dschmidt opened a new pull request, #3540: URL: https://github.com/apache/tika/pull/3540
[TIKA-5036](https://issues.apache.org/jira/browse/TIKA-5036), SketchUp part (Blender is #3539). SketchUp stores a PNG preview in every `.skp` file. `SketchUpParser` (cad module) emits it as the `THUMBNAIL`. The file starts with "SketchUp Model" and the version as length-prefixed UTF-16 strings; after that, the preview sits in one of two places: - up to SketchUp 2020, in a `CDib` record about 2 KB into the file: the record's length fields, then the PNG; - from SketchUp 2021, in a zip that starts a few bytes after the version. Its `meta/model_thumbnail.png` shows the model alone on a transparent background and is taken; `meta/preview_thumbnail.png` (the view with background and axes) is the fallback, as some files lack it. `application/vnd.sketchup.skp` is new, detected by the header. No SketchUp type is registered anywhere; this is the one file(1) has used since 2018 for the same header. Checked against 107 real files from public repositories, SketchUp 3 to 2026: all are detected, with and without a file name, and each thumbnail matches, byte for byte, the PNG an independent extraction finds. An open question on the `*.skp` glob: Tika maps it to `application/vnd.koan` (SSEYO Koan), as IANA does, and two types cannot share a glob. This PR leaves it there, so a SketchUp file is recognized by its content, and a `.skp` without the header is still Koan. Moving the glob to SketchUp, keeping it with Koan, or something else entirely: I don't have a strong opinion and would go with whatever you prefer. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
