Add release tagging script.

Signed-off-by: James Peach <[email protected]>

hack/make-release-tag.sh

@@ -0,0 +1,43 @@
+#! /usr/bin/env bash
+
+# make-release-tag.sh: Tag main with a release version.
+
+readonly PROGNAME=$(basename "$0")
+readonly OLDVERS="$1"
+readonly NEWVERS="$2"
+
+if [ -z "$OLDVERS" ] || [ -z "$NEWVERS" ]; then
+    printf "Usage: %s OLDVERS NEWVERS\n" "$PROGNAME"
+    exit 1
+fi
+
+set -o errexit
+set -o nounset
+set -o pipefail
+
+git tag -F - "$NEWVERS" <<EOF
+Tag $NEWVERS release.
+
+$(git shortlog "$OLDVERS..HEAD")
+EOF
+
+printf "Created tag '%s'\n" "$NEWVERS"
+
+# People set up their remotes in different ways, so we need to check
+# which one to dry run against. Choose a remote name that pushes to the
+# projectcontour org repository (i.e. not the user's Github fork).
+readonly REMOTE=$(git remote -v | awk '$2~/projectcontour\/lint/ && $3=="(push)" {print $1}' | head -n 1)
+if [ -z "$REMOTE" ]; then
+    printf "%s: unable to determine remote for %s\n" "$PROGNAME" "projectcontour/lint"
+    exit 1
+fi
+
+readonly BRANCH=$(git branch --show-current)
+
+printf "Testing whether commit can be pushed\n"
+git push --dry-run "$REMOTE" "$BRANCH"
+
+printf "Testing whether tag '%s' can be pushed\n" "$NEWVERS"
+git push --dry-run "$REMOTE" "$NEWVERS"
+
+printf "Run 'git push %s %s' to push the commit and then 'git push %s %s' to push the tag if you are happy\n" "$REMOTE" "$BRANCH" "$REMOTE" "$NEWVERS"