See what the CSS tests render¶
Find a layout or CSS feature, compare the printed pages, and download the actual Fullbleed PDF. The default view shows a maintainer-run comparison of the published Fullbleed 2.5.43 Linux wheel against the pinned, independent IronPress corpus.
Choose a package version to inspect its own results. Shared case links keep the selected version, filters and fixture. Earlier 2.5.39 results, 2.5.34 results, and 2.5.22 results remain available with their evidence records.
The comparison gate fails. Passing fixtures meet the comparator's policy; they do not establish complete CSS support or pixel-identical output. Failures and disputed references retain their original labels. Read the method and limits.
The description labeled “Upstream fixture” is the corpus author’s original text. Its support notes describe upstream test context; use the selected Fullbleed result and saved document to assess this version.
Loading the fixture index…
Read a result¶
- PASS: the fixture meets the unchanged upstream comparison policy.
- FAIL: a comparison or rendering check failed. Inspect the actual pages; the label alone does not identify the cause or establish missing content.
- REFERENCE-DISPUTED: the candidate differs from a reference marked disputed by the upstream corpus. These cases are not counted as passes.
The reference's own status is shown separately, including when a fixture with a disputed reference receives a PASS verdict. The images use the original report's rasterization, and each PDF download preserves the tested bytes. Reference notes are upstream review text; mentions of IronPress in those notes describe that project's reference review.
The corpus uses static HTML/CSS and explicit assets. Its results do not establish support for every CSS feature, arbitrary websites, browser JavaScript, accessibility standards, or PDF conformance. Try representative data and your own fonts in the HTML/CSS playground before choosing a layout.