A small but important bug-fix release.
🐛 Fixed
-
Images served from extensionless URLs are no longer broken or corrupted. References like
<img src=".../attachments/228/content">— common with OpenProject wiki attachments and similar CMS "content" endpoints — are now downloaded as raw binary, with the original image format preserved. Matchingsrcsetcandidates are rewritten to the local copy too, so responsive images keep working offline.Previously the crawler classified resources by URL file extension: an extensionless attachment URL was either skipped during discovery (leaving a broken link) or, in older versions, mistaken for an HTML page and corrupted during text decoding. (#49)
Thanks to @aanno for the detailed report.
Full changelog: v2.6.0...v2.6.1