From 414161ebfaa58218ffac7907b4d8ef2babb236cb Mon Sep 17 00:00:00 2001
From: Jes Cok <xigua67damn@gmail.com>
Date: Wed, 7 Feb 2024 04:56:26 +0000
Subject: [PATCH] encoding/xml: rewrite func procInst

This CL tries to make function procInst more exact,
also adds test cases, however, including tricky ones.

Change-Id: If421299fc84d136e56a25dba7a4919c4424702c8
GitHub-Last-Rev: b9a3192718ae5535d66fddd260c27b48d93b4af1
GitHub-Pull-Request: golang/go#64336
Reviewed-on: https://go-review.googlesource.com/c/go/+/544475
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
---
 src/encoding/xml/xml.go      | 27 +++++++++++++++++++--------
 src/encoding/xml/xml_test.go |  7 +++++++
 2 files changed, 26 insertions(+), 8 deletions(-)

diff --git a/src/encoding/xml/xml.go b/src/encoding/xml/xml.go
index 73eedad290f..6b8f2e79784 100644
--- a/src/encoding/xml/xml.go
+++ b/src/encoding/xml/xml.go
@@ -2045,16 +2045,27 @@ func procInst(param, s string) string {
 	// TODO: this parsing is somewhat lame and not exact.
 	// It works for all actual cases, though.
 	param = param + "="
-	_, v, _ := strings.Cut(s, param)
-	if v == "" {
+	lenp := len(param)
+	i := 0
+	var sep byte
+	for i < len(s) {
+		sub := s[i:]
+		k := strings.Index(sub, param)
+		if k < 0 || lenp+k >= len(sub) {
+			return ""
+		}
+		i += lenp + k + 1
+		if c := sub[lenp+k]; c == '\'' || c == '"' {
+			sep = c
+			break
+		}
+	}
+	if sep == 0 {
 		return ""
 	}
-	if v[0] != '\'' && v[0] != '"' {
+	j := strings.IndexByte(s[i:], sep)
+	if j < 0 {
 		return ""
 	}
-	unquote, _, ok := strings.Cut(v[1:], v[:1])
-	if !ok {
-		return ""
-	}
-	return unquote
+	return s[i : i+j]
 }
diff --git a/src/encoding/xml/xml_test.go b/src/encoding/xml/xml_test.go
index 42f5f5f8a60..4bec4e7f1e9 100644
--- a/src/encoding/xml/xml_test.go
+++ b/src/encoding/xml/xml_test.go
@@ -830,6 +830,13 @@ var procInstTests = []struct {
 	{`version="1.0" encoding='utf-8' `, [2]string{"1.0", "utf-8"}},
 	{`version="1.0" encoding=utf-8`, [2]string{"1.0", ""}},
 	{`encoding="FOO" `, [2]string{"", "FOO"}},
+	{`version=2.0 version="1.0" encoding=utf-7 encoding='utf-8'`, [2]string{"1.0", "utf-8"}},
+	{`version= encoding=`, [2]string{"", ""}},
+	{`encoding="version=1.0"`, [2]string{"", "version=1.0"}},
+	{``, [2]string{"", ""}},
+	// TODO: what's the right approach to handle these nested cases?
+	{`encoding="version='1.0'"`, [2]string{"1.0", "version='1.0'"}},
+	{`version="encoding='utf-8'"`, [2]string{"encoding='utf-8'", "utf-8"}},
 }
 
 func TestProcInstEncoding(t *testing.T) {