check ruleset files in HTTPS-Everywhere for duplicate target definitions

check_https_everywhere_duplicate_rulesets.go

package main

// This script is designed to check ruleset files in HTTPS-Everywhere for duplicate target definitions

import (
	"encoding/xml"
	"flag"
	"path/filepath"
	"io/ioutil"
	"log"
)

type Target struct {
	Host string `xml:"host,attr"`
}
type Ruleset struct {
	XMLName xml.Name `xml:"ruleset"`
	Targets []Target `xml:"target"`
}

func main() {
	var (
		rulesDirectory  = flag.String("rules-dir", "src/chrome/content/rules", "path to rules.xml directory")
	)

	flag.Parse()
	
	existingHosts := make(map[string]string)
	
	files, err := filepath.Glob(filepath.Join(*rulesDirectory, "*.xml"))
	if err != nil {
		log.Fatal(err)
	}
	for _, filename := range files {
		base := filepath.Base(filename)
		body, err := ioutil.ReadFile(filename)
		if err != nil {
			log.Fatalf("err %s", err)
		}
		var r Ruleset
		err = xml.Unmarshal(body, &r)
		if err != nil {
			log.Fatalf("err %s", err)
		}
		for _, t := range r.Targets {
			if existingFile, ok := existingHosts[t.Host]; ok {
				log.Printf("%s found in %s and %s", t.Host, existingFile, base)
			}
			existingHosts[t.Host] = base
		}
	}
	log.Printf("%d hosts checked", len(existingHosts))
}