# CommonMark spec.txt v0.31.2 examples we deliberately do not target yet.
#
# Format: one example number per line. Add a comment line above each entry
# explaining why we are not targeting it. This file is informational; the
# allowlist is what gates regression.
#
# The entries below are examples that previously passed by accident — under
# the prior CommonMark flavor defaults the relevant constructs (inline links,
# reference links, fenced code, etc.) weren't enabled at all, so the parser
# fell through to plain-text and matched the spec's expected paragraph
# rendering. Turning on those constructs (which CommonMark requires) exposed
# pre-existing parser laxity in the link / reference-definition / autolink
# code paths. Each needs its own focused parser fix; tracked here so the
# follow-up scope stays visible.

# Reference-definition parser is too lenient: accepts trailing junk after the
# title (`[foo]: /url "title" ok` is not a valid def, must fall back to
# paragraph) and also lets a def interrupt a paragraph (`Foo\n[bar]: /baz`
# should stay inside the paragraph).
209
213

# Code-span vs link precedence: a backtick run inside a link's text/dest must
# end the code span before the link is closed, but the parser currently
# greedily closes the link first.
342
525

# Inline link parser is too lenient on URL/title syntax (CommonMark stricter
# than Pandoc-markdown):
#  - 488: bare URL cannot contain spaces
#  - 490: URL cannot span lines
#  - 493: angle-bracketed URL rejects `\>` escapes
#  - 508: title cannot contain unescaped same-quote char
#  - 523: emphasis run inside a link's bracketed text must close before `]`
#  - 546: link label rejects an unescaped `[` inside it
488
490
493
508
523
546

# Autolink validator accepts strings that aren't valid URI / email autolinks:
#  - 606: backslash escapes are not allowed inside email autolinks
#  - 609: a 1-char scheme is too short (CommonMark requires 2-32 chars)
606
609
