal errors
for other cases.

Closes Github Issue #50
 2015-03-17 15:52:21 -04:00
TwoPoint7182818
911036c2a6 Make order-of-magnitude more specific 
(order-of-magnitude n) is only negative if n < 1. Hence it is non-negative for all positive integers.
 2015-03-15 20:48:46 -04:00
Asumu Takikawa
8106d318f1 Fix first-order check for TR's opaque object/c 
Closes PR 15003
 2015-03-13 16:52:34 -04:00
Asumu Takikawa
197cdeac9b Fix contract, remove comment signature 2015-03-12 01:35:54 -04:00
Asumu Takikawa
d19bed0e8b Fix inaccurate data definition comment 2015-03-10 15:18:03 -04:00
Asumu Takikawa
8e6710438f Support prefabs for unsafe-struct-ref 2015-03-10 15:18:02 -04:00
Asumu Takikawa
5d708ab53f Add typechecking for prefab structs 
Adds a new Prefab type constructor along with support for
the use of `struct` with the #:prefab keyword.
 2015-03-10 15:14:51 -04:00
Asumu Takikawa
50f2271917 Fix minor contract violation in REPL messages 2015-03-09 15:59:02 -04:00
Asumu Takikawa
c00cc17e07 Add cond-contracts to name-env functions 
This helped for debugging, especially since some of
the comments were inaccurate. Removed contract comments
that aren't more specific than the code contracts.
 2015-03-09 15:59:02 -04:00
Asumu Takikawa
fb0c96e7f0 Forgot to add this change to the previous commit 
(put requires in the right phase)
 2015-03-09 15:57:22 -04:00
Asumu Takikawa
96e741dc1b Add srcloc instead of a quoted syntax for casts 2015-03-06 16:09:31 -05:00
Sam Tobin-Hochstadt
a3e25231a6 Look up unrenamed versions of type aliases as well. 
Fixes bug reported by Alexander Knauth.
 2015-03-06 12:13:40 -05:00
Sam Tobin-Hochstadt
7a67243d68 Fix contract provide fixup to handle contracts at the beginning. 2015-03-06 09:41:29 -05:00
Matthew Flatt
011649a654 move contract submodules to start of enclosing module 
Putting the contract submodules early enough allows them to be used
for `module*` submodules.

Repairs a problem reported by Alexander:
 http://lists.racket-lang.org/users/archive/2015-March/066164.html
 2015-03-06 09:41:29 -05:00
Alexis King
c79b61a751 Adjust named-let to use the return type annotation whenever possible 
Fixes #43
 2015-03-04 15:17:35 -08:00
Asumu Takikawa
33543ce054 Use absent in some row poly class contracts 
When exporting row polymorphic functions from TR, just
use absent clauses to ensure that TR won't accidentally
try to add pre-existing fields/methods. No sealing is
needed because the typechecker enforces parameteric use
of the class.
 2015-03-04 16:26:35 -05:00
Asumu Takikawa
8d0c352dcc Add a custom object contract for use in TR 
This corresponds to the more strict object contracts
from the OOPSLA paper. Also use `object/c-opaque` in
TR contract generation
 2015-03-04 16:26:35 -05:00
Asumu Takikawa
e2fd3b6653 Enable opaque class contracts in TR 2015-03-04 13:28:06 -05:00
Asumu Takikawa
e80f17cbad Adjust contract generation for some Instance types 2015-03-04 13:27:31 -05:00
Andrew Kent
ea942c2110 fixed base/value overlap bug 2015-03-03 23:20:15 -05:00
Alexis King
3f67cfea7c Use contract-neg-party to accurately handle contracted applications 
Fixes #38
 2015-03-03 16:15:09 -08:00
Andrew Kent
766bd83a98 removed subtle latent bug in tc-let-unit 
there was a values mismatch bug hiding in here that hadn't ever (to my knowledge) been triggered -- perhaps this code is dead?
 2015-03-01 12:27:51 -05:00