Rewrite the markdown parser to follow CommonMark - #35
Merged
Merged
Conversation
The parser recognised a fixed set of constructs line by line, which left out most of the syntax the spec defines and got several of the covered cases wrong - inline code appeared verbatim in the output, *single asterisks* were not emphasis, and blockquotes and list items could hold only inline text. It now follows the two phase strategy CommonMark describes. FOBlockParser reads the source line by line and keeps a stack of open container blocks, asking each block polymorphically whether it continues (#continueOn:), what it may contain (#canContain:) and whether it takes raw lines. FOInlineParser then turns the text of the leaf blocks into inline nodes, using a delimiter stack for emphasis and a bracket stack for links, so emphasis nests and link text is parsed. New syntax: inline code, setext headings, thematic breaks, backslash escapes, HTML entities, hard and soft line breaks, autolinks, inline raw HTML, link reference definitions, reference links, loose and tight lists, ordered lists with a start number, and blockquotes and list items holding arbitrary blocks. FOHTMLRenderVisitor now produces the HTML the spec asks for, newlines included, and percent encodes link destinations. Because every block now ends with a newline, the four Foliage-Core tests that asserted rendered pages verbatim were adjusted. The two Foliage additions are kept and covered by tests: ```language=x alongside the plain ```x info string, and HTML attributes after an image destination. Measured against the 655 examples of the CommonMark 0.31.2 specification, all of which pass. FOCommonMarkSpecTest keeps 65 of them in the repository as a standing regression net. The GFM extensions (tables, task lists, strikethrough, extended autolinks) are not implemented yet. Two behaviour changes show up in existing blog sources: a code fence must now be closed by a fence at least as long as the one that opened it, and a ** run between a letter and punctuation no longer opens emphasis. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every block now ends with a newline, which the four series tests asserting rendered output did not expect. testAbstractPrefersHandWrittenDataFileOverAutoDerivedOne was already failing on main: "Add a per-post series teaser box, render series text as HTML" changed #abstract to render HTML and updated FOSeriesTest, but left this one expecting the plain source text. It now expects the same HTML the other series tests do. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
noha
force-pushed
the
feature/commonmark-core
branch
from
September 11, 2026 22:02
ec15c91 to
09d7e1b
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The parser recognised a fixed set of constructs line by line, which left out most of the syntax the spec defines and got several of the covered cases wrong - inline code appeared verbatim in the output, single asterisks were not emphasis, and blockquotes and list items could hold only inline text.
It now follows the two phase strategy CommonMark describes. FOBlockParser reads the source line by line and keeps a stack of open container blocks, asking each block polymorphically whether it continues (#continueOn:), what it may contain (#canContain:) and whether it takes raw lines. FOInlineParser then turns the text of the leaf blocks into inline nodes, using a delimiter stack for emphasis and a bracket stack for links, so emphasis nests and link text is parsed.
New syntax: inline code, setext headings, thematic breaks, backslash escapes, HTML entities, hard and soft line breaks, autolinks, inline raw HTML, link reference definitions, reference links, loose and tight lists, ordered lists with a start number, and blockquotes and list items holding arbitrary blocks.
FOHTMLRenderVisitor now produces the HTML the spec asks for, newlines included, and percent encodes link destinations. Because every block now ends with a newline, the four Foliage-Core tests that asserted rendered pages verbatim were adjusted. The two Foliage additions are kept and covered by tests:
language=x alongside the plainx info string, and HTML attributes after animage destination.
Measured against the 655 examples of the CommonMark 0.31.2 specification, all
of which pass. FOCommonMarkSpecTest keeps 65 of them in the repository as a
standing regression net. The GFM extensions (tables, task lists, strikethrough,
extended autolinks) are not implemented yet.
Two behaviour changes show up in existing blog sources: a code fence must now be
closed by a fence at least as long as the one that opened it, and a ** run
between a letter and punctuation no longer opens emphasis.