bug#28528: Split command problems.

[Assaf Gordon]

Hello,

On 2017-09-29 04:11 AM, Nick Farrow wrote:
> still with the same hash differences. 
>
> split -b 3276800 -d largefile largefile.

I'm unable to reproduce this issue.

I've tried the above command on two different machines,
one of which is an RPi3 similar to yours (armv7l with split version
8.23), the other an amd64 machine with split version 8.25 .

For completeness, these are the commands I've used:

    $ dd if=/dev/urandom of=largefile bs=1M count=100
    $ sha1sum largefile
    087fd7665a93875e69af2b9b2fd04aed6681b944  largefile

    $ split -b 3276800 -d largefile largefile.
    $ sha1sum largefile.* > hash-amd64

    [ copy 'largefile' to the RPi3 machine, snf run there: ]

    $ sha1sum largefile
    087fd7665a93875e69af2b9b2fd04aed6681b944  largefile

    $ split -b 3276800 -d largefile largefile.
    $ sha1sum largefile.* > hash-rpi3

The resulting two hash files (hash-amd64 and hash-rpi3) are identical in
my case - indicating the generated split files are identical on both
machines with using "split -b".

As an additional verification, concatenating the split files should
result in the same sha1 checksum as the original file, on both machines
(this should be true even if the split used differnt "-b X" value)

    $ cat largefile.* | sha1sum
    087fd7665a93875e69af2b9b2fd04aed6681b944  -


If you observe different result, please try the above commands
and let us know which ones differ (e.g. is the checksum of the input
file the same on both machines, is the checksum of the concatenated
files the same on both machines).


regards