fix(parser): trim initial spaces from fenced code block

Author: iamgio

quarkdown-core/src/main/kotlin/com/quarkdown/core/lexer/patterns/BaseMarkdownBlockTokenRegexPatterns.kt

@@ -109,7 +109,7 @@ open class BaseMarkdownBlockTokenRegexPatterns {
             wrap = ::FencesCodeToken,
             regex =
                 RegexBuilder(
-                    "^ {0,3}fencesstart[ \\t]*lang?[ \\t]*caption?[ \\t]*customid?$" +
+                    "^( {0,3})fencesstart[ \\t]*lang?[ \\t]*caption?[ \\t]*customid?$" +
                         "(?s)(.+?)" +
                         "fencesend[ \\t]*$",
                 ).withReference("fencesstart", "(?<fenceschar>[`~]){3,}")

quarkdown-core/src/main/kotlin/com/quarkdown/core/parser/BlockTokenParser.kt

@@ -99,15 +99,26 @@ class BlockTokenParser(
 
     override fun visit(token: FencesCodeToken): Node {
         val groups = token.data.groups.iterator(consumeAmount = 2)
+        val initialSpaces = groups.next().length // Amount of spaces before the fence.
         val language = token.data.namedGroups["fencescodelang"]
         val caption = token.data.namedGroups["fencescodecaption"]?.trim()
         val referenceId = token.data.namedGroups["fencescodecustomid"]?.trim()
 
+        // Removes, at most, the initial spaces from each line (GFM #101).
+        val content =
+            groups
+                .next()
+                .lineSequence()
+                .map { it.replace("^ {0,$initialSpaces}".toRegex(), "") }
+                .joinToString(separator = "\n")
+                .removePrefix("\n")
+                .removeSuffix("\n")
+
         return Code(
             language = language?.takeIf { it.isNotBlank() }?.trim(),
             caption = caption?.trimDelimiters(),
             referenceId = referenceId,
-            content = groups.next().trim(),
+            content = content,
         )
     }
 

quarkdown-core/src/test/kotlin/com/quarkdown/core/BlockParserTest.kt

@@ -204,6 +204,10 @@ class BlockParserTest {
             assertEquals("Code line 1\n    Code line 2", content)
             assertNull(language)
         }
+        with(nodes.next()) {
+            assertEquals("Code line 1\nCode line 2\n Code line 3\n  Code line 4", content)
+            assertNull(language)
+        }
         with(nodes.next()) {
             assertEquals("Code", content)
             assertEquals("text", language)

quarkdown-core/src/test/resources/parsing/fencescode.md

@@ -16,6 +16,13 @@ Code line 1
     Code line 2
 ```
 
+  ```
+ Code line 1
+  Code line 2
+   Code line 3
+    Code line 4
+ ```
+
 ```text
 Code
 ```

quarkdown-test/src/test/kotlin/com/quarkdown/test/CodeTest.kt

@@ -81,6 +81,28 @@ class CodeTest {
         }
     }
 
+    // #259
+    @Test
+    fun `indented block`() {
+        execute(
+            """
+            - indented context .br
+                ```
+                aaaa
+                bbbb
+                cccc
+                ```
+            """.trimIndent(),
+        ) {
+            assertEquals(
+                "<ul><li>indented context <br />" +
+                    "<pre><code>aaaa\nbbbb\ncccc</code></pre>" +
+                    "</li></ul>",
+                it,
+            )
+        }
+    }
+
     // #32
     @Test
     fun `long block`() {