pkg/email: improve parsing of splitted lines

Allow:

full-commit-title-on-next-line

This allows commit titles between 70 and 80 cols with gmail.

Also be more permissive wrt spaces and tabs.

2 files changed, 80 insertions, 29 deletions

diff --git a/pkg/email/parser.go b/pkg/email/parser.go
index 67c0159cd..c59736a1b 100644
--- a/pkg/email/parser.go
+++ b/pkg/email/parser.go
@@ -172,42 +172,69 @@ func extractCommand(body []byte) (cmd, args string) {
 		return
 	}
 	cmdPos += len(commandPrefix)
+	for cmdPos < len(body) && body[cmdPos] == ' ' {
+		cmdPos++
+	}
 	cmdEnd := bytes.IndexByte(body[cmdPos:], '\n')
 	if cmdEnd == -1 {
 		cmdEnd = len(body) - cmdPos
 	}
-	cmdLine := strings.TrimSpace(string(body[cmdPos : cmdPos+cmdEnd]))
-	if cmdLine == "" {
-		return
+	if cmdEnd1 := bytes.IndexByte(body[cmdPos:], '\r'); cmdEnd1 != -1 && cmdEnd1 < cmdEnd {
+		cmdEnd = cmdEnd1
 	}
-	split := strings.Split(cmdLine, " ")
-	cmd = split[0]
-	var aargs []string
-	if len(split) > 1 {
-		aargs = split[1:]
+	if cmdEnd1 := bytes.IndexByte(body[cmdPos:], ' '); cmdEnd1 != -1 && cmdEnd1 < cmdEnd {
+		cmdEnd = cmdEnd1
 	}
-	// Text emails are split at 80 columns are the transformation is irrevesible.
-	// Try to restore args for some commands that don't have spaces in args.
-	want := 0
+	cmd = string(body[cmdPos : cmdPos+cmdEnd])
+	// Some email clients split text emails at 80 columns are the transformation is irrevesible.
+	// We try hard to restore what was there before.
+	// For "test:" command we know that there must be 2 tokens without spaces.
+	// For "fix:"/"dup:" we need a whole non-empty line of text.
 	switch cmd {
 	case "test:":
-		want = 2
+		args = extractArgsTokens(body[cmdPos+cmdEnd:], 2)
 	case "test_5_arg_cmd":
-		want = 5
+		args = extractArgsTokens(body[cmdPos+cmdEnd:], 5)
+	case "fix:", "dup:":
+		args = extractArgsLine(body[cmdPos+cmdEnd:])
 	}
-	for pos := cmdPos + cmdEnd + 1; len(aargs) < want && pos < len(body); {
+	return
+}
+
+func extractArgsTokens(body []byte, num int) string {
+	var args []string
+	for pos := 0; len(args) < num && pos < len(body); {
 		lineEnd := bytes.IndexByte(body[pos:], '\n')
 		if lineEnd == -1 {
 			lineEnd = len(body) - pos
 		}
 		line := strings.TrimSpace(string(body[pos : pos+lineEnd]))
+		for {
+			line1 := strings.Replace(line, "  ", " ", -1)
+			if line == line1 {
+				break
+			}
+			line = line1
+		}
 		if line != "" {
-			aargs = append(aargs, strings.Split(line, " ")...)
+			args = append(args, strings.Split(line, " ")...)
 		}
 		pos += lineEnd + 1
 	}
-	args = strings.TrimSpace(strings.Join(aargs, " "))
-	return
+	return strings.TrimSpace(strings.Join(args, " "))
+}
+
+func extractArgsLine(body []byte) string {
+	pos := 0
+	for pos < len(body) && (body[pos] == ' ' || body[pos] == '\t' ||
+		body[pos] == '\n' || body[pos] == '\r') {
+		pos++
+	}
+	lineEnd := bytes.IndexByte(body[pos:], '\n')
+	if lineEnd == -1 {
+		lineEnd = len(body) - pos
+	}
+	return strings.TrimSpace(string(body[pos : pos+lineEnd]))
 }
 
