summaryrefslogtreecommitdiff
path: root/dhall/src/semantics/nze (unfollow)
Commit message (Collapse)AuthorFilesLines
2021-01-22chore: fix clippy warningsNadrieril1-4/+4
2020-12-07Resolve imports and alternatives outside of the ast traversalNadrieril1-0/+8
2020-12-07Defer name errors to typecheckingNadrieril1-0/+1
We aren't supposed to inspect anything before alternatives are chosen
2020-12-07Pass import results via the global contextNadrieril1-1/+4
2020-12-07Thread cx everywhere else imports are readNadrieril1-2/+2
2020-12-07Thread cx through normalizationNadrieril3-133/+145
2020-11-03Remove NirInternalNadrieril2-50/+23
2020-11-03Normalize `with` by mutation.Nadrieril3-5/+42
This is Cow-style mutation: we clone only what's shared and then mutate it. This it more legible and more efficient than the immutable version.
2020-11-03Try to keep ownership of `Nir`s when convenientNadrieril1-1/+1
The goal is that we might avoid cloning internals in the common case where a value is not shared
2020-04-06Move BinOp and Builtin definitions in the relevant moduleNadrieril1-4/+3
2020-04-06Move builtins module up a levelNadrieril1-2/+3
2020-04-06Extract operation-related code to a new moduleNadrieril2-295/+9
2020-04-06Only need to store OpKind in NirNadrieril2-82/+77
2020-04-06Fix normalization bugNadrieril1-49/+49
2020-04-06Avoid a few allocationsNadrieril2-16/+20
2020-04-06Factor our operations in normalizationNadrieril2-300/+257
2020-04-06Split off operations from main expr enumNadrieril2-100/+128
2020-03-31Brutally make all of dhall pubNadrieril5-35/+35
2020-03-31Nir::normalize isn't usefulNadrieril1-76/+0
It pretends to normalize but actually can't normalize under lambdas. The correct way to normalize a Nir is to convert it to Hir.
2020-03-31Improve handling of builtin types in NirNadrieril2-25/+32
2020-03-31Move conversion to SimpleVal/Ty to simple moduleNadrieril1-81/+0
2020-03-31Remove top-level Expr aliasesNadrieril1-5/+5
2020-03-31Rename LitKind to NumKindNadrieril2-40/+40
2020-03-31Add SimpleType to mirror SimpleValueNadrieril1-1/+47
2020-03-31Add SimpleValue type to facilitate deserializationNadrieril1-0/+34
2020-03-17Remove never_type featureNadrieril1-13/+14
2020-03-17Run clippyNadrieril3-28/+29
2020-03-05Implement some normalization simplificationsNadrieril1-76/+80
2020-03-02Normalization for ProjectionByExprBasile Henry1-3/+34
2020-02-19Avoid re-typechecking after importNadrieril1-0/+1
2020-02-18Remove useless `normalize` option from ToExprOptionsNadrieril1-3/+0
2020-02-17Remove spans from NirNadrieril1-15/+7
That info was too unreliable: we need clean tracking mechanisms if we want to improve error messages
2020-02-17Rename Value to NirNadrieril5-257/+239
2020-02-16Avoid recomputing universes in tckNadrieril1-3/+3
2020-02-16Start requiring Universe to build a TypeNadrieril1-1/+1
2020-02-15Introduce proper Type structNadrieril1-1/+4
2020-02-15Desugar Completion during resolutionNadrieril1-11/+9
2020-02-15Factor some tck code to avoid needing get_type_tyexprNadrieril2-3/+3
2020-02-13Simplify conversions to/from TyExprNadrieril1-17/+3
2020-02-11Introduce LitKind to factor out common enum nodesNadrieril2-57/+50
2020-02-11Remove the Embed variant from ExprKindNadrieril2-15/+7
2020-02-10Remove need for EmbedNadrieril1-0/+6
This was an archaic leftover from copying the Haskell datatypes anyway
2020-02-09Remove types from NzEnvNadrieril3-30/+32
2020-02-09Remove all types from ValueNadrieril2-74/+9
2020-02-09Take Hir for typecheckNadrieril1-2/+5
2020-02-09Remove extra types stored in ValueNadrieril2-25/+14
2020-02-09Remove most TyExpr from normalizationNadrieril2-76/+65
2020-02-09Add Hir as untyped alternative to TyExprNadrieril2-7/+7
2020-02-09Remove move type propagation through ValueNadrieril2-154/+54
2020-02-09Remove most reliance on types stored in ValueNadrieril3-38/+53