[BreezeDelegate]

Fixes #153.

Formatting now stops on failed commands instead of reporting success, and refreshes the partition table before wipefs/mkfs so the new geometry is visible first.

Tested with failure-injection and an exFAT loop device (fsck.exfat -n, blkid).

[clefebvre]

Can you drop the udevadm settle line and keep only partprobe?

• partprobe is what makes the kernel re-read the new table, and it's now checked, so a failure there still stops the format. That seems to be the part that matters for USB Stick Formatter 1.6.6 creates invalid exFAT filesystems on multiple drives #153.
• --exit-if-exists=<partition> stops waiting as soon as the node exists, and the kernel creates /dev/sdX1 as soon as the partition is added. As I read it, the wait often returns immediately and doesn't wait for udev to finish anything.
• Since execute() now uses check_call, udevadm settle returns non-zero if the udev queue isn't empty within 10s. On a busy system that would abort a format that would have worked.

Could you check that #153 is still fixed with only partprobe added, and let us know if it isn't? If the wait turns out to be needed, a plain udevadm settle --timeout=10 that doesn't abort on failure would be better than a fatal one.

[BreezeDelegate]

Removed the udevadm settle call, leaving partprobe before wipefs and mkfs.exfat. I checked the command order and failure handling with tests: formatting stops if partprobe or mkfs.exfat fails.

I couldn't rerun the USB test on this host, so I haven't verified #153 on a device with this revision.