From c1f6a3a3b050de905aee4537b925a736ebd584a8 Mon Sep 17 00:00:00 2001
From: Simon Let <simon.a.let@gmail.com>
Date: Sun, 11 Aug 2019 18:08:34 +0200
Subject: [PATCH] minor changes, cleanup

---
 go.mod                                    |  1 -
 go.sum                                    |  2 --
 sanitize-history/resh-sanitize-history.go | 14 +++++++++-----
 sanitizer_data/whitelist.txt              |  3 +++
 4 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/go.mod b/go.mod
index 9ff85f1..28d5ece 100644
--- a/go.mod
+++ b/go.mod
@@ -4,6 +4,5 @@ go 1.12
 
 require (
 	github.com/BurntSushi/toml v0.3.1
-	github.com/mattn/go-shellwords v1.0.5
 	github.com/whilp/git-urls v0.0.0-20160530060445-31bac0d230fa
 )
diff --git a/go.sum b/go.sum
index 629918a..72fef1f 100644
--- a/go.sum
+++ b/go.sum
@@ -1,6 +1,4 @@
 github.com/BurntSushi/toml v0.3.1 h1:WXkYYl6Yr3qBf1K79EBnL4mak0OimBfB0XUf9Vl28OQ=
 github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
-github.com/mattn/go-shellwords v1.0.5 h1:JhhFTIOslh5ZsPrpa3Wdg8bF0WI3b44EMblmU9wIsXc=
-github.com/mattn/go-shellwords v1.0.5/go.mod h1:3xCvwCdWdlDJUrvuMn7Wuy9eWs4pE8vqg+NOMyg4B2o=
 github.com/whilp/git-urls v0.0.0-20160530060445-31bac0d230fa h1:rW+Lu6281ed/4XGuVIa4/YebTRNvoUJlfJ44ktEVwZk=
 github.com/whilp/git-urls v0.0.0-20160530060445-31bac0d230fa/go.mod h1:2rx5KE5FLD0HRfkkpyn8JwbVLBdhgeiOb2D2D9LLKM4=
diff --git a/sanitize-history/resh-sanitize-history.go b/sanitize-history/resh-sanitize-history.go
index 9dbe9ea..c2bb51a 100644
--- a/sanitize-history/resh-sanitize-history.go
+++ b/sanitize-history/resh-sanitize-history.go
@@ -38,8 +38,8 @@ func main() {
 
 	showVersion := flag.Bool("version", false, "Show version and exit")
 	showRevision := flag.Bool("revision", false, "Show git revision and exit")
-	trimHashes := flag.Int("trim-hashes", 12, "Trim hashes to N characters (default: 12), 0 turns off trimming")
-	outputPath := flag.String("output", "", "Output file")
+	trimHashes := flag.Int("trim-hashes", 12, "Trim hashes to N characters, '0' turns off trimming")
+	outputPath := flag.String("output", "", "Output file (default: use stdout)")
 
 	flag.Parse()
 
@@ -177,11 +177,13 @@ func (s *sanitizer) sanitizeRecord(record *common.Record) error {
 }
 
 func (s *sanitizer) sanitizeCmdLine(cmdLine string) (string, error) {
+	const optionEndingChars = "=;)"
+	const optionAllowedChars = "-_"
 	sanCmdLine := ""
 	buff := ""
 
 	// simple options shouldn't be sanitized
-	// 1) whitespace 2) "-" or "--" 3) letters, digits, "-", "_" 4) ending whitespace or "="
+	// 1) whitespace 2) "-" or "--" 3) letters, digits, "-", "_" 4) ending whitespace or any of "=;)"
 	var optionDetected bool
 
 	prevR3 := ' '
@@ -190,7 +192,7 @@ func (s *sanitizer) sanitizeCmdLine(cmdLine string) (string, error) {
 	for _, r := range cmdLine {
 		switch optionDetected {
 		case true:
-			if unicode.IsSpace(r) || r == '=' || r == ';' {
+			if unicode.IsSpace(r) || strings.ContainsRune(optionEndingChars, r) {
 				// whitespace, "=" or ";" ends the option
 				// => add option unsanitized
 				optionDetected = false
@@ -199,7 +201,8 @@ func (s *sanitizer) sanitizeCmdLine(cmdLine string) (string, error) {
 					buff = ""
 				}
 				sanCmdLine += string(r)
-			} else if unicode.IsLetter(r) == false && unicode.IsDigit(r) == false && r != '-' && r != '_' {
+			} else if unicode.IsLetter(r) == false && unicode.IsDigit(r) == false &&
+				strings.ContainsRune(optionAllowedChars, r) == false {
 				// r is not any of allowed chars for an option: letter, digit, "-" or "_"
 				// => sanitize
 				if len(buff) > 0 {
@@ -217,6 +220,7 @@ func (s *sanitizer) sanitizeCmdLine(cmdLine string) (string, error) {
 		case false:
 			// split command on all non-letter and non-digit characters
 			if unicode.IsLetter(r) == false && unicode.IsDigit(r) == false {
+				// TODO: decide if we want to split on tokens
 				// split token
 				if len(buff) > 0 {
 					sanToken, err := s.sanitizeCmdToken(buff)
diff --git a/sanitizer_data/whitelist.txt b/sanitizer_data/whitelist.txt
index e765f43..74b4625 100644
--- a/sanitizer_data/whitelist.txt
+++ b/sanitizer_data/whitelist.txt
@@ -104,6 +104,7 @@ com
 combinedeltarpm
 comm
 command
+convert
 coredumpctl
 cp
 cpgr
@@ -781,6 +782,7 @@ view
 vigr
 vipw
 visudo
+vlc
 vmstat
 w
 wait
@@ -812,6 +814,7 @@ xzfgrep
 xzgrep
 xzless
 xzmore
+yaourt
 yes
 ypdomainname
 yum