From 6208f327107dd4d28aac77bc1bc41591f53a9bda Mon Sep 17 00:00:00 2001
From: Aadhavan Srinivasan <aadhavan@twomorecents.org>
Date: Wed, 27 Nov 2024 11:48:04 -0500
Subject: [PATCH] Added support for numeric ranges: <5-38> will match all
 numbers between 5 and 38, inclusive on both ends. Also print line number on
 which matches occur, if we are in printing (and single line) mode

---
 main.go | 67 ++++++++++++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 64 insertions(+), 3 deletions(-)

diff --git a/main.go b/main.go
index 607d261..ef07279 100644
--- a/main.go
+++ b/main.go
@@ -39,7 +39,55 @@ See: https://blog.cernera.me/converting-regular-expressions-to-postfix-notation-
 */
 func shuntingYard(re string) []postfixNode {
 	re_postfix := make([]rune, 0)
-	re_runes := []rune(re) // Convert the string to a slice of runes to allow iteration through it
+	// Convert the string to a slice of runes to allow iteration through it
+	re_runes_orig := []rune(re) // This is the rune slice before the first parsing loop (which detects and replaces numeric ranges)
+	re_runes := make([]rune, 0)
+	// Check for numeric range. If we are at the start of a numeric range,
+	// skip to end and construct the equivalent regex for the range.
+	// The reason this is outside the loop below, is that it actually modifies
+	// the given regex (we 'cut' the numeric range and 'paste' an equivalent regex).
+	// It also makes the overall parsing easier, since I don't have to worry about the numeric range
+	// anymore.
+	// Eventually, I might be able to add it into the main parsing loop, to reduce the time
+	// complexity.
+	// A numeric range has the syntax: <num1-num2>. Ir matches all numbers in this range.
+	for i := 0; i < len(re_runes_orig); i++ {
+		c := re_runes_orig[i]
+		if c == '<' && (i == 0 || (re_runes_orig[i-1] != '\\' && re_runes_orig[i-1] != '?')) {
+			i++ // Step over opening angle bracket
+			tmpStr := ""
+			hyphenFound := false
+			for i < len(re_runes_orig) && re_runes_orig[i] != '>' {
+				if !unicode.IsDigit(re_runes_orig[i]) {
+					if re_runes_orig[i] != '-' || (hyphenFound) {
+						panic("ERROR: Invalid numeric range.")
+					}
+				}
+				if re_runes_orig[i] == '-' {
+					hyphenFound = true
+				}
+				tmpStr += string(re_runes_orig[i])
+				i++
+			}
+			// End of string reached and last character doesn't close the range
+			if i == len(re_runes_orig) && re_runes_orig[len(re_runes_orig)-1] != '>' {
+				panic("ERROR: Numeric range not closed.")
+			}
+			if len(tmpStr) == 0 {
+				panic("ERROR: Empty numeric range.")
+			}
+			// Closing bracket will be skipped when the loop variable increments
+			var rangeStart int
+			var rangeEnd int
+			fmt.Sscanf(tmpStr, "%d-%d", &rangeStart, &rangeEnd)
+			regex := range2regex(rangeStart, rangeEnd)
+			re_runes = append(re_runes, []rune(regex)...)
+
+		} else {
+			re_runes = append(re_runes, c)
+		}
+	}
+
 	/* 	Add concatenation operators.
 	Only add a concatenation operator between two characters if both the following conditions are met:
 		1. 	The first character isn't an opening parantheses or alteration operator (or an escape character)
@@ -541,6 +589,7 @@ func main() {
 	var test_runes []rune // Rune-slice representation of test_str
 	var err error
 	var linesRead bool // Whether or not we have read the lines in the file
+	lineNum := 0       // Current line number
 	// Create reader for stdin and writer for stdout
 	reader := bufio.NewReader(os.Stdin)
 	out := bufio.NewWriter(os.Stdout)
@@ -554,6 +603,7 @@ func main() {
 		if !(*multiLineFlag) {
 			// Read every string from stdin until we encounter an error. If the error isn't EOF, panic.
 			test_str, err = reader.ReadString('\n')
+			lineNum++
 			if err != nil {
 				if err == io.EOF {
 					linesRead = true
@@ -582,8 +632,19 @@ func main() {
 		test_runes = []rune(test_str)
 		matchIndices := findAllMatches(startState, test_runes)
 		if *printMatchesFlag {
-			for _, idx := range matchIndices {
-				fmt.Printf("%s\n", idx.toString())
+			// if we are in single line mode, print the line on which
+			// the matches occur
+			if len(matchIndices) > 0 {
+				if !(*multiLineFlag) {
+					fmt.Fprintf(out, "Line %d:\n", lineNum)
+				}
+				for _, idx := range matchIndices {
+					fmt.Fprintf(out, "%s\n", idx.toString())
+				}
+				err := out.Flush()
+				if err != nil {
+					panic(err)
+				}
 			}
 			continue
 		}