From bf11728f7d7dc4d53f1df16905f02f25d9ee25ce Mon Sep 17 00:00:00 2001
From: Saeed Rasooli <saeed.gnu@gmail.com>
Date: Sun, 17 Mar 2024 19:50:44 +0330
Subject: [PATCH] add support for \d and \w

---
 README.md              |  4 ++++
 lib/charclass_names.go |  5 +++++
 lib/generate_test.go   | 27 +++++++++++++++++++++++++++
 lib/lex.go             |  4 ++++
 4 files changed, 40 insertions(+)

diff --git a/README.md b/README.md
index 2fc521c..d58dc6a 100644
--- a/README.md
+++ b/README.md
@@ -40,6 +40,10 @@ Then `repassgen` binary file will be created in current directory.
 - [x] `[:ascii:]` ASCII characters
 - [x] [Unicode code points](https://www.regular-expressions.info/unicode.html), like `[\u00e0-\u00ef]{5}`
 - [x] Group references `\1`, `\2`, etc
+- [x] `\d` Digits
+- [x] `\w` Word characters: alphanumeric and underline, same as `[a-zA-Z0-9_]`
+
+
 
 # Aditional Features (not part of regexp)
 - [x] Combined multiple named/manual character classes, for example:
diff --git a/lib/charclass_names.go b/lib/charclass_names.go
index 04741d5..84a9193 100644
--- a/lib/charclass_names.go
+++ b/lib/charclass_names.go
@@ -1,5 +1,10 @@
 package passgen
 
+// \w == [a-zA-Z0-9_]
+var wordChars = []rune(
+	`abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_`,
+)
+
 var charClasses = map[string][]rune{
 	// POSIX character classes, https://www.regular-expressions.info/posixbrackets.html
 	"alpha": []rune("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"),
diff --git a/lib/generate_test.go b/lib/generate_test.go
index 1e676ef..f4840d9 100644
--- a/lib/generate_test.go
+++ b/lib/generate_test.go
@@ -777,6 +777,33 @@ func TestGenerate(t *testing.T) {
 			return true
 		},
 	})
+	test(&genCase{
+		Pattern: `\d{4}`,
+		PassLen: [2]int{4, 4},
+		Entropy: [2]float64{13.2, 13.3},
+		Validate: func(p string) bool {
+			for _, c := range p {
+				if c < '0' || c > '9' {
+					return false
+				}
+			}
+			return true
+		},
+	})
+	const wordChars = `abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_`
+	test(&genCase{
+		Pattern: `\w{4}`,
+		PassLen: [2]int{4, 4},
+		Entropy: [2]float64{23.9, 23.91},
+		Validate: func(p string) bool {
+			for _, c := range p {
+				if !strings.ContainsRune(wordChars, c) {
+					return false
+				}
+			}
+			return true
+		},
+	})
 	testErr(&genErrCase{
 		Pattern: `$base64(gh)`,
 		Error:   `         ^ value error: invalid hex number "gh"`,
diff --git a/lib/lex.go b/lib/lex.go
index fdf2f6e..e3c0781 100644
--- a/lib/lex.go
+++ b/lib/lex.go
@@ -92,6 +92,10 @@ func lexBackslash(s *State) (LexType, error) {
 			return nil, s.errorUnknown("incomplete buffer: %s", string(s.buffer))
 		}
 		return lexRootUnicodeWide, nil
+	case 'd':
+		return processRange(s, []rune("0123456789"))
+	case 'w':
+		return processRange(s, wordChars)
 	}
 	err := s.addOutputOne(backslashEscape(c))
 	if err != nil {