Commit graph

297 commits

Author SHA1 Message Date
b1eefb0f4d move prettyTag to Quox.Pretty 2023-09-20 21:58:42 +02:00
ee22486e97 rename BindName.name to .val 2023-09-20 21:58:27 +02:00
08fb686bf6 move Scoped to separate module 2023-09-20 21:58:04 +02:00
cf3ed604a4 move Quox.Syntax.Var to just Quox.Var 2023-09-20 21:56:59 +02:00
4704dd0441 remove on-hold dir 2023-09-20 21:55:03 +02:00
dc076b636d fix warnings 2023-09-19 18:13:45 +02:00
80b1b3581a use ST from base 2023-09-19 13:05:01 +02:00
ebde478adc add η for pairs in zero contexts 2023-09-19 00:41:17 +02:00
bb8d2464af add fst and snd 2023-09-18 21:53:38 +02:00
e6c06a5c81 pass the subject quantity through equality etc
in preparation for non-linear η laws
2023-09-18 21:53:38 +02:00
3fe9b96f05 make function types with an empty domain subsingletons
this is useful for the base cases of W types when i try those again

closes #23
2023-09-17 20:10:51 +02:00
244b33d786 fix some comments 2023-09-17 19:11:20 +02:00
b85dcb5402 η for box
fixes #27
2023-09-17 19:11:12 +02:00
e1257560b7 Show for contexts, etc 2023-09-17 19:09:54 +02:00
4c88918ade stop throwing names away 2023-09-17 19:08:49 +02:00
8221d71416 some refactors 2023-09-17 14:41:20 +02:00
7b53d56072 a few basic fv tests to make sure it's not reversed or whatever 2023-09-16 13:34:11 +02:00
fa14ce1a02 add FreeVars, and split only on used dvars in Equal 2023-09-12 09:56:49 +02:00
9973f8d07b refactor elim equality error stuff 2023-09-12 06:48:51 +02:00
1e8932690b untangle big mutual block in Equal 2023-08-28 22:07:57 +02:00
d5d30ee198 loosen pushCoe's type slightly 2023-08-28 20:03:06 +02:00
add2eb400c make Elim.compare0 able to pass a type to isSubSing
it now recovers from (most) errors and always returns a type, so that
isSubSing doesn't have to recalculate it

it already assumed the inputs had the same type. now it just leans on
that assumption harder
2023-08-28 20:00:54 +02:00
6f9d31aa0a add displacement to Definition 2023-08-28 19:59:36 +02:00
32f6e5a3b1 make displace total (with a few asserts) 2023-08-28 19:57:02 +02:00
72609bc12f Elim.compare0 infers the type
instead of calling computeElimType over and over. now there's just one
at the start
2023-08-27 19:05:25 +02:00
3e3bf1b67f factor out this case !mode of {..} stuff 2023-08-27 19:04:30 +02:00
2340b14407 partly improve coercions over constant lines
still needs a real quality check, or something, for stuff like
e : (x ≡ x : A) ⊢ coe (𝑖 ⇒ e @𝑖) x
2023-08-27 18:28:08 +02:00
edfe30ff63 update compare0 for type-directed whnf 2023-08-26 21:32:15 +02:00
ba77c45c64 always print the direction in coe/comp 2023-08-26 21:19:40 +02:00
f3f74d581a fix Main 2023-08-26 21:07:10 +02:00
22db2724ce make coercion computation type-directed like it should be 2023-08-26 21:00:19 +02:00
0bcb8c24db make an optional Loc non-optional 2023-08-26 20:59:39 +02:00
a221380d61 more effect stuff, incl. ST 2023-08-25 18:59:54 +02:00
4b6b3853a1 make uses of eff more consistent 2023-08-24 19:55:57 +02:00
8264a1bb81 split up whnf module 2023-08-24 18:42:26 +02:00
a24ebe0702 tycasePi etc don't actually need a scope of (S d) 2023-08-24 17:45:37 +02:00
688204f1a4 make some things private 2023-08-24 17:45:20 +02:00
09e39d6224 add some comments 2023-08-24 17:45:12 +02:00
00d92d3f25 add missing parens in pretty printer 2023-08-12 10:25:07 +02:00
932469a91e make quantities optional and default to 1 2023-07-18 23:12:04 +02:00
349cf2f477 remove unused Tighten impl 2023-07-17 18:10:13 +02:00
b6264f388d fix #11 the easy way
tightening just pushes substitutions all the way through. bleh
2023-07-17 03:50:16 +02:00
612fb33663 bump again 2023-07-13 21:28:39 +02:00
fa09aaf228 squash warnings 2023-06-23 18:32:05 +02:00
f0d3529f63 fix subtype stuff for Eq 2023-06-22 22:20:12 +02:00
cd330c1092 remove a noLoc 2023-06-11 19:25:38 +02:00
865772d512 remove stale todos 2023-06-11 19:25:32 +02:00
00e79d4264 quote names in Show 2023-05-25 18:34:13 +02:00
3bbf0366c8 make 0 in ★₀ optional 2023-05-21 20:34:05 +02:00
282565c7a3 Whnf ⇒ CanWhnf; WhnfM ⇒ Eff Whnf 2023-05-21 20:34:05 +02:00
2af8ee20ea those were not meant to stay there 2023-05-21 20:34:05 +02:00
42aa07c9c8 crude but effective stratification 2023-05-21 20:34:05 +02:00
e4a20cc632 remove redundancy in equality check 2023-05-20 21:38:23 +02:00
64de93a13c remove square brackets around type lines
(parens are needed if they are anything other than a `term arg`)
2023-05-16 18:14:42 +02:00
d631b86be3 make p,q in coe/comp optional and default to @0 @1 2023-05-15 20:06:40 +02:00
7b93a913c7 rewrite pretty printer 2023-05-15 17:13:14 +02:00
f6abf084b3 qty lub is total actually (usually ω) 2023-05-12 17:28:29 +02:00
8d6ae6cc32 move location to the start of type errors 2023-05-02 19:03:05 +02:00
d5f4a012c5 add source locations to inner syntax 2023-05-02 03:06:25 +02:00
30fa93ab4e refactor core syntax slightly to derive Eq/Show
add a new `WithSubst tm env to` record that packages a `tm from`
with a `Subst env from to`, and write instances for just that. the
rest of the AST can be derived
2023-04-27 21:37:20 +02:00
7e079a9668 add file locations to Parser.Syntax
they're immediately thrown away currently. but one step at a time
2023-04-26 06:12:03 +02:00
b5f42cde64 remove big mutual blocks in parser 2023-04-26 02:28:08 +02:00
adebfe090c error message tweaks 2023-04-24 23:19:15 +02:00
b74ffa0077 rewrite parser
previously it backtracked too much, so instead of giving a useful
parse error, it just said "expected end of input" at the beginning of
the problem toplevel. which, if it's a namespace, could be way off.
2023-04-24 22:25:04 +02:00
a4ffd74625 fix the most embarrassing quantity mistake ever 2023-04-20 19:29:57 +02:00
3f06e8d68b allow multiple names in a binder
e.g. "(x y : ℕ) × plus x y ≡ 10 : ℕ"

