suzanne.soy/macrotypes
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
266 Commits 11 Branches 0 Tags 2.8 MiB
6cb15a06da
Commit Graph

189 Commits

Author SHA1 Message Date
Stephen Chang
5c5f500f39 clean up requires/provides, see other mlish language additions below 
- fix bug where recursive tyvar did not have #%type tag
- records
- optional field names in type definitions
- reuse prefixes by default
- list fns
- provide explicit inst (need it for ho poly fns)
- if now allows non-false instead of bool
- small inference fixes
- add rw ocaml tests, ch6, up to "polymorphic variants"
 2016-03-10 00:39:01 -05:00
Stephen Chang
dc3767c844 mlish: add nbody test 2016-03-08 16:25:02 -05:00
Stephen Chang
159bd56bf3 add more iterators and matrix tests 2016-03-08 14:50:37 -05:00
Stephen Chang
2066dbc577 mlish: add require/provide, boxes, set!, more iterators 
- add tests: fibo, hash, k-nucleotide
 2016-03-07 23:54:47 -05:00
Stephen Chang
f5a043b7e6 mlish: add hash, fl prims, str prims, output, and some other prims 
- allow begin exprs to have any type (not just unit)
- add fasta test
 2016-03-07 00:31:25 -05:00
Stephen Chang
ee6c0c11f7 mlish: add for/list, when, printf, multi expr bodies 
- add fannkuch test
 2016-03-05 10:12:32 -05:00
Stephen Chang
468cf075d6 mlish: add vectors, sequences, and iteration 
- add ary and ack test
- fix bug in uparrow-s (in typecheck.rkt) extra #'
 2016-03-04 16:15:14 -05:00
Stephen Chang
1ad357b55e separate out mlish queens test 2016-03-04 14:20:02 -05:00
Stephen Chang
965778d9f0 add mlish tests 2016-03-04 14:15:28 -05:00
Stephen Chang
cd6c8920ab mlish: add concurrency prims and test 2016-03-03 23:38:21 -05:00
Stephen Chang
8b11b0fb60 fix missing clauses check; code cleanup 2016-03-02 17:08:47 -05:00
Stephen Chang
487c8fedc5 mlish: fix define-type and match 
- resolves #1
  - define-type-constructor accepts arbitrary extra "info" that is stored
    with the type
    - get-extra-info extras this info
  - define type stores extra info that is essentially a mu type
    containing all variant info
  - match must manually "unfold" this info, ie iso-recursive
- all tests passing
- mlish tests implements full nqueens example
  - no explicit instantiation of polymorphic functions
  - some polymorphic constructors require annotations when no inference possible
    - a lone nil with no other information
  - lambda params require annotations
  - top lvl fns require full signature
 2016-03-02 15:34:51 -05:00
Stephen Chang
3839ea51c4 some code cleanup 2016-02-29 15:36:57 -05:00
Stephen Chang
ab9f96efe4 fix match guards and ordering 2016-02-29 15:33:22 -05:00
Stephen Chang
f0aba48497 - add folding compuate+instantiate-or-inferral of args in app 
- start guard support in match (doesnt work yet)
 2016-02-29 14:21:02 -05:00
Stephen Chang
87cf55e7ae fix mlish match 
- properly propagate expected-type
- need cond that checks all cases at runtime
 - need to compute accessors and predicate
- add tests
 2016-02-29 00:18:54 -05:00
Stephen Chang
a8d461ea0d fix more app err msgs in tests; all tests passing 2016-02-27 17:31:20 -05:00
Stephen Chang
9000e126ff fix mlish tests to use new app err msg abstractions 2016-02-27 16:26:15 -05:00
Stephen Chang
