core/trinitry/update.sh

#!/usr/bin/env sh

# Library {{{
mktmp() {
	ensure_tmp_dir
	mktemp -p "$LIB_TEMP_DIR_FOR_SCRIPT"
}
ensure_tmp_dir() {
	if ! [ -d "$LIB_TEMP_DIR_FOR_SCRIPT" ]; then
		LIB_TEMP_DIR_FOR_SCRIPT="$(mktemp -d)"
		export LIB_TEMP_DIR_FOR_SCRIPT
	fi
}
remove_tmp_dir() {
	# The test is here because some scripts still delete this on their own
	if [ -d "$LIB_TEMP_DIR_FOR_SCRIPT" ]; then
		rm -r "$LIB_TEMP_DIR_FOR_SCRIPT"
	fi
}

trap remove_tmp_dir EXIT
ensure_tmp_dir # ensure that the variable has been set, even in subshells
# }}}

tmp="$(mktmp)"

curl https://raw.githubusercontent.com/minimaxir/big-list-of-naughty-strings/master/blns.txt |
	awk '!/^#/' |
	awk '!/^[\s\t]*$/' |
	awk -v s="'" '!/.*s.*/' |

	# This entry contains duplicated spaces, just ignore it
	awk '!/Power/' |
	iconv -c -f utf-8 -t ascii \
		>"$tmp"

counter=0

cat <<EOF
// DO NOT EDIT
// This file is automatically generated by the 'update.sh' file, with data from:
// https://raw.githubusercontent.com/minimaxir/big-list-of-naughty-strings/master/blns.txt

#[cfg(test)]
mod test {
    use crate::Trinitry;
EOF

while read -r name; do
	tmp2="$(mktmp)"
	printf "%s" "$name" >"$tmp2"
	if rg '^[A-Za-z0-9_.-]+$' "$tmp2" -q; then
		cat <<EOF
    #[test]
    fn parse_$counter() {
        let p = Trinitry::new(r##"$name"##).unwrap_or_else(|e| {
            panic!("{}", e);
        });
        assert_eq!(r##"$name"##, &p.to_string());
    }
EOF
	elif rg '^[A-Za-z0-9_.-]+ [A-Za-z0-9_.(){}<>?!+^@&*~|=,/\\ -]*$' "$tmp2" -q; then
		cat <<EOF
    #[test]
    fn parse_$counter() {
        let p = Trinitry::new(r##"$name"##).unwrap_or_else(|e| {
            panic!("{}", e);
        });
        assert_eq!(r##"$name"##, &p.to_string());
    }
EOF
	else
		cat <<EOF
    #[test]
    fn parse_$counter() {
        let p = Trinitry::new(r##"$name"##);
        assert!(p.is_err());
    }
EOF
	fi
	counter=$((counter + 1))
done <"$tmp"

cat <<EOF
}
EOF

# vim: ft=sh