#288 Fix unstable lexing with substyled keyword and unterminated string.

Merge backtracking loops and only check line end styles - don't exit from backtracking when non-string state found. Respect isPHPScript if backtracking reaches start. Remove redundant styler.StartAt(startPos);
ScintillaOrg · Nov 23, 2024 · 8a6e5e7 · 8a6e5e7
1 parent 942ac3f
commit 8a6e5e7
Show file tree

Hide file tree

Showing 6 changed files with 29 additions and 14 deletions.
diff --git a/doc/LexillaHistory.html b/doc/LexillaHistory.html
@@ -611,6 +611,10 @@ <h3>
 	<a href="https://github.com/ScintillaOrg/lexilla/issues/286">Issue #286</a>.
 	</li>
 	<li>
+	PHP: Fix unstable lexing with substyled keyword and unterminated string.
+	<a href="https://github.com/ScintillaOrg/lexilla/issues/288">Issue #288</a>.
+	</li>
+	<li>
 	TOML: Don't treat keys without values as errors.
 	<a href="https://github.com/ScintillaOrg/lexilla/pull/283">Pull request #283</a>.
 	</li>

diff --git a/lexers/LexHTML.cxx b/lexers/LexHTML.cxx
@@ -668,6 +668,10 @@ constexpr bool isPHPStringState(int state) noexcept {
 	    (state == SCE_HPHP_COMPLEX_VARIABLE);
 }
 
+constexpr bool StyleNeedsBacktrack(int state) noexcept {
+	return InTagState(state) || isPHPStringState(state);
+}
+
 enum class AllowPHP : int {
 	None, // No PHP
 	PHP, // <?php and <?=
@@ -1190,7 +1194,6 @@ void SCI_METHOD LexerHTML::Lex(Sci_PositionU startPos, Sci_Position length, int
 	if (isPHPScript && (startPos == 0)) {
 		initStyle = SCE_HPHP_DEFAULT;
 	}
-	styler.StartAt(startPos);
 	std::string lastTag;
 	std::string prevWord;
 	PhpNumberState phpNumber;
@@ -1201,23 +1204,18 @@ void SCI_METHOD LexerHTML::Lex(Sci_PositionU startPos, Sci_Position length, int
 	int makoComment = 0;
 	std::string djangoBlockType;
 	// If inside a tag, it may be a script tag, so reread from the start of line starting tag to ensure any language tags are seen
-	if (InTagState(state)) {
-		while ((startPos > 0) && (InTagState(styler.StyleIndexAt(startPos - 1)))) {
+	// PHP string can be heredoc, must find a delimiter first. Reread from beginning of line containing the string, to get the correct lineState
+	if (StyleNeedsBacktrack(state)) {
+		while ((startPos > 0) && (StyleNeedsBacktrack(styler.StyleIndexAt(startPos - 1)))) {
 			const Sci_Position backLineStart = styler.LineStart(styler.GetLine(startPos-1));
 			length += startPos - backLineStart;
 			startPos = backLineStart;
 		}
-		state = (startPos > 0) ? styler.StyleIndexAt(startPos - 1) : SCE_H_DEFAULT;
-	}
-	// String can be heredoc, must find a delimiter first. Reread from beginning of line containing the string, to get the correct lineState
-	if (isPHPStringState(state)) {
-		while (startPos > 0 && (isPHPStringState(state) || !isLineEnd(styler[startPos - 1]))) {
-			startPos--;
-			length++;
-			state = styler.StyleIndexAt(startPos);
+		if (startPos > 0) {
+			state = styler.StyleIndexAt(startPos - 1);
+		} else {
+			state = isPHPScript ? SCE_HPHP_DEFAULT : SCE_H_DEFAULT;
 		}
-		if (startPos == 0)
-			state = SCE_H_DEFAULT;
 	}
 	styler.StartAt(startPos);
 

diff --git a/test/examples/hypertext/Issue288.php b/test/examples/hypertext/Issue288.php
@@ -0,0 +1,4 @@
+<?
+nl2br("\n); // unterminated string
+?>
+")
diff --git a/test/examples/hypertext/Issue288.php.folded b/test/examples/hypertext/Issue288.php.folded
@@ -0,0 +1,5 @@
+ 2 400   0 + <?
+ 0 401   0 | nl2br("\n); // unterminated string
+ 0 401   0 | ?>
+ 0 401   0 | ")
+ 0 401   0 | 
diff --git a/test/examples/hypertext/Issue288.php.styled b/test/examples/hypertext/Issue288.php.styled
@@ -0,0 +1,4 @@
+{18}<?{118}
+{198}nl2br{127}({119}"\n); // unterminated string
+?>
+"{127}){118}
diff --git a/test/examples/hypertext/SciTE.properties b/test/examples/hypertext/SciTE.properties
@@ -33,7 +33,7 @@ substyles.hypertext.96=1
 substylewords.96.1.*=parse
 # PHP
 substyles.hypertext.121=1
-substylewords.121.1.*=decrypt
+substylewords.121.1.*=decrypt nl2br
 
 fold=1
 fold.html=1