From e324a09fb83501830c0678eb8a3194125250c3c3 Mon Sep 17 00:00:00 2001
From: Pierre Le Marre <dev@wismill.eu>
Date: Mon, 17 Jun 2024 16:59:02 +0200
Subject: [PATCH] core: Improve Case & Unfold

---
 unicode-data/lib/Unicode/Char/Case.hs        |  5 ++++-
 unicode-data/lib/Unicode/Char/Case/Compat.hs | 11 ++++++-----
 unicode-data/lib/Unicode/Internal/Unfold.hs  | 15 +++++++++------
 3 files changed, 19 insertions(+), 12 deletions(-)

diff --git a/unicode-data/lib/Unicode/Char/Case.hs b/unicode-data/lib/Unicode/Char/Case.hs
index c23b3719..3580fb52 100644
--- a/unicode-data/lib/Unicode/Char/Case.hs
+++ b/unicode-data/lib/Unicode/Char/Case.hs
@@ -249,4 +249,7 @@ step = \case
         where
             -- Mask for a single Unicode code point: (1 << 21) - 1
             mask = 0x1fffff
-            cp = fromIntegral (s .&. mask)
+            -- [NOTE] As of GHC 9.4, Int64 is represented internally by Int64#,
+            -- so the previous code `fromIntegral s .&. mask` leads to
+            -- unefficient generated code.
+            cp = fromIntegral s .&. mask
diff --git a/unicode-data/lib/Unicode/Char/Case/Compat.hs b/unicode-data/lib/Unicode/Char/Case/Compat.hs
index f01fa92e..0801dcc2 100644
--- a/unicode-data/lib/Unicode/Char/Case/Compat.hs
+++ b/unicode-data/lib/Unicode/Char/Case/Compat.hs
@@ -31,8 +31,9 @@ import qualified Unicode.Internal.Char.UnicodeData.SimpleUpperCaseMapping as C
 -- Title case is used by a small number of letter ligatures like the
 -- single-character form of /Lj/.
 --
--- It matches characters with general category 'UppercaseLetter' and
--- 'TitlecaseLetter'.
+-- It matches characters with general category
+-- 'Unicode.Char.General.UppercaseLetter' and
+-- 'Unicode.Char.General.TitlecaseLetter'.
 --
 -- See: 'Unicode.Char.Case.isUpperCase' for the /full upper/ case predicate.
 --
@@ -48,11 +49,11 @@ isUpper c =
         UC.UppercaseLetter -> True
         UC.TitlecaseLetter -> True
         _                  -> False
-    where cp = ord c
+    where !cp = ord c
 
 -- | Selects lower-case alphabetic Unicode characters (letters).
 --
--- It matches characters with general category 'LowercaseLetter'.
+-- It matches characters with general category 'Unicode.Char.Case.LowercaseLetter'.
 --
 -- See: 'Unicode.Char.Case.isLowerCase' for the /full/ lower case predicate.
 --
@@ -67,7 +68,7 @@ isLower c =
     case UC.generalCategoryPlanes0To3 cp of
         UC.LowercaseLetter -> True
         _                  -> False
-    where cp = ord c
+    where !cp = ord c
 
 -- | Convert a letter to the corresponding upper-case letter, if any.
 -- Any other character is returned unchanged.
diff --git a/unicode-data/lib/Unicode/Internal/Unfold.hs b/unicode-data/lib/Unicode/Internal/Unfold.hs
index c8ae6268..b1629035 100644
--- a/unicode-data/lib/Unicode/Internal/Unfold.hs
+++ b/unicode-data/lib/Unicode/Internal/Unfold.hs
@@ -1,6 +1,5 @@
 {-# LANGUAGE CPP                       #-}
 {-# LANGUAGE ExistentialQuantification #-}
-{-# LANGUAGE LambdaCase                #-}
 
 -- |
 -- Module      : Unicode.Internal.Unfold
@@ -19,6 +18,8 @@ module Unicode.Internal.Unfold
     , toList
     ) where
 
+import GHC.Base (build)
+
 -- | An @Unfold a b@ is a generator of a stream of values of type @b@ from a
 -- seed of type @a@.
 --
@@ -57,8 +58,10 @@ toList :: Unfold a a -> a -> [a]
 toList (Unfold step inject) input =
     case inject input of
         Stop -> [input]
-        Yield b s -> b : go (step s)
-    where
-    go = \case
-        Yield b s -> let !s' = step s in b : go s'
-        Stop      -> []
+        Yield b s0 -> b : build
+                        ( \c n ->
+                            let go s = case step s of
+                                    Yield a s' -> a `c` go s'
+                                    Stop -> n
+                            in go s0
+                        )