Expose Unicode version

Author: wismill

Changelog.md

@@ -3,6 +3,7 @@
 ## 0.3.0 (December 2021)
 
 - Support for big-endian architectures.
+- Added `unicodeVersion`.
 - Added `GeneralCategory` data type and corresponding `generalCategoryAbbr`,
   `generalCategory` functions.
 - Added the following functions to `Unicode.Char.General`:

lib/Unicode/Char.hs

@@ -32,6 +32,7 @@ module Unicode.Char
     , module Unicode.Char.Numeric
     , module Unicode.Char.Normalization
     , module Unicode.Char.Identifiers
+    , unicodeVersion
 
     -- * Re-export
     , ord
@@ -40,9 +41,16 @@ module Unicode.Char
 where
 
 import Data.Char (chr, ord)
+import Data.Version (Version, makeVersion)
 import Unicode.Char.Case
 import Unicode.Char.General
 import Unicode.Char.General.Compat
 import Unicode.Char.Identifiers
 import Unicode.Char.Numeric
 import Unicode.Char.Normalization
+
+-- | Version of Unicode standard used by @unicode-data@.
+--
+-- @since 0.3.0
+unicodeVersion :: Version
+unicodeVersion = makeVersion [14, 0, 0]

ucd.sh

@@ -53,8 +53,9 @@ download_files() {
     done
 }
 
-# Compile and run ucd2haskell
+# Generate the Haskell files.
 run_generator() {
+    # Compile and run ucd2haskell
     cabal run --flag ucd2haskell ucd2haskell -- \
           --input ./ucd/ \
           --output ./lib/ \
@@ -68,6 +69,9 @@ run_generator() {
           --core-prop XID_Continue \
           --core-prop Pattern_Syntax \
           --core-prop Pattern_White_Space
+    # Update unicodeVersion in Unicode.Char
+    VERSION_AS_LIST=$(echo "$VERSION" | sed "s/\./, /g")
+    sed -ri "s/^(unicodeVersion = makeVersion \[)[^]]*\]/\1$VERSION_AS_LIST\]/" "lib/Unicode/Char.hs"
 }
 
 # Print help text