From c8c68ec0ae29e7a5802aecae97811081f2e950da Mon Sep 17 00:00:00 2001 From: John MacFarlane Date: Wed, 23 Sep 2026 20:16:34 -0700 Subject: [PATCH] Markdown reader: reset sourcepos in parseWithString'... ...when parsing a block quote or list item. Otherwise it can happen that by the time parseWithString' is called, the position has already been set to the next file on the command line. Closes #11888. Closes #11888. --- pandoc.cabal | 3 ++ src/Text/Pandoc/Readers/Markdown.hs | 20 ++++++++----- test/command/11888.md | 46 +++++++++++++++++++++++++++++ test/command/11888/section1/lab.md | 11 +++++++ test/command/11888/section2/lab.md | 8 +++++ test/command/11888/section3/lab.md | 7 +++++ test/command/9635.md | 2 +- 7 files changed, 89 insertions(+), 8 deletions(-) create mode 100644 test/command/11888.md create mode 100644 test/command/11888/section1/lab.md create mode 100644 test/command/11888/section2/lab.md create mode 100644 test/command/11888/section3/lab.md diff --git a/pandoc.cabal b/pandoc.cabal index e4e3694b2..c0b37526a 100644 --- a/pandoc.cabal +++ b/pandoc.cabal @@ -281,6 +281,9 @@ extra-source-files: test/command/11486/scroll.revealjs test/command/11498.png test/command/11301-styles.opendocument + test/command/11888/section1/lab.md + test/command/11888/section2/lab.md + test/command/11888/section3/lab.md test/asciidoc-reader.adoc test/asciidoc-reader.native test/asciidoc-reader-include.rb diff --git a/src/Text/Pandoc/Readers/Markdown.hs b/src/Text/Pandoc/Readers/Markdown.hs index 7c3573a32..c5153f6dc 100644 --- a/src/Text/Pandoc/Readers/Markdown.hs +++ b/src/Text/Pandoc/Readers/Markdown.hs @@ -815,6 +815,7 @@ emailBlockQuote = try $ do blockQuote :: PandocMonad m => MarkdownParser m (F Blocks) blockQuote = do + pos <- getPosition raw <- emailBlockQuote (mbAlert, raw') <- (do guardEnabled Ext_alerts @@ -830,7 +831,8 @@ blockQuote = do _ -> pure (Nothing, raw)) <|> pure (Nothing, raw) -- parse the extracted block, which may contain various block elements: - contents <- parseFromString' parseBlocks $ T.intercalate "\n" raw' <> "\n\n" + contents <- parseFromString' (setPosition pos >> parseBlocks) + $ T.intercalate "\n" raw' <> "\n\n" return $ case mbAlert of Nothing -> B.blockQuote <$> contents @@ -972,11 +974,12 @@ listItem fourSpaceRule start = try $ do state <- getState let oldContext = stateParserContext state setState $ state {stateParserContext = ListItemState} + pos <- getPosition (first, continuationIndent) <- rawListItem fourSpaceRule start continuations <- many (listContinuation continuationIndent) -- parse the extracted block, which may contain various block elements: let raw = T.concat (first:continuations) - contents <- parseFromString' parseBlocks raw + contents <- parseFromString' (setPosition pos >> parseBlocks) raw updateState (\st -> st {stateParserContext = oldContext}) exts <- getOption readerExtensions return $ B.fromList . taskListItemFromAscii exts . B.toList <$> contents @@ -1017,8 +1020,9 @@ defListStart = do definitionListItem :: PandocMonad m => MarkdownParser m (F (Inlines, [Blocks])) definitionListItem = try $ do + pos <- getPosition rawLine' <- anyLine - term <- parseFromString' (trimInlinesF <$> inlines) rawLine' + term <- parseFromString' (setPosition pos >> (trimInlinesF <$> inlines)) rawLine' isTight <- (False <$ blanklines) <|> pure True fourSpaceRule <- (True <$ guardEnabled Ext_four_space_rule) <|> pure False contents <- many1 $ listItem fourSpaceRule defListStart @@ -1224,8 +1228,9 @@ lineBlock :: PandocMonad m => MarkdownParser m (F Blocks) lineBlock = do guardEnabled Ext_line_blocks try $ do + pos <- getPosition lines' <- lineBlockLines >>= - mapM (parseFromString' (trimInlinesF <$> inlines)) + mapM (parseFromString' (setPosition pos >> (trimInlinesF <$> inlines))) return $ B.lineBlock <$> sequence lines' -- @@ -1312,8 +1317,9 @@ rawTableLine indices = do tableLine :: PandocMonad m => [Int] -> MarkdownParser m (F [Blocks]) -tableLine indices = rawTableLine indices >>= - fmap sequence . mapM (parseFromString' (mconcat <$> many plain)) +tableLine indices = do + raw <- rawTableLine indices + sequence <$> mapM (parseFromString' (mconcat <$> many plain)) raw -- Parse a multiline table row and return a list of blocks (columns). multilineRow :: PandocMonad m @@ -1398,7 +1404,7 @@ multilineTableHeader headless = try $ do then [] else map (T.unlines . map trim) rawHeadsList heads <- fmap sequence $ - mapM (parseFromString' (mconcat <$> many plain).trim) rawHeads + mapM (parseFromString' (mconcat <$> many plain) . trim) rawHeads return (fmap (:[]) heads, aligns, indices') -- Parse a grid table: starts with row of '-' on top, then header diff --git a/test/command/11888.md b/test/command/11888.md new file mode 100644 index 000000000..93b4960bc --- /dev/null +++ b/test/command/11888.md @@ -0,0 +1,46 @@ +``` +% pandoc -f markdown+rebase_relative_paths command/11888/section1/lab.md command/11888/section2/lab.md command/11888/section3/lab.md +^D +