 func parseBody(r io.Reader, headers mail.Header) (body []byte, attachments [][]byte, err error) {
diff --git a/pkg/email/parser_test.go b/pkg/email/parser_test.go
index e4e79e253..727dafc02 100644
--- a/pkg/email/parser_test.go
+++ b/pkg/email/parser_test.go
@@ -19,6 +19,12 @@ func TestExtractCommand(t *testing.T) {
 				t.Logf("got   : %q %q", cmd, args)
 				t.Fail()
 			}
+			cmd, args = extractCommand([]byte(strings.Replace(test.body, "\n", "\r\n", -1)))
+			if cmd != test.cmd || !reflect.DeepEqual(args, test.args) {
+				t.Logf("expect: %q %q", test.cmd, test.args)
+				t.Logf("got   : %q %q", cmd, args)
+				t.Fail()
+			}
 		})
 	}
 }
@@ -118,24 +124,24 @@ var extractCommandTests = []struct {
 		body: `Hello,
 
 line1
-#syz  foo  bar baz 	`,
-		cmd:  "foo",
+#syz  fix:  bar baz 	`,
+		cmd:  "fix:",
 		args: "bar baz",
 	},
 	{
 		body: `Hello,
 
 line1
-#syz foo bar  	 baz
+#syz fix:  bar  	 baz
 line 2
 `,
-		cmd: "foo",
+		cmd: "fix:",
 		args: "bar  	 baz",
 	},
 	{
 		body: `
 line1
-> #syz foo bar   baz
+> #syz fix: bar   baz
 line 2
 `,
 		cmd:  "",
@@ -173,7 +179,7 @@ locking/core
 		body: `
 #syz test_5_arg_cmd arg1
 
- arg2 arg3
+ arg2  arg3
  
 arg4
 arg5
@@ -206,6 +212,24 @@ arg2
 		cmd:  "test_5_arg_cmd",
 		args: "arg1 arg2",
 	},
+	{
+		body: `
+#syz fix:
+arg1 arg2 arg3
+arg4 arg5
+ 
+`,
+		cmd:  "fix:",
+		args: "arg1 arg2 arg3",
+	},
+	{
+		body: `
+#syz  fix: arg1 arg2 arg3
+arg4 arg5 
+`,
+		cmd:  "fix:",
+		args: "arg1 arg2 arg3",
+	},
 }
 
 type ParseTest struct {
@@ -223,7 +247,7 @@ Content-Type: text/plain; charset="UTF-8"
 
 text body
 second line
-#syz command 	 arg1 arg2 arg3 	
+#syz fix: 	 arg1 arg2 arg3 	
 last line
 -- 
 You received this message because you are subscribed to the Google Groups "syzkaller" group.
@@ -240,7 +264,7 @@ For more options, visit https://groups.google.com/d/optout.`,
 			Cc:        []string{"bob@example.com"},
 			Body: `text body
 second line
-#syz command 	 arg1 arg2 arg3 	
+#syz fix: 	 arg1 arg2 arg3 	
 last line
 -- 
 You received this message because you are subscribed to the Google Groups "syzkaller" group.
@@ -249,7 +273,7 @@ To post to this group, send email to syzkaller@googlegroups.com.
 To view this discussion on the web visit https://groups.google.com/d/msgid/syzkaller/abcdef@google.com.
 For more options, visit https://groups.google.com/d/optout.`,
 			Patch:       "",
-			Command:     "command",
+			Command:     "fix:",
 			CommandArgs: "arg1 arg2 arg3",
 		}},
 
@@ -280,7 +304,7 @@ From: Bob <bob@example.com>
 To: syzbot <bot@example.com>, Alice <alice@example.com>
 Content-Type: text/plain
 
-#syz  command   	 
+#syz  invalid   	 
 text body
 second line
 last line`,
@@ -289,12 +313,12 @@ last line`,
 			Subject:   "test subject",
 			From:      "\"Bob\" <bob@example.com>",
 			Cc:        []string{"alice@example.com", "bob@example.com", "bot@example.com"},
-			Body: `#syz  command   	 
+			Body: `#syz  invalid   	 
 text body
 second line
 last line`,
 			Patch:       "",
-			Command:     "command",
+			Command:     "invalid",
 			CommandArgs: "",
 		}},