On 6/4/2014 1:35 PM, Antonio Diaz Diaz wrote:
I mean passing an option to configure as in:
./configure --enable-passthrough && make && make install
This way everybody else gets a normal ddrescue built.
I am not yet sure how to do that, but I would assume that it would still
involve using conditional compilation code like I mentioned. But I could
be wrong. I am sure it could be figured out.
But my patch will not do anything different unless the user actually
uses an option in the ddrescue command line. Just like --direct must be
manually selected, so would --scsi-passthrough and --ata-passthrough,
and I would not include an easy single letter that would enable it so it
would have to be fully typed out (could even add a
"yes-i-know-what-i-am-doing" switch for added safety). I would not even
consider making a patch that did not make the passthrough an option that
had to be intentionally selected.
Ddrescue already works like this. It is the kernel the one returning a
total block failure when it finds a bad sector. I wish the kernel
behaved better in case of read error. (Maybe some kernel version
behaves as you describe. I don't know).
We will definitely talk about this more when I am finally ready for it.
Scott
_______________________________________________
Bug-ddrescue mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-ddrescue
