29623138702
60%

Ran 18 Jul 2026 12:28AM UTC

Jobs 1

Files 30

Run time 1min

Badge

Embed ▾

Committed 17 Jul 2026 08:09AM UTC coverage: 59.83%. Remained the same

Build # 29623138702

Build Type

push

github

Committed by

web-flow

Commit Message

fix: prevent MarkdownHeaderLevelInferrer regex from consuming newlines (#481)

The header pattern used `\s+` between the `#` characters and the heading
text, which matches any whitespace including newlines.  When a header line
contained *only* trailing whitespace (e.g. `## `) the regex engine could
consume the trailing space *and* the following newline in one `\s+` match,
then lazily expand `(.+?)` over the next header line.  This caused a
whitespace-only pseudo-header and the following real header to be treated as
a single match, rewriting `## \n## Section\nContent` to the invalid
`# ## Section\nContent` (a hash character embedded inside the heading text).

Fix: replace `\s+` / `(?:\s*)` with `[ \t]+` / `(?:[ \t]*)` so that
the pattern is restricted to horizontal whitespace only and cannot span line
boundaries.  All 14 existing tests continue to pass; two new regression tests
cover the fixed cases.

Co-authored-by: JSap0914 <JSap0914@users.noreply.github.com>
Co-authored-by: Stefano Fiorucci <stefanofiorucci@gmail.com>

Coverage Stats

423 of 707 relevant lines covered (59.83%)

0.6 hits per line

Coverage Regressions

Lines	Coverage	∆	File
2	96.88	0.0%	components/preprocessors/md_header_level_inferrer.py

Jobs

ID	Job ID	Ran	Files	Coverage
1	29623138702.1	18 Jul 2026 12:28AM UTC	30	59.83	GitHub Action Run

deepset-ai / haystack-experimental / 29623138702
60%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 29623138702

deepset-ai / haystack-experimental / 29623138702 60%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Coverage Regressions

Jobs

Source Files on build 29623138702

deepset-ai / haystack-experimental / 29623138702
60%

README BADGES
x