Since PyMuPDF 1.23.9, upstream has provided a new fitz implementation for PyMuPDF. We were hesitant to use it for two reasons:
It was too late in our release to switch to a new PyMuPDF implementation, as we were concerned for any last minute bugs.
PyMuPDF was writing some logs to stdout (see #700).
Now that 0.6.0 is out, we can switch to the new fitz implementation, once we silence some problematic calls.
Note that while experimenting with the new PyMuPDF implementation, we needed some helpers that were not available in Dangerzone (debug logs from second container, faster way to build images), so we add those in this PR.
Since PyMuPDF 1.23.9, upstream has provided a new
fitz
implementation for PyMuPDF. We were hesitant to use it for two reasons:Now that 0.6.0 is out, we can switch to the new
fitz
implementation, once we silence some problematic calls.Note that while experimenting with the new PyMuPDF implementation, we needed some helpers that were not available in Dangerzone (debug logs from second container, faster way to build images), so we add those in this PR.
Fixes #700