Heading 1

+
+pic 1 + +
+
+

a blockquote

+
+pic 2 + +
+

with an image reference

+
+pic 3 + +
+
+

Heading 2

+
+pic 3 + +
+
    +

  1. A list

    +
    +pic 4 + +
    2. +
+

Heading 3

+
+pic 5 + +
+
    +

  1. A list

    +
    +pic 6 + +
    2. +
+ +``` diff --git a/test/command/11888/section1/lab.md b/test/command/11888/section1/lab.md new file mode 100644 index 000000000..f0cccd2b8 --- /dev/null +++ b/test/command/11888/section1/lab.md @@ -0,0 +1,11 @@ +# Heading 1 + +![pic 1](media/Pic1_1.png) + +> a blockquote +> +> ![pic 2](media/Pic1_2.png) +> +> with an image reference +> +> ![pic 3](media/Pic1_3.png) \ No newline at end of file diff --git a/test/command/11888/section2/lab.md b/test/command/11888/section2/lab.md new file mode 100644 index 000000000..b3ae52b75 --- /dev/null +++ b/test/command/11888/section2/lab.md @@ -0,0 +1,8 @@ +# Heading 2 + +![pic 3](media/Pic2_1.png) + +1. A list + + ![pic 4](media/Pic2_2.png) + diff --git a/test/command/11888/section3/lab.md b/test/command/11888/section3/lab.md new file mode 100644 index 000000000..e19e111e6 --- /dev/null +++ b/test/command/11888/section3/lab.md @@ -0,0 +1,7 @@ +# Heading 3 + +![pic 5](media/Pic3_1.png) + +1. A list + + ![pic 6](media/Pic3_2.png) \ No newline at end of file diff --git a/test/command/9635.md b/test/command/9635.md index 088a5b6d2..5b5ca7950 100644 --- a/test/command/9635.md +++ b/test/command/9635.md @@ -6,7 +6,7 @@ okay ^D -2> [WARNING] Div at _chunk line 1 column 1 unclosed at _chunk line 5 column 1, closing implicitly. +2> [WARNING] Div at line 1 column 1 unclosed at line 5 column 1, closing implicitly.

that is not closed