git.iko.soy/elk
on 577fc17 help sign in
history ofcompiler/Type/Unify.hs
all commits
eaea367Get rid of debugging print that snuck inEvan Czaplicki 0dc89bfcosmetic: make actuallyUnify less monolithicEvan Czaplicki a12554fcosmetic: try to make -Wall happierEvan Czaplicki b16e876Add "alias" as a field in tvars and TermN'sEvan Czaplicki eae58a7Use canonical names during type inferenceEvan Czaplicki 3902049Merge branch 'master' into canonicalizeEvan Czaplicki 7862d20resolve #538, correct error for appendableEvan Czaplicki a88ed26Switch imports from SourceSyntax to ASTEvan Czaplicki 9dd5dffMake AST more general and try to give its phases better namesEvan Czaplicki c966776Add -W flag to many files, mostly getting rid of dead importsEvan Czaplicki 3686c82Add note about no sharing between top-level and let-bound type variablesEvan Czaplicki 2874797Go to original error style. I think it's nicer.Evan Czaplicki 6ff938cImprove message for errors when unifying rigid type variablesEvan Czaplicki f73a6ffAdd an occurs check to finally resolve #294Evan Czaplicki dce5e3eAdd String as an appendable typeEvan Czaplicki 76b2e56Add Strings as a primitive in the compilerEvan Czaplicki 1609e61Fix accidental creation of rigid variables when type-checking mixed record types when a type alias is givenEvan Czaplicki 47aa48dPreemptively add "number" as a kind of flexible variable that can unify with an (Is Number)Evan Czaplicki 3b27395Fix issue unifying comparables brought up by Max on the listsEvan Czaplicki 53ef6d3Give a better error message when trying to unify rigid type variables in a bad wayEvan Czaplicki 20e2ff0Properly handle the qualified `Text` typeEvan Czaplicki 8006d8eReport line numbers in type errorsEvan Czaplicki 3065f58Support appendable typesEvan Czaplicki a1b7336Make comparable work fully and clean up codeEvan Czaplicki fd63affCorrect the code that unifies recordsEvan Czaplicki c1d5f86Simplify type error messagesEvan Czaplicki 42065ebGenerate fewer intermediate variables during record unificationEvan Czaplicki 2a51470Implement unification for recordsEvan Czaplicki 9fe6c9aChange functions for creating type variables, distinguish names of different constraint type variablesEvan Czaplicki 609cc3dUnify any constrained variables with the same super-typeEvan Czaplicki b6144aeProperly unify constrained variables, serious changes here!!!Evan Czaplicki 4e7ef5fAdd definitions for functions that act on numbers and comparablesEvan Czaplicki 82ca695Add support for "number" type which can be either an Int or FloatEvan Czaplicki 4208b5cRemove some debugging stuff and generally clean up the code. There was one potential error in which a debug needed to be a liftIO, the computation it performed was not optional.Evan Czaplicki b5887f4Get type inference working in the basic case. Begin working on printing errors in a prettier way.Evan Czaplicki 497d478Get the solver working on basic programs. It outputs pretty types for the variables in the program. Need to test further and start doing some benchmarking.Evan Czaplicki ca62ee6Create type pools, have an organized model of state to flow through the State Transformer during constraint solving and variable unification.Evan Czaplicki 4c51159Create a Type.Pool module specifically for variable pools. Start using a state transformer to thread the pool and errors through.Evan Czaplicki ca8441fContinue getting the new type-checker in order.Evan Czaplicki