archive/tar: fix issues with readGNUSparseMap1x0

Motivations:
* Use of strconv.ParseInt does not properly treat integers as 64bit,
preventing this function from working properly on 32bit machines.
* Use of io.ReadFull does not properly detect truncated streams
when the file suddenly ends on a block boundary.
* The function blindly trusts user input for numEntries and allocates
memory accordingly.
* The function does not validate that numEntries is not negative,
allowing a malicious sparse file to cause a panic during make.

In general, this function was overly complicated for what it was
accomplishing and it was hard to reason that it was free from
bounds errors. Instead, it has been rewritten and relies on
bytes.Buffer.ReadString to do the main work. So long as invariants
about the number of '\n' in the buffer are maintained, it is much
easier to see why this approach is correct.

Change-Id: Ibb12c4126c26e0ea460ea063cd17af68e3cf609e
Reviewed-on: https://go-review.googlesource.com/15174
Reviewed-by: Russ Cox <rsc@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>

archive/tar/reader_test.go

@@ -298,9 +298,7 @@ var untarTests = []*untarTest{
 	},
 	{
 		file: "testdata/issue11169.tar",
-		// TODO(dsnet): Currently the library does not detect that this file is
-		// malformed. Instead it incorrectly believes that file just ends.
-		// err:  ErrHeader,
+		err:  ErrHeader,
 	},
 }
 
@@ -983,3 +981,55 @@ func TestReadHeaderOnly(t *testing.T) {
 		}
 	}
 }
+
+func TestParsePAXRecord(t *testing.T) {
+	var medName = strings.Repeat("CD", 50)
+	var longName = strings.Repeat("AB", 100)
+
+	var vectors = []struct {
+		input     string
+		residual  string
+		outputKey string
+		outputVal string
+		ok        bool
+	}{
+		{"6 k=v\n\n", "\n", "k", "v", true},
+		{"19 path=/etc/hosts\n", "", "path", "/etc/hosts", true},
+		{"210 path=" + longName + "\nabc", "abc", "path", longName, true},
+		{"110 path=" + medName + "\n", "", "path", medName, true},
+		{"9 foo=ba\n", "", "foo", "ba", true},
+		{"11 foo=bar\n\x00", "\x00", "foo", "bar", true},
+		{"18 foo=b=\nar=\n==\x00\n", "", "foo", "b=\nar=\n==\x00", true},
+		{"27 foo=hello9 foo=ba\nworld\n", "", "foo", "hello9 foo=ba\nworld", true},
+		{"27 ☺☻☹=日a本b語ç\nmeow mix", "meow mix", "☺☻☹", "日a本b語ç", true},
+		{"17 \x00hello=\x00world\n", "", "\x00hello", "\x00world", true},
+		{"1 k=1\n", "1 k=1\n", "", "", false},
+		{"6 k~1\n", "6 k~1\n", "", "", false},
+		{"6_k=1\n", "6_k=1\n", "", "", false},
+		{"6 k=1 ", "6 k=1 ", "", "", false},
+		{"632 k=1\n", "632 k=1\n", "", "", false},
+		{"16 longkeyname=hahaha\n", "16 longkeyname=hahaha\n", "", "", false},
+		{"3 somelongkey=\n", "3 somelongkey=\n", "", "", false},
+		{"50 tooshort=\n", "50 tooshort=\n", "", "", false},
+	}
+
+	for _, v := range vectors {
+		key, val, res, err := parsePAXRecord(v.input)
+		ok := (err == nil)
+		if v.ok != ok {
+			if v.ok {
+				t.Errorf("parsePAXRecord(%q): got parsing failure, want success", v.input)
+			} else {
+				t.Errorf("parsePAXRecord(%q): got parsing success, want failure", v.input)
+			}
+		}
+		if ok && (key != v.outputKey || val != v.outputVal) {
+			t.Errorf("parsePAXRecord(%q): got (%q: %q), want (%q: %q)",
+				v.input, key, val, v.outputKey, v.outputVal)
+		}
+		if res != v.residual {
+			t.Errorf("parsePAXRecord(%q): got residual %q, want residual %q",
+				v.input, res, v.residual)
+		}
+	}
+}