bug #17159 [Yaml] recognize when a block scalar is left (xabbuh)

This PR was merged into the 2.3 branch.

Discussion
----------

[Yaml] recognize when a block scalar is left

| Q             | A
| ------------- | ---
| Bug fix?      | yes
| New feature?  | no
| BC breaks?    | no
| Deprecations? | no
| Tests pass?   | yes
| Fixed tickets | #17116
| License       | MIT
| Doc PR        |

The parser did not recognize when the block scalar was completely parsed
and thus treated following comments as they need to be kept leading to
parse errors on the following lines.

Commits
-------

3a22165 [Yaml] recognize when a block scalar is left

Author: fabpot

Parser.php

@@ -303,7 +303,11 @@ private function getCurrentLineIndentation()
     private function getNextEmbedBlock($indentation = null, $inSequence = false)
     {
         $oldLineIndentation = $this->getCurrentLineIndentation();
-        $insideBlockScalar = $this->isBlockScalarHeader();
+        $blockScalarIndentations = array();
+
+        if ($this->isBlockScalarHeader()) {
+            $blockScalarIndentations[] = $this->getCurrentLineIndentation();
+        }
 
         if (!$this->moveToNextLine()) {
             return;
@@ -340,17 +344,26 @@ private function getNextEmbedBlock($indentation = null, $inSequence = false)
 
         $isItUnindentedCollection = $this->isStringUnIndentedCollectionItem();
 
-        if (!$insideBlockScalar) {
-            $insideBlockScalar = $this->isBlockScalarHeader();
+        if (empty($blockScalarIndentations) && $this->isBlockScalarHeader()) {
+            $blockScalarIndentations[] = $this->getCurrentLineIndentation();
         }
 
         $previousLineIndentation = $this->getCurrentLineIndentation();
 
         while ($this->moveToNextLine()) {
             $indent = $this->getCurrentLineIndentation();
 
-            if (!$insideBlockScalar && $indent === $previousLineIndentation) {
-                $insideBlockScalar = $this->isBlockScalarHeader();
+            // terminate all block scalars that are more indented than the current line
+            if (!empty($blockScalarIndentations) && $indent < $previousLineIndentation && trim($this->currentLine) !== '') {
+                foreach ($blockScalarIndentations as $key => $blockScalarIndentation) {
+                    if ($blockScalarIndentation >= $this->getCurrentLineIndentation()) {
+                        unset($blockScalarIndentations[$key]);
+                    }
+                }
+            }
+
+            if (empty($blockScalarIndentations) && !$this->isCurrentLineComment() && $this->isBlockScalarHeader()) {
+                $blockScalarIndentations[] = $this->getCurrentLineIndentation();
             }
 
             $previousLineIndentation = $indent;
@@ -366,7 +379,7 @@ private function getNextEmbedBlock($indentation = null, $inSequence = false)
             }
 
             // we ignore "comment" lines only when we are not inside a scalar block
-            if (!$insideBlockScalar && $this->isCurrentLineComment()) {
+            if (empty($blockScalarIndentations) && $this->isCurrentLineComment()) {
                 continue;
             }
 
@@ -523,7 +536,7 @@ private function parseBlockScalar($style, $chomping = '', $indentation = 0)
             $previousLineIndented = false;
             $previousLineBlank = false;
 
-            for ($i = 0; $i < count($blockLines); $i++) {
+            for ($i = 0; $i < count($blockLines); ++$i) {
                 if ('' === $blockLines[$i]) {
                     $text .= "\n";
                     $previousLineIndented = false;
@@ -618,7 +631,7 @@ private function isCurrentLineComment()
         //checking explicitly the first char of the trim is faster than loops or strpos
         $ltrimmedLine = ltrim($this->currentLine, ' ');
 
-        return $ltrimmedLine[0] === '#';
+        return '' !== $ltrimmedLine && $ltrimmedLine[0] === '#';
     }
 
     /**

Tests/ParserTest.php

@@ -737,7 +737,9 @@ public function testCommentLikeStringsAreNotStrippedInBlockScalars($yaml, $expec
 
     public function getCommentLikeStringInScalarBlockData()
     {
-        $yaml1 = <<<'EOT'
+        $tests = array();
+
+        $yaml = <<<'EOT'
 pages:
     -
         title: some title
@@ -752,7 +754,7 @@ public function getCommentLikeStringInScalarBlockData()
 
             footer # comment3
 EOT;
-        $expected1 = array(
+        $expected = array(
             'pages' => array(
                 array(
                     'title' => 'some title',
@@ -771,8 +773,9 @@ public function getCommentLikeStringInScalarBlockData()
                 ),
             ),
         );
+        $tests[] = array($yaml, $expected);
 
-        $yaml2 = <<<'EOT'
+        $yaml = <<<'EOT'
 test: |
     foo
     # bar
@@ -787,7 +790,7 @@ public function getCommentLikeStringInScalarBlockData()
         # bar
         baz
 EOT;
-        $expected2 = array(
+        $expected = array(
             'test' => <<<'EOT'
 foo
 # bar
@@ -814,11 +817,47 @@ public function getCommentLikeStringInScalarBlockData()
                 ),
             ),
         );
+        $tests[] = array($yaml, $expected);
 
-        return array(
-            array($yaml1, $expected1),
-            array($yaml2, $expected2),
+        $yaml = <<<EOT
+foo:
+  bar:
+    scalar-block: >
+      line1
+      line2>
+  baz:
+# comment
+    foobar: ~
+EOT;
+        $expected = array(
+            'foo' => array(
+                'bar' => array(
+                    'scalar-block' => 'line1 line2>',
+                ),
+                'baz' => array(
+                    'foobar' => null,
+                ),
+            ),
         );
+        $tests[] = array($yaml, $expected);
+
+        $yaml = <<<'EOT'
+a:
+    b: hello
+#    c: |
+#        first row
+#        second row
+    d: hello
+EOT;
+        $expected = array(
+            'a' => array(
+                'b' => 'hello',
+                'd' => 'hello',
+            ),
+        );
+        $tests[] = array($yaml, $expected);
+
+        return $tests;
     }
 
     public function testBlankLinesAreParsedAsNewLinesInFoldedBlocks()