Regression tests: Canoscan LiDE 700F
Regression testing was performed on the Canoscan LiDE 700F. The following was discovered:
- BUG: Backend generates an I/O error on return to home after scan. If the return distance is very short, then the error does not appear. For -y between 0mm-13mm, return is fully successful. For -y between 14mm-73mm the scanner does not reach home at all: it falls about 10mm short. If -y is very large, then the scanner enters a weird state where it is stationary but buzzing, part way on its travel back home. The only way to recover is to power cycle the scanner.
- USABILITY: When starting to scan further down the page, using a non-zero value for -t, the scanner moves to the start scan position at scanning speed, rather then fast-forwarding. For some of the very high resolutions, this can make a very small scan take an extremely long time. Verified with the Windows driver that fast forwarding to the start position is possible.
- BUG: In a shell loop, running for multiple resolution scans, scanimage sometimes "hangs" at the beginning of a scan session. This is most often observed at the beginning of a 1200dpi scan. My guess would be that when the previous scanimage process terminated, the scanner was not yet quite ready to start another scan. This might be related to 1. above. Because of 1. above, I was restricted to scans that required only a very short return to home.
Functionality successfully tested:
- Button sensors: COPY, SCAN, FILE (PDF button), EMAIL (E-MAIL button)
- Lineart, Color and Gray tested.
- All resolutions tested for all modes.
- -x, -y, -l, -t options for positioning the scan.
- Other options tested: --color-filter (red, green, blue: verified visually), --depth (8, 16), --preview
Version 1.0.29
Edited by Ralph Little