[Eapp1e]

Summary

An AsciiDoc listing or literal block opens on four or more delimiter characters — the Asciidoctor documentation names the length "minimum delimiter" and its is_delimited_block? accepts any uniform run. The backend tested only the exact strings ---- and ...., so a block with a longer marker was not opened: the markers stayed in the document as paragraphs and the code between them became prose.

$ python tmp/repro.py     # [source,python] / ----- / x = 1 / ----- / ..... / raw / .....
paragraph  text='----- x = 1 -----'
paragraph  text='....'  text='raw'  text='....'

Fix

Two regexes replace the exact-string tests:

-            if stripped in {"....", "----"}:
+            if _VERBATIM_DELIMITER_RE.fullmatch(stripped):
@@
-            if source_attr is not None and i + 1 < n and lines[i + 1].strip() == "----":
+            if (
+                source_attr is not None
+                and i + 1 < n
+                and _LISTING_DELIMITER_RE.fullmatch(lines[i + 1].strip())
+            ):
                 block_data = []
-                block_delimiter = "----"
+                block_delimiter = lines[i + 1].strip()

_VERBATIM_DELIMITER_RE is -{4,}|\.{4,}, and _LISTING_DELIMITER_RE is -{4,} because only a listing block takes a [source,lang] attribute. The closing check keeps comparing against the recorded opener, which matches Asciidoctor's behaviour of closing on an identical line. This mirrors the rule the same file already applies to comment blocks (_COMMENT_BLOCK_DELIMITER_RE = re.compile(r"/{4,}")).

Tests

tests/test_backend_asciidoc.py::test_delimiter_longer_than_four_characters_opens_the_block converts [source,python] + ----- + x = 1 + ----- and a ..... literal block, then asserts both become code items, that the language is Python, and that no marker text leaks into the body.

On the tree without the fix: 1 failed, 62 passed; with the fix: 63 passed.

$ python -m pytest -o addopts="" -p no:cacheprovider tests/test_backend_asciidoc.py -q

tests/test_backend_text_encoding.py (55 passed) and tests/test_input_doc.py (43 passed) are unchanged. No file under tests/data uses a delimiter longer than four characters, so no reference data changes.

Fixes #5044

[github-actions]

✅ DCO Check Passed

Thanks @Eapp1e, all your commits are properly signed off. 🎉

[mergify]

Merge Protections

🟢 Merge protection satisfied — ready to merge.

Show 1 satisfied protection

🟢 Enforce conventional commit

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

[github-actions]

AI triage (advisory)

Checked commit 3395211bb0. The analysis read the diff only. It did not run code from this pull request.

The PR fixes the AsciiDoc backend so that listing and literal block delimiters of five or more characters (e.g. -----, .....) open and close a verbatim block, matching Asciidoctor's rule that four characters is a minimum. Two module-level regexes replace the exact-string tests in _iter_blocks, and a focused test verifies the fix.

Topic labels added: bug, asciidoc.

CI safety: no concern found. A maintainer must still approve CI.

4 related pull request(s) in the same code

PR	State	Signal	Assessment
#5039	open	changes the same code	related: PR 5039 fixes caption parent placement for pictures, tables, and code blocks; it touches _add_picture_item and _add_literal_block, not the delimiter-matching logic changed here.
#5010	open	changes the same code	related: PR 5010 adds row-grouping by column count and cell-span parsing for tables; it does not touch the verbatim block delimiter logic changed here.
#5008	open	changes the same code	related: PR 5008 fixes _parse_picture so a leading named attribute is not read as alt text; it does not touch the verbatim block delimiter logic changed here.
#3891	open	changes the same code	related: PR 3891 fixes inline-group whitespace in _flush_text_data; it does not touch the verbatim block delimiter logic changed here.

_{The ai:* labels are advisory. They do not approve CI and they do not add tests:full. A new push runs this check again.}