[#strings] glsub support for overlapping and simplification

1 files changed, 20 insertions, 22 deletions

diff --git a/awk-edits/#strings.awk b/awk-edits/#strings.awk
index 3365df2..5e214cb 100644
--- a/awk-edits/#strings.awk
+++ b/awk-edits/#strings.awk
@@ -49,7 +49,8 @@ function gesub(regexp, exregexp, replacement, target) {
 # Behaves kind of like gsub, however substitution of regexp with replacement
 # only happens, when regexp is preceded by lbregexp and succeeded by laregexp.
 # Only regexp get's replaced, the matches of lbregexp and laregexp are left
-# as-is. Overlapping of any form isn't supported.
+# as-is. Overlapping of the lookahead and lookbehind substrings is supported,
+# but any sort of overlapping which includes regexp isn't.
 # This function is quite slow, try to use as little as possible.
 # Returns the result string, rather than doing it inplace.
 function glsub(lbregexp, regexp, laregexp, replacement, target) {
@@ -57,29 +58,26 @@ function glsub(lbregexp, regexp, laregexp, replacement, target) {
 		target = $0;
 	}
 
-	surlen = split(target, surrounds, lbregexp regexp laregexp);
-	result = surrounds[1];
-	# POSIX compliance, seps parameter in split isn't available
-	target = substr(target, length(surrounds[1]) + 1);
-	for (i = 2; i <= surlen; i++) {
-		if (surrounds[i] == "") {
-			match(target, lbregexp regexp laregexp);
-			si = RLENGTH + 1;
-		}
-		else {
-			si = index(target, surrounds[i]);
+	result = ""
+	match(target, lbregexp regexp);
+	while (RLENGTH > -1) {
+		result = result substr(target, 1, RSTART - 1);
+		matched = substr(target, RSTART, RLENGTH);
+		target = substr(target, RSTART + RLENGTH);
+		
+		match(target, laregexp);
+		if (RSTART == 1) {
+			match(matched, regexp);
+			result = result substr(matched, 1, RSTART - 1);
+			matched = substr(matched, RSTART);
+			sub(/.*/, replacement, matched);
 		}
-		csep = substr(target, 1, si - 1);
-		target = substr(target, length(surrounds[i]) + si);
-
-		match(csep, regexp);
-		behind = substr(csep, 1, RSTART - 1);
-		ahead = substr(csep, RSTART + RLENGTH);
-		csep = substr(csep, RSTART, RLENGTH);
+		result = result matched;
 
-		sub(/.*/, replacement, csep);
-		result = result behind csep ahead surrounds[i];
+		match(target, lbregexp regexp);
 	}
 
-	return result;
+	return result target;
+}
+
 }