gotenberg/gotenberg v9.1.0

9.1.0

on GitHub

New Features

• extract-attachments (#1718). /forms/pdf/extract-attachments returns the files a PDF embeds, every one or the one name gives, as <pdf>_<attachment>: invoice.pdf gives invoice_factur-x.xml. Several files come back in a ZIP. A PDF without attachments, or without the named one, answers 400. Thanks @Betelgeuse1.
• extract-text. /forms/pdf/extract-text returns the text of each PDF as a .txt file, pages joined by a form feed. pages selects pages; invisible, artifacts and annotations add invisible text, artifacts such as headers and stamps, and annotation text. Gotenberg does no OCR: a scanned page reads as empty.

Both are stages too: they come last in a stages list, after encrypt, sign or timestamp if any, and the response holds the PDFs and the extracted files.

Bug Fixes

• backdrop-filter: blur() rendered blank on linux/ppc64le (#1491). The ppc64le image pinned Chromium 146, which has the bug; it now installs the same Chromium as the other platforms. Thanks @florianaiautomation.
• Data race on concurrent stages. Stages that write many files at once, such as split or render on several PDFs, could register their output names concurrently.

Chore

• linux/ppc64le moves from Chromium 146 to 156, from Debian's security updates.
• Updated Go dependencies.