When I run gocr on an image that has pixels of all the same color, it dies with a floating point exception (division by zero). Reproducible: Always Steps to Reproduce: 1. gocr -i /tmp/bad.pnm Actual Results: # thresholdValue out of range 255..255, reset to 255 Floating point exception Expected Results: Graceful exit.
Created attachment 127098 [details] Example image that causes the problem.
Created attachment 127099 [details, diff] Patch that fixes the problem.
Should this really be assigned to TeX herd?
Read the metadata.xml
nice, bug confirmed. But, has this been sent upstream ? I'd prefer adding the patch as a backport rather than a fix as it changes the (bad) behavior of the program
Submitted bug upstream. Sorry for being lazy and hoping that someone else would do it.
This has been fixed upstream in 0.47.
(In reply to comment #7) > This has been fixed upstream in 0.47. bumped, thanks