tools/update-unicode-data.sh: Add a script to handle the unicode updates

We need to perform multiple actions on unicode updates, so let's handle
all the commands in a single script so that we don't have to remember
all the details each time.

1 files changed, 38 insertions, 0 deletions

diff --git a/tools/update-unicode-data.sh b/tools/update-unicode-data.sh
new file mode 100755
index 000000000..2d751db15
--- /dev/null
+++ b/tools/update-unicode-data.sh
@@ -0,0 +1,38 @@
+#!/usr/bin/env bash
+
+if [ ! -d "$1" ]; then
+    echo "Usage $(basename "$0") UCD-directory [version]"
+    exit 1
+fi
+
+ucd=$(realpath "$1")
+version=$2
+glib_dir=$(git -C "$(dirname "$0")" rev-parse --show-toplevel)
+
+# shellcheck disable=SC2144 # we only want to match a file like this
+if ! [ -f "$ucd"/UnicodeData*.txt ] || ! [ -f "$ucd"/CaseFolding.*txt ]; then
+    echo "'$ucd' does not look like an Unicode Database directory";
+fi
+
+if [ -z "$version" ]; then
+    readme=("$ucd"/ReadMe*.txt)
+    version=$(sed -n "s,.*Version \([0-9.]\+\) of the Unicode Standard.*,\1,p" \
+        "${readme[@]}")
+
+    if [ -z "$version" ]; then
+        echo "Invalid version found"
+        exit 1
+    fi
+fi
+
+cd "$glib_dir" || exit 1
+
+echo "Updating generated code to Unicode version $version"
+set -xe
+
+(cd glib && ./gen-unicode-tables.pl -both "$version" "$ucd")
+glib/tests/gen-casefold-txt.py "$version" \
+    "$ucd"/CaseFolding*.txt > glib/tests/casefold.txt
+glib/tests/gen-casemap-txt.py "$version" \
+    "$ucd"/UnicodeData*.txt \
+    "$ucd"/SpecialCasing*.txt > glib/tests/casemap.txt