14474 commits

This branch

This branch

All branches

Author	SHA1	Message	Date
Folkert de Vries	fc446d307d	changes after code review	2019-11-18 14:55:03 +01:00
Folkert de Vries	1fba0ba879	add note about cargo update	2019-11-18 14:44:10 +01:00
Richard Feldman	7f66c4a10b	Merge pull request #8 from rtfeldman/parse-block-string ... Parse BlockString	2019-11-17 17:39:13 -05:00
Richard Feldman	50a6817bea	Merge pull request #11 from rtfeldman/fix-underscore-arg ... Fix underscore argument parsing	2019-11-17 06:09:49 -05:00
Richard Feldman	b46bd68c24	Add a clarifying comment.	2019-11-16 22:39:37 -05:00
Richard Feldman	942cf8b781	Require space between function argument patterns.	2019-11-16 22:39:31 -05:00
Richard Feldman	60ee53080a	Reproduce malformed argument bug	2019-11-16 22:39:25 -05:00
Richard Feldman	1e8423a368	Merge pull request #2 from rtfeldman/fix-identity ... Generate type var names when pretty printing	2019-11-16 22:38:27 -05:00
Richard Feldman	11acfd8d2a	Fix invalid function arg name in test	2019-11-16 22:30:38 -05:00
Richard Feldman	0b61fcdc25	Clarify some comments.	2019-11-16 22:30:38 -05:00
Richard Feldman	349f060495	Don't generate names that collide with rigids	2019-11-16 22:30:38 -05:00
Richard Feldman	a2cc417a34	Disable unrelated failing tests for now	2019-11-16 22:30:38 -05:00
Richard Feldman	036a72e6e5	Fix generated type var order	2019-11-16 22:30:38 -05:00
Richard Feldman	47fdedc63e	Avoid unnecessary work.	2019-11-16 22:30:38 -05:00
Richard Feldman	c5882fe667	Fix some bugs in type var name gen	2019-11-16 22:30:38 -05:00
Richard Feldman	394f33de8a	Add more function inference tests	2019-11-16 22:30:38 -05:00
Richard Feldman	b26c4393cb	Add more identity inference tests	2019-11-16 22:30:38 -05:00
Richard Feldman	7382dc93ba	Use unwrap_or_else over unwrap	2019-11-16 22:30:38 -05:00
Richard Feldman	377ba81795	Improve Debug display of union find structure	2019-11-16 22:30:38 -05:00
Richard Feldman	75166878e8	Give a more descriptive panic when lookup fails	2019-11-16 22:30:38 -05:00
Richard Feldman	90bc97b6d9	Format parens better in function types	2019-11-16 22:30:38 -05:00
Richard Feldman	c2e966ca3d	Make name_root support multiple letters	2019-11-16 22:30:38 -05:00
Richard Feldman	dd014b7794	Drop obsolete comment	2019-11-16 22:30:38 -05:00
Richard Feldman	90b8db4fbe	Restore always_function test	2019-11-16 22:30:38 -05:00
Richard Feldman	cb47a2f16f	Add anonymous identity function test	2019-11-16 22:30:38 -05:00
Richard Feldman	9ee5fad6ab	Remove unnecessary ref	2019-11-16 22:29:36 -05:00
Richard Feldman	2c3307d4ce	Drop obsolete FlatType::BinOp	2019-11-16 22:29:36 -05:00
Richard Feldman	bf2e75a535	Name all type vars	2019-11-16 22:29:36 -05:00
Richard Feldman	b6f862b7d6	Cleanup	2019-11-16 22:29:36 -05:00
Richard Feldman	ffe088e686	Name some Reason variants better	2019-11-16 22:29:36 -05:00
Richard Feldman	a335632630	Rename identity test to identity_function	2019-11-16 22:29:36 -05:00
Richard Feldman	d82b7815c2	Canonicalize patterns with scope.symbol	2019-11-16 22:29:11 -05:00
Richard Feldman	912fe3b593	Fix pattern constraint gen	2019-11-16 22:29:11 -05:00
Folkert	b13c9295e3	add BlockStrLiteral pattern in the parsing ast	2019-11-16 14:20:31 +01:00
Folkert	2765e3d284	use Roc syntax in comment	2019-11-16 12:48:40 +01:00
Folkert	68eb120da8	desugar if into case ... - I think this runs into the same "cloning disaster" described in other branches - should "False" be qualified? could it ever be locally shadowed? In that case it's `(&["Bool"], "False")`? - is there a default/empty region that is preferable for the `False` and `_` patterns? - desugaring works, but generating constraints for variants (needed for False and _) doesn't yet. so the test would trigger a panic currently - is there a reason for False and _ (True and _ is the more obvious choice)	2019-11-16 12:40:11 +01:00
Richard Feldman	4af0ee0e64	Merge pull request #5 from rtfeldman/desugar-apply ... Desugar pizza into Apply	2019-11-16 00:10:29 +00:00
Richard Feldman	f0945a3e50	Merge pull request #6 from rtfeldman/can-blockstr ... Can blockstr	2019-11-16 00:09:55 +00:00
Folkert	c39e060ddc	Parse BlockString ... a couple of open questions: - shouldn't patterns (for formatting) make the distinction between block and line? - can the `arena.alloc` be removed? the Vec is already allocated by bumpalo - for error reporting, is it better to distinguish block and line?	2019-11-15 18:08:01 +01:00
Folkert	42932367e1	revert accidental push to trunk	2019-11-15 13:13:20 +01:00
Folkert	6d982001e8	add commented test ... desugaring works, but generating constraints for variants (needed for False and _) doesn't yet.	2019-11-15 13:08:17 +01:00
Folkert	b1f2a47a58	desugar if into case ... - I think this runs into the same "cloning disaster" described in other branches - should "False" be qualified? could it ever be locally shadowed? In that case it's `(&["Bool"], "False")`? - is there a default/empty region that is preferable for the `False` and `_` patterns?	2019-11-15 13:03:11 +01:00
Folkert	1ea0264174	changes after review ... trust the borrow checker: reallocation of Expr is not needed	2019-11-15 12:18:06 +01:00
Folkert	a8bb717068	changes after review	2019-11-15 11:56:21 +01:00
Richard Feldman	6c0e771287	Merge pull request #1 from rtfeldman/format-records ... Add formatting for records	2019-11-15 03:34:26 +01:00
Richard Feldman	20240df507	Replace Operator with BinOp and UnaryOp	2019-11-15 02:26:47 +00:00
Richard Feldman	307c2bbd6b	Merge remote-tracking branch 'origin/trunk' into format-records	2019-11-15 02:24:31 +00:00
Richard Feldman	eb0dce2854	Finish multiline record field formatting	2019-11-15 02:13:51 +00:00
Richard Feldman	67d512f5fb	Remove block comments	2019-11-15 02:01:40 +00:00
Richard Feldman	d662fe44ba	Merge branch 'clippy' into trunk	2019-11-15 01:31:15 +00:00