Alex Pyrgiotis

Debian Trixie has started failing as well with the same error: https://app.circleci.com/pipelines/github/freedomofpress/dangerzone/2781/workflows/3660952c-d4ca-4b5b-a2a6-e2b41d417d95/jobs/39053?invite=true#step-106-1987_107

One factor that may contribute to this is the difference in the base images. The Fedora image differs if you build the development environment through Podman instead of Docker. In...

Well, turns out that FUSE is the culprit, when it's used to mount directories to the containers. If we move the `rpm-build/` directory outside the mounted volume, then the permissions...

This issue has bit us again. In order to fix it, I propose doing the following: 1. Detect in our RPM .spec file if the created package contains any files...

# On Packaging ## Windows The [official Tesseract site](https://tesseract-ocr.github.io/tessdoc/Installation.html#windows) points at the [Mannheim University Library](https://github.com/UB-Mannheim/tesseract/wiki) for the latest installers on Tesseract. These installers do not seem to be a solution...

We had a breakthrough in this front. Turns out that installing Tesseract-OCR is not a prerequisite for using the OCR capabilities of PyMuPDF. We were mistaken that this was the...

Added some extra tasks: > - [ ] Download OCR language data under `share/` for inclusion on Windows / MacOS, as well as when we do development. > - [...

## Design overview In this section, we'll explain how we plan to implement the on-host conversion feature. As a reminder, here's in pseudocode the current conversion flow: ```python p =...

# The Debian situation While implementing the above solution on #748, we realized that of all platforms, we have a big problem in Debian-based ones specifically. Here's a list of...

> I am curious why this is just a Debian/Ubuntu issue and not Fedora? Is their PyMuPDF package setup differently/better? Fedora enables the OCR feature in their PyMuPDF package, whereas...