Merge pull request #12 from satta/broken-crash

improve robustness to broken input files

.travis.yml

@@ -1,9 +1,9 @@
 language: go
 sudo: false
 before_script:
-- go get ./...
+- go get -t ./...
 - go vet ./...
-script: make release
+script: make release test
 deploy:
   provider: releases
   api_key:

bloom.go

@@ -68,6 +68,10 @@ func (s *BloomFilter) Read(input io.Reader) error {
 	}
 
 	s.k = binary.LittleEndian.Uint64(bs8)
+	maxInt := uint64(int(^uint(0) >> 1))
+	if s.k >= maxInt {
+		return fmt.Errorf("value of k (number of hash functions) is too high (%d), must be less than maximum int (%d)", s.k, maxInt)
+	}
 
 	if _, err := io.ReadFull(input, bs8); err != nil {
 		return err
@@ -189,7 +193,6 @@ const g uint64 = 18446744073709550147
 // Fingerprint returns the fingerprint of a given value, as an array of index
 // values.
 func (s *BloomFilter) Fingerprint(value []byte, fingerprint []uint64) {
-
 	hv := fnv.New64()
 	hv.Write(value)
 	hn := hv.Sum64() % m

io_test.go

@@ -6,6 +6,7 @@ package bloom
 import (
 	"io/ioutil"
 	"os"
+	"regexp"
 	"testing"
 )
 
@@ -36,6 +37,22 @@ func TestFromReaderFile(t *testing.T) {
 	checkResults(t, bf)
 }
 
+func TestFromReaderCorruptFile(t *testing.T) {
+	f, err := os.Open("testdata/broken.bloom")
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer f.Close()
+	_, err = LoadFromReader(f, false)
+	if err == nil {
+		t.Fatal("error expected")
+	}
+	r, _ := regexp.Compile("is too high")
+	if !r.MatchString(err.Error()) {
+		t.Fatalf("wrong error message: %s", err.Error())
+	}
+}
+
 func testFromSerialized(t *testing.T, gzip bool) {
 	bf := Initialize(100, 0.0001)
 	for _, v := range []string{"foo", "bar", "baz"} {