[Bug-ddrescue] Bug in cygwin version 1.21 of ddrescue and a proposition to differentiate errors

[Joe Kickman]

Hello, Antonio!
1. I've found a bug about device names shown in the version 1.21 of ddrescue IN 
CYGWIN:
...\>ddrescue.exe /dev/sda /dev/sdb --ask ./copylog2.log -s1s -f
GNU ddrescue 1.21
About to copy 512 Bytes from /dev/sda [WDC WD1600JS-00NCB1] to /dev/sdb [WDC 
WD1600JS-00NCB1WDC WD1600JS-00MHB0].
Proceed (y/N)?^C

The second device has the name of the first, then its own.
Native Linux versions never had such bugs.

2. Is it possible to differentiate read errors and add an option to wait for 
the source drive to become ready again? For example, copying 160 Gb from a 
badly damaged seagate drive often caused freezes. It helped to power cycle the 
drive and resume copying from a little bit  further position [+ or -]6000 LBA. 
So, an option to --wait-ready <seconds_until_exit> or so together with 
-K<skip-size> would be very helpful:
ddrescue /dev/sdb ~/usbdrive/driveclone.img ~/usbdrive/clone-disk.log 
--wait-ready <20> -K6144s

Status: drive not ready, waiting: 20, 19 ,18, 17, 16, 15, 14, .. .. .. .. ..

It will allow not to restart ddrescue each time the drive self-disconnects or 
being power-cycled.

3. How the -K option works? Either it often misunderstood my input such as 
-K102400s or -K2000000s and set skip size not as I expected - or I have 
forgotten that used -b4096 instead of default -b512 ?? What is the right syntax 
for it, which units does it accept?

Thanks.
______________
"Joe Kickman"