fixes #2
2023-04-19 21:37:51 +02:00
b666bc20cf do scope checking in FromParser where it belongs 2023-04-18 22:55:23 +02:00
55c0bf9974 fix skipping files 2023-04-18 00:10:53 +02:00
4578b30c79 namespaces work now 2023-04-18 00:10:53 +02:00
4db373a84f use constraints when checking δ
when checking δ 𝑖 ⇒ s, add 𝑖=ε to Ψ instead of checking s‹ε/𝑖›.
this has the same effect but an error message will show "𝑖, 𝑖=ε" in
the context
2023-04-17 22:02:40 +02:00
ac85dc9352 don't put a ∷ℕ on nat literals that's silly 2023-04-17 22:02:40 +02:00
06b159973f pretty printing fixes 2023-04-17 21:45:05 +02:00
682965eebd 0∨1 is not undefined it's ω 2023-04-17 21:42:33 +02:00
6428d39ce1 semicolons between decls in namespaces 2023-04-17 21:41:00 +02:00
3fb8580f85 re-add tightening and use it when messing with scopes
e.g. "coe [_ ⇒ A] @p @q s" should immediately reduce to "s",
but if the "_ ⇒ A" happened to use an SY it didn't.

this will still happen if a wrong SY sneaks in but the alternative is
re-traversing the term over and over every time whnf runs
2023-04-17 20:56:31 +02:00
a5ccf0215a coercions and compositions 2023-04-15 15:13:01 +02:00
a42e82c355 type-case 2023-04-03 17:46:23 +02:00
1211272420 factor out some pretty printing stuff 2023-04-02 15:52:55 +02:00
e1dbf272df rename EqualE to just Equal & add runEqual 2023-04-02 15:52:55 +02:00
38dbd275a1 add case0 to syntax since that is possible sometimes 2023-04-02 15:52:55 +02:00
ba2818a865 remove IsQty interface 2023-04-01 19:16:43 +02:00
924fd991f9 fix equality types 2023-04-01 16:02:02 +02:00
15f6f4c8a4 fix nat elim quantities 2023-04-01 16:01:53 +02:00
036e2bd4a5 fix case-box typing 2023-04-01 16:01:31 +02:00
1fce4d80f6 add box patterns to the parser oops 2023-04-01 15:59:16 +02:00
a17752f31c fix up tests 2023-03-31 23:43:25 +02:00
5e220da2f4 a half-implemented verified dimeq 2023-03-31 19:34:24 +02:00
1ab0e42605 print a slightly better error if 'case' has no qty 2023-03-31 19:33:40 +02:00
c8fbd73ea4 use names when pretty printing contexts 2023-03-31 19:30:55 +02:00
ad942b2fd8 printing for most of FromParserError 2023-03-31 19:29:15 +02:00
2b2f79fca9 fix some typing mistakes 2023-03-31 19:26:55 +02:00
36609713ac mtl ⇒ eff 2023-03-31 19:26:24 +02:00
8a9b4c23dd box type 2023-03-31 19:26:24 +02:00
7d36a7ff54 allow matching at 0 where appropriate
(for pairs, and for enums with 0 or 1 constructors)
2023-03-27 00:08:09 +02:00
137962c176 add missing %default total 2023-03-27 00:07:39 +02:00
773f6372ea quantities in case don't need to be *exactly* the same
...as long as they are all compatible with the target.
for example, given ω.n : ℕ:
```
  case double_it? return ℕ of {
    'true  ⇒ plus n n;
    'false ⇒ n
  }
```
2023-03-27 00:01:32 +02:00
f620dda639 fix error message 2023-03-26 16:15:30 +02:00
5df2a4538c more tests 2023-03-26 16:15:19 +02:00
5560cb6708 move 'enum' to Syntax.Base 2023-03-26 16:14:58 +02:00
e6c4203b46 print ; between case branches 2023-03-26 16:13:36 +02:00