Timo Kokkonen

I don't think DPI information is part of the JPEG image. So it likely comes from some marker. jpegoptim normally displays markers found in the image when its run, so...

Seems like bug in MemorySanitizer, doesn't seem to be any "use-of-unitialized-value" going on here? ``` 374 if (cmarker->marker == EXIF_JPEG_MARKER && cmarker->data_length >= EXIF_IDENT_STRING_LEN) { (gdb) print *cmarker $5 =...

After looking into this bit more, it would seem like this is "issue" with libjpeg (or its just issue with MemorySanitizer....) Call to _jpeg_read_header(&cinfo, TRUE)_ succeeds, but it ends up...

Something like this might work better: ``` find . -type f \( -iname "*.jpg" -o -iname "*.jpeg" \) -exec jpeginfo -c --csv {} \; | grep -E "WARNING|ERROR" | cut...

@Summertime, thanks for reporting this. Clearly a bug in the retry logic. Are you able to test if following change fixes this? #210 Initial fix is found in the ```automode_fix```...

Man page is included in the repository: https://github.com/tjko/jpeginfo/blob/master/jpeginfo.1 I added note about the exit code in the man page (https://github.com/tjko/jpeginfo/commit/6bcb601dd063102d4654f2724681cdd156561c1e)