d880b9d668 clean up app err msgs; need to fix tests 2016-02-26 16:54:12 -05:00
Stephen Chang
893d457bad add top lvl defines and infer instantation 
- need to err msg when instantiation fails
 2016-02-25 17:02:46 -05:00
Stephen Chang
6fa962f83f all constructor type instantiations (in tests) inferred 
- id type names and variant names (for 0arg constructors) supported
 2016-02-24 19:02:51 -05:00
Stephen Chang
dbfd52bf0f start mlish lang 2016-02-22 19:23:14 -05:00
Stephen Chang
a82a833b43 add stlc+effect.rkt 2016-01-12 15:29:56 -05:00
Stephen Chang
b6c8bb5880 infer some lambda param types from body 2015-12-21 18:55:42 -05:00
Stephen Chang
aa930c1180 add more infer examples 2015-12-17 17:05:25 -05:00
Stephen Chang
75da160c2d finish nqueens example for infer.rkt 2015-12-17 16:16:09 -05:00
Stephen Chang
fbc5934675 implement list functions using infer.rkt: annotations still required for top lvl fns and unapplied lams 2015-12-16 19:39:18 -05:00
Stephen Chang
a41a4cdd62 start exploring (local) inference 2015-12-11 17:01:25 -05:00
Stephen Chang
7999f1cc8c stlc+cons: inference testing 2015-12-09 17:30:11 -05:00
Stephen Chang
6f7edcc431 use string-prefix from racket/string 2015-10-27 16:29:22 -04:00
Stephen Chang
b52f220dcc occurrence: use tups instead of records 2015-10-22 17:12:51 -04:00
Stephen Chang
4152db6575 add occurrence and overloading to run-all-tests 2015-10-22 17:12:11 -04:00
stchang
12d73ee2a9 Merged in overload (pull request #6) 
Overload -- initial draft
 2015-10-22 17:07:38 -04:00
Ben Greenman
51e6e1b3ae [overload] remove design/ folder 2015-10-22 17:07:05 -04:00
stchang
67d28645f9 Merged in o+ (pull request #4) 
Occurence types, improved
 2015-10-22 17:05:05 -04:00
Ben Greenman
e038d220f1 [overload] now improved with identifier macros 2015-10-22 15:20:16 -04:00
Stephen Chang
f7635af9b5 add stlc fail tests 2015-10-22 13:22:35 -04:00
Stephen Chang
b4bc922dda add stlc+sub tests 2015-10-22 13:22:22 -04:00
Ben Greenman
c299e58a04 [overload] fix export-as, export 'resolve', fix subtyping 2015-10-22 05:44:17 -04:00
Ben Greenman
0b587697f4 [overloading] initial revision, overloading via parameter 2015-10-21 00:47:21 -04:00
Ben Greenman
e6e538c3ed [overload] doc'd the problem with lambda, gosh I miss let 2015-10-20 18:06:59 -04:00
Ben Greenman
b3e5a1b5cf [overload] add run-time desugar to design doc 2015-10-20 16:42:13 -04:00
Ben Greenman
3cbafd3772 [overload] got \psi-types and signature init 2015-10-20 00:49:55 -04:00
Ben Greenman
f228ace20f [overload] outline tests 2015-10-20 00:49:23 -04:00
Ben Greenman
e9cefe3298 [overload] design doc for \psi types 2015-10-19 20:44:39 -04:00
Ben Greenman
3877a28a75 [o+] some examples from ICFP'10 paper 2015-10-19 13:53:15 -04:00
Ben Greenman
d1a18786ac [occurrence] filters for lists 2015-10-17 15:13:08 -04:00
Ben Greenman
1887d99abb [occurrence] filters for products 2015-10-16 18:56:54 -04:00
Ben Greenman
36bec40650 regression tests for subs? 2015-10-15 15:37:48 -04:00
Ben Greenman
36d568b512 [o+] allow unions as filters 2015-10-15 15:26:27 -04:00
Ben Greenman
dfaab96a29 [o+] check for same-arity functions in unions 2015-10-15 14:31:31 -04:00
Ben Greenman
5c8300a538 [o+] check & filter functions (by arity) 2015-10-15 14:08:22 -04:00
