You are not logged in.
My system is fully upgraded, and the wonderful OCR application tesseract-5.4.0-1, in the repositories since last 6 jun, does not work. It generates a “floating point exception” error, even on a simple image containing only a few lines of text. I had to downgrade to the previous version 5.3.4-2, which works correctly. I don't know if anyone else has detected the issue. Thanks in advance for your support.
Last edited by dulzamacis (2024-06-12 11:18:29)
Offline
Could you share some reproduction instructions and logs? I think this could be a problem that needs reporting upstream ![]()
Offline
OK. This is the image of the process that has failed: https://i.imghippo.com/files/LhI5s1718147369.png
And this is the core dump back trace file: backtraceInGdb.txt
Last edited by V1del (2024-06-11 23:44:05)
Offline
There's been an update to tesseract, can you reproduce with the new version? The trace matches: https://github.com/tesseract-ocr/tesseract/issues/4257 which should be fixed in 5.4.1.
Also please don't embed huge images and less so if the only information they contain is text, just post it as text.
Last edited by V1del (2024-06-11 23:48:30)
Offline
Thank you very much for the information. I have verified that the issue is solved with the new version.
Offline