https://bugs.kde.org/show_bug.cgi?id=524001
Noah Davis <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |UPSTREAM Status|CONFIRMED |RESOLVED --- Comment #3 from Noah Davis <[email protected]> --- (In reply to mwyeatts from comment #2) > I'm not sure if this is due to Arch not including that config file in the > package, and the proper fix is to add it? Or maybe Spectable can pass that > flag for certain languages? In https://invent.kde.org/plasma/spectacle/-/merge_requests/589, we determined that this bug was caused by a bug in tessdata_best, which tessdata's latest models are based on. Arch Linux distributes tessdata while Debian, Fedora, openSUSE and Ubuntu distribute tessdata_fast, which explains why I couldn't reproduce the bug on openSUSE Tumbleweed. Official documentation says this: > Most users will want tessdata_fast and that is what will be shipped as part > of Linux distributions. The bug should probably be fixed in tessdata_best instead of Spectacle and Arch Linux should probably switch to tessdata_fast. There is already a bug report for this issue on Tesseract's side. If you want, you can report a bug to Arch Linux, requesting that they switch from tessdata to tessdata_fast. Relevant links: - https://github.com/tesseract-ocr/tesseract/issues/2702 - https://github.com/tesseract-ocr/tessdata_best - https://github.com/tesseract-ocr/tessdata - https://github.com/tesseract-ocr/tessdata_fast - https://tesseract-ocr.github.io/tessdoc/Data-Files -- You are receiving this mail because: You are watching all bug changes.
