[Bug-ddrescue] Sector-sized rounding

[????????? ???????????]

Hello, Antonio.

I've got one more idea: add an option to round both -s and -i to greater or 
lower values, rounded up or down to the sector size given by -b<value> ( 
--read-less --read-more ???).
For example, if run -b4096 -i100k -s10k --read-less
- the program will read from position 102400 up to 110592 (8192 bytes).
For -b4096 -i100k -s10k --read-more
- the program will read from position 98304 up to 110592 (12288 bytes).
Or, add this functionality to the --direct option.
As you adviced, I wrote my previous ideas to the address@hidden
This will also be posted there.

> BTW, I think you should have written to address@hidden so that 
> other users can see your ideas and give feedback about them.
> 
> 
>> 1) Please, add an option to quit or, better, to pause the program or
>> to reset/reinit/power_cycle the input hard drive when skipping slow
>> areas. If made to quit - I can use a shell script to pause and
>> restart the copying (in cycle).
> 
> Quitting is the most flexible option. As you say, it allows to script 
> whatever action the drive may need.
> 
> Or maybe the syntax of --min-read-rate could be extended to:
> 
>   --min-read-rate=<bytes>,<pause>
> 
> where <pause> is the amount of seconds to pause after each slow read. A 
> <pause> of 0 could make ddrescue quit.
> 
> 
>> 2) The second my question is about using direct mode on disks with advanced 
>> format:
>> -d (direct mode) did not work in any of the combinations:
>> -d -c1
>> -d -b4096 -c1
>> -d -c8
>> -d -c1 -b4096
>> Both input and output devices have native sector sizes of 4kb (emulated to 
>> OS as 512 bytes) and size of 3000Gb.
> 
> I can't reproduce the problem. I do not own any device with 4096-byte 
> sectors, but all your commands work fine on my 512-byte devices.
> 
> If you can find the cause, then I could fix the problem (if it is a 
> problem in ddrescue).
> 
> 
>> 3) Can you add (return/implement back) the option of
>> log_saving_time_interval or just make it static, about 30 seconds.
>> Sometimes the system hangs - and some progress is lost.
> 
> Ddrescue never had a 'log_saving_time_interval' option, but such option 
> can be implemented. Making it static at 30 seconds is not a good idea 
> because saving a big mapfile to a slow device may take a significant 
> part of that time, slowing the rescue.
> 
> 
>> Have seen new versions (up to 1.22pre) ! Maybe, emergency save of the
>> mapfile will partially help with my question number 3, to visualize.
> 
> I don't think so.
> 
> 
>> 4) I can translate the program and documentation to Russian - will it
>> be helpful for the project?
> 
> Thanks. I warn you that translating the manual is en enormous amount of 
> work, which systematically becomes out of date because nobody wants to 
> keep translating new versions. See the lzip French manual:
> 
> http://www.nongnu.org/lzip/manual/lzip_manual.html    (lzip 1.18)
> http://www.nongnu.org/lzip/manual/lzip_manual_fr.html (lzip 1.11)
> 
> But it would be nice to have the web page[1] translated to Russian. It 
> is already translated to Spanish, French and Italian.
> 
> [1] http://www.gnu.org/software/ddrescue/ddrescue.html
> 
> 
>> 5) Is the following implemented?
>>> The only safe solution, in addition to leaving direct mode as it is now, 
>>> would be to modify direct mode so that it does not attempt partial 
>>> sector reads.
> 
> Yes, it was implemented in ddrescue 1.20:
> 
> http://www.gnu.org/software/ddrescue/manual/ddrescue_manual.html#Direct-disc-access
> "NOTE: Direct disc access can copy arbitrary domains by reading whole 
> sectors and then writing only the requested part. This is the only case 
> where ddrescue will try to read data outside of the rescue domain."
> 
> 
> Best regards,
> Antonio.
>