From 2fb5366952643d3cf742f4f212d431f4b8ccb167 Mon Sep 17 00:00:00 2001
From: Lilleman auf Larv <lilleman@larvit.se>
Date: Sun, 18 Oct 2026 12:05:27 +0200
Subject: [PATCH] =?UTF-8?q?Answer=20stability=20round=201:=201=20?=
 =?UTF-8?q?=F0=9F=94=B4,=202=20=F0=9F=9F=A1;=20read=20a=20quote=20marker?=
 =?UTF-8?q?=20behind=20a=20split=20tab's=20leftover=20columns=20as=20text,?=
 =?UTF-8?q?=20and=20pin=20the=20tab=20cases?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 src/markdown/parse/blocks.test.ts | 26 ++++++++++++++++++++++++++
 src/markdown/parse/blocks.ts      |  7 ++++---
 2 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/src/markdown/parse/blocks.test.ts b/src/markdown/parse/blocks.test.ts
index 0e945e1..c597bdb 100644
--- a/src/markdown/parse/blocks.test.ts
+++ b/src/markdown/parse/blocks.test.ts
@@ -152,3 +152,29 @@ test('stops a blank line at an empty item, and gives a code block in an item the
     { kind: 'code', language: '', position: { line: 1, offset: 0 }, text: '\nx' },
   ])
 })
+
+function outline(blocks: readonly Block[]): string {
+  return blocks
+    .map((block) => {
+      if (block.kind === 'blockquote') return `quote(${outline(block.blocks)})`
+      if (block.kind === 'bulletList') return `list(${block.items.map(outline).join(',')})`
+      if (block.kind === 'code') return `code(${JSON.stringify(block.text)})`
+      if (block.kind === 'paragraph') return `paragraph(${block.text.split('\n').map((line) => line.trim()).join('|')})`
+      return block.kind
+    })
+    .join(' ')
+}
+
+test('reads a marker a tab pushes to four columns past its container as text, not as a marker', () => {
+  const read = (markdown: string): string => outline(parseBlocks(markdown, flavourClaims.lossless).blocks)
+  assert.equal(read('> a\n \t> b\n'), 'quote(paragraph(a|> b))')
+  assert.equal(read('> > a\n>\t\t> b\n'), 'quote(quote(paragraph(a|> b)))')
+  assert.equal(read('- > a\n\t\t> b\n'), 'list(quote(paragraph(a|> b)))')
+  assert.equal(read('x\n\t# h\n'), 'paragraph(x|# h)')
+})
+
+test('gives what follows an item the columns of the tab a cut splits', () => {
+  const read = (markdown: string): string => outline(parseBlocks(markdown, flavourClaims.lossless).blocks)
+  assert.equal(read('- ```\n\tx\n  ```\n'), 'list(code("  x"))')
+  assert.equal(read('- a\n \t- b\n'), 'list(paragraph(a) list(paragraph(b)))')
+})
diff --git a/src/markdown/parse/blocks.ts b/src/markdown/parse/blocks.ts
index ea17138..3e7415f 100644
--- a/src/markdown/parse/blocks.ts
+++ b/src/markdown/parse/blocks.ts
@@ -63,8 +63,9 @@ type ContainerStart = { kind: 'blockquote'; rest: Line } | { fresh: boolean; ind
 type ParagraphLine = { position: SourcePosition; text: string }
 
 // The line from an absolute column on: a tab a cut splits keeps the stop it is measured against, and its leftover columns
-// stand as `spaces` before `text`, so a cut never builds a string. At most 3, which a cut of `largestOpenerIndentation` takes whole.
-// They are blank, so blankness reads `text` alone; a block takes the line through `lineText`.
+// stand as `spaces` before `text`, so a cut shares `text`. At most 3, and blank, so blankness reads `text` alone. A cut of
+// `largestOpenerIndentation` leaves some only on a line indented as code, which no marker follows: a marker check reads
+// `spaces` first, and a block takes the line through `lineText`.
 type Line = { column: number; spaces: number; text: string }
 
 type LeafOpener = { index: number; position: SourcePosition }
@@ -201,7 +202,7 @@ function continuesContainer(container: EdgeContainer, line: Line): Line | undefi
 }
 
 function blockquoteRest(opener: Line): Line | undefined {
-  if (!opener.text.startsWith('>')) return undefined
+  if (opener.spaces > 0 || !opener.text.startsWith('>')) return undefined
   return removeColumns({ column: opener.column + 1, spaces: 0, text: opener.text.slice(1) }, 1)
 }