Ben Greenman
e2b1eaa06b bugfix in stlc+sub: subs assumed length=? 2015-10-15 14:00:42 -04:00
Stephen Chang
48b625c2f4 define star as rename-transformer to #%type, in fomega 2015-10-14 15:39:11 -04:00
Stephen Chang
4dd2b66d2b generalize fomega ty= to avoid specific types; add tests 
- fixes fomega bug where kind annotations were not being compared
- fix subst bug: must tranfer props (ie types) when replacing an id
- add exist tests to check proper subst in ty=
 2015-10-14 14:57:27 -04:00
Stephen Chang
a07fa92d25 clean up fomega2/3, remove many type-eval 
- remove expand-fn optional arg in infer fn
 2015-10-14 14:57:27 -04:00
stchang
94b610c93b Merged in occurrence (pull request #3) 
Occurrence Types: first draft
 2015-10-13 21:08:05 -04:00
Ben Greenman
bc2e83c449 [occurence] simple filters complete 2015-10-13 15:16:24 -04:00
Ben Greenman
d7913f7753 [occurrence] subtyping, normal form, and VERY BASIC filters. Having trouble propogating variables. 2015-10-13 12:51:24 -04:00
Ben Greenman
c5d6bfbf9e [occurrence] type eval 2015-10-11 00:22:16 -04:00
Stephen Chang
bf126449b4 define-type-constructor supports annotated bound vars 
- use this to impl forall in fomega
- define-type-constructor default arity is =1
- define explicit normalize fn in fomega -- speeds up tests 15%
- move same-type to define-syntax-category
 2015-10-09 16:59:48 -04:00
Ben Greenman
5497b347df [occurrence] simple type evaluator, assumes no recursive unions 2015-10-09 02:24:09 -04:00
Stephen Chang
dad2c26c49 add existential tests demonstrating hidden type 2015-10-08 13:22:26 -04:00
Stephen Chang
33996e71ac rec-iso, reduce ty=? to just focus on lamda 2015-10-08 13:21:54 -04:00
Ben Greenman
050c43b3a2 [occurrence] constructor + basic tests 2015-10-06 16:32:26 -04:00
Stephen Chang
0aa2857e74 define-syntax-category defines default type=?; other cleanup 
- extended type=? in rec-iso only does 1 subst
- simplify current-type? in fomega by defining in terms of current-kind?
 2015-10-06 14:28:08 -04:00
Stephen Chang
2278adcf95 replace provide/requires with "extends...except"/"reuse...from" forms 
- introduce basic define-typed-syntax instead of define-syntax
- add fomega3 --- same as fomega2, without the duplication
  - app and lam are both types and terms
  - arrow is both type and kind
  - same tests as fomega2
 2015-10-01 16:21:29 -04:00
Stephen Chang
cb7a49ab06 use #lang typecheck 2015-09-23 17:01:13 -04:00
AlexKnauth
149897653a use make-variable-like-transformer 2015-09-21 18:09:58 -04:00
AlexKnauth
82b7ec004d add version of make-variable-like-transformer 2015-09-21 17:54:25 -04:00
Stephen Chang
fc2150dc0d fix subtyping bug in if; add current-join 2015-09-16 15:04:48 -04:00
Stephen Chang
3ea9fc5db0 rec-iso: refactor requires so stlc+tup is base lang 2015-09-16 15:04:13 -04:00
Stephen Chang
a3b9cbc712 fix define-type-constructor to properly error if used at runtime 2015-09-10 14:29:23 -04:00
Stephen Chang
d1814cb57a code cleanup; add define-sub-relation 
- all tests passing
 2015-08-28 16:57:38 -04:00
Stephen Chang
0255003ff7 all tests passing with define-syntax-category (except some err msgs in fomega2) 2015-08-27 19:30:46 -04:00
Stephen Chang
d9e1c2febb fomega working with define-syntax-category 2015-08-27 17:23:03 -04:00
Stephen Chang
efaef8b60e progress commit: stuck on fomega+define-syntax-category 2015-08-21 19:28:36 -04:00
Stephen Chang
336b6ea1f3 fix bugs in fomega2 2015-08-20 14:49:33 -04:00
Stephen Chang
6feff7d3b3 completed: type valiation; new type constructor pattern matching 2015-08-19 19:29:07 -04:00
Stephen Chang
1a831454f4 progress commit: use typechecking system to typecheck types 
- ported (tests passing) up to stlc+sub.rkt
 2015-08-14 19:37:09 -04:00
Stephen Chang
3d4fe8f71a new define-type-constructor: exist-tests.rkt passing 2015-08-07 18:12:47 -04:00
Stephen Chang
0dde065949 new define-type-constructor: support binding forms 
- tests passing up to stlc+rec-iso.rkt
 2015-08-07 17:26:53 -04:00
Stephen Chang
75b4dae95f use new define-type-constructor and #%type in stlc+box 2015-08-04 18:24:33 -04:00
Stephen Chang
6e91837992 use new define-type-constructor and #%type in stlc+cons 2015-08-04 18:21:02 -04:00
Stephen Chang
c151463a6f use #%type in stlc+reco+var.rkt; tests passing 2015-08-04 17:31:36 -04:00
Stephen Chang
a12e85d4bc fix type constructor pattern matching and extraction 
- tests passing up to stlc+reco+var tup and proj (not but variants)
 2015-07-31 17:33:45 -04:00
Stephen Chang
b687cffb0a extend define-tycon with "get"- extract matched type by pat var name 2015-07-30 14:12:34 -04:00
Stephen Chang
9973e1c705 start require "extends" macro in typecheck.rkt 2015-07-29 14:27:24 -04:00
Stephen Chang
8d6eaf227e ext-stlc-tests.rkt passing 2015-07-28 19:32:16 -04:00
Stephen Chang
655c4acd5a cleanup typecheck.rkt 2015-07-28 19:14:27 -04:00
Stephen Chang
b958ee6947 add #%plain-type; cleanup 
- see notes.txt for details on #%plain-type
- convert \vdash to meta2 macro instead of function
  - cleaner syntax -- no #', use colon
- fix some src locs
- up stlc+lit.rkt working, everything else broken
 2015-07-28 17:49:29 -04:00
Stephen Chang
a44d946ffb add #%type annotation; add type->str; type constructor pattern match 
- #%type annotation distinguishes types from terms
- is-type? checks #%type
- type->str fixes type printing
  - wasnt recursively calling get-orig
- use delay in "type" syntax-class norm attribute
  - fixes eager eval problem from before
- add more invalid type tests
  - only tested up to stlc+lit.rkt
- nicer type constructors
  - more concise, intuitive pat matching language
  - use for-meta 2 macros
 2015-07-28 15:25:34 -04:00
Stephen Chang
daecc5c624 add tycon pattern matching to stlc 
- fix stlc+lit tests
- current-promote is currently broken
- add err msg checking unit rackunit-typechecking
- works with new expander
 2015-07-24 16:26:04 -04:00
Stephen Chang
0e1a848465 add fsub test; typecheck? needs to call current-promote 2015-07-20 18:03:33 -04:00
Stephen Chang
d0c61a5dc0 add current-promote; fsub avoids extending app and proj 2015-07-10 18:14:33 -04:00
Stephen Chang
5435021ee6 complete fsub 2015-07-10 15:43:08 -04:00
Stephen Chang
03453362b5 start fsub 2015-07-09 15:37:50 -04:00
Stephen Chang
176b48208d cleanup 2015-07-09 15:34:36 -04:00
Stephen Chang
f0d314b46d consolidate typecheck.rkt: impl everything in terms of infer fn 2015-07-09 15:30:12 -04:00
Stephen Chang
b0a8d192ef finish exist.rkt 2015-07-09 14:00:36 -04:00