[Git][ghc/ghc][wip/T25281] Comments

[Simon Peyton Jones (@simonpj)]

Simon Peyton Jones pushed to branch wip/T25281 at Glasgow Haskell Compiler / GHC


Commits:
8156fe53 by Simon Peyton Jones at 2024-09-26T15:11:27+01:00
Comments

- - - - -


3 changed files:

- compiler/GHC/HsToCore/Expr.hs
- compiler/GHC/HsToCore/Pmc.hs
- compiler/GHC/Tc/Instance/Class.hs


Changes:

=====================================
compiler/GHC/HsToCore/Expr.hs
=====================================
@@ -558,6 +558,20 @@ dsExpr (SectionR x _ _)   = dataConCantHappen x
 *                                                                      *
 ********************************************************************* -}
 
+{- Note [Desugaring applications]
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+When we come across an application (f e1 .. en) we collect up
+all the desugared arguments, and then dispatch on the function f.
+(Including the nullary case where n=0.)
+
+There are several special cases to handle
+
+* HsRecSel: a record selector gets warnings if it might fail.
+* HsVar:    special magic for `noinline`
+* HsVar:    special magic for `seq`
+
+-}
+
 dsApp :: HsExpr GhcTc -> DsM CoreExpr
 dsApp e = ds_app e [] []
 
@@ -621,8 +635,8 @@ ds_app (HsRecSel _ fld_occ@(FieldOcc { foExt = fun_id })) hs_args core_args
          diagnosticDs $ DsIncompleteRecordSelector (idName fun_id) cons_trimmed
                                             (cons_trimmed /= cons_wo_field)
 
-       -- Type-based check
-       --   ToDo: explain
+       -- Type-based check.  See GHC.HsToCore.Pmc (IRS4)
+       -- in Note [Detecting incomplete record selectors]
        ; case filterOut isTypeArg core_args of
            (arg:_) -> pmcRecSel fun_id arg
            _       -> return ()


=====================================
compiler/GHC/HsToCore/Pmc.hs
=====================================
@@ -198,9 +198,8 @@ pmcMatches origin ctxt vars matches = {-# SCC "pmcMatches" #-} do
         formatReportWarnings ReportMatchGroup ctxt vars result)
       return (NE.toList (ldiMatchGroup (cr_ret result)))
 
-{-
-Note [Detecting incomplete record selectors]
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+{- Note [Detecting incomplete record selectors]
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 A record selector occurrence is incomplete iff. it could fail due to
 being applied to a data type constructor not present for this record field.
 
@@ -211,16 +210,16 @@ e.g.
 There are 4 parts to detecting and warning about
 incomplete record selectors to consider:
 
-  - Computing which constructors a general application of a record field will succeed on,
-    and which ones it will fail on. This is stored in the `sel_cons` field of
-    `IdDetails` datatype, which is a part of an `Id` and calculated when renaming a
-    record selector in `mkOneRecordSelector`
+(IRS1) Computing which constructors a general application of a record field will
+    succeed on, and which ones it will fail on. This is stored in the `sel_cons`
+    field of `IdDetails` datatype, which is a part of an `Id` and calculated
+    when renaming a record selector in `mkOneRecordSelector`
 
-  - Emitting a warning whenever a `HasField` constraint is solved.
+(IRS2) Emitting a warning whenever a `HasField` constraint is solved.
     This is checked in `matchHasField` and emitted only for when
     the constraint is resolved with an implicit instance rather than a
     custom one (since otherwise the warning will be emitted in
-      the custom implementation anyways)
+    the custom implementation anyways)
 
     e.g.
       g :: HasField "x" t Int => t -> Int
@@ -229,7 +228,7 @@ incomplete record selectors to consider:
       f :: T -> Int
       f = g -- warning will be emitted here
 
-  - Emitting a warning for a general occurrence of the record selector
+(IRS3) Emitting a warning for a general occurrence of the record selector
     This is done during the renaming of a `HsRecSel` expression in `dsExpr`
     and simply pulls the information about incompleteness from the `Id`
 
@@ -237,7 +236,7 @@ incomplete record selectors to consider:
       l :: T -> Int
       l a = x a -- warning will be emitted here
 
-  - Emitting a warning for a record selector `sel` applied to a variable `y`.
+(IRS4) Emitting a warning for a record selector `sel` applied to a variable `y`.
     In that case we want to use the long-distance information from the
     pattern match checker to rule out impossible constructors
     (See Note [Long-distance information]). We first add constraints to
@@ -255,6 +254,7 @@ incomplete record selectors to consider:
 pmcRecSel :: Id       -- ^ Id of the selector
           -> CoreExpr -- ^ Core expression of the argument to the selector
           -> DsM ()
+-- See (IRS4) in Note [Detecting incomplete record selectors]
 pmcRecSel sel_id arg
   | RecSelId{ sel_cons = (cons_w_field, _ : _) } <- idDetails sel_id = do
       !missing <- getLdiNablas


=====================================
compiler/GHC/Tc/Instance/Class.hs
=====================================
@@ -1298,7 +1298,8 @@ matchHasField dflags short_cut clas tys
                              ; unless (null $ snd $ sel_cons $ idDetails sel_id)
                                  $ addDiagnostic $ TcRnHasFieldResolvedIncomplete name
                                  -- Only emit an incomplete selector warning if it's an implicit instance
-                                 -- See Note [Detecting incomplete record selectors] in GHC.HsToCore.Pmc
+                                 -- See (IRS2) in Note [Detecting incomplete record selectors]
+                                 -- in GHC.HsToCore.Pmc
                              ; return OneInst { cir_new_theta   = theta
                                               , cir_mk_ev       = mk_ev
                                               , cir_canonical   = EvCanonical



View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/8156fe535b2318c6720ed9260b209756369309f5

-- 
View it on GitLab: https://gitlab.haskell.org/ghc/ghc/-/commit/8156fe535b2318c6720ed9260b209756369309f5
You're receiving this email because of your account on gitlab.haskell.org.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-commits/attachments/20240926/1f133cea/attachment-0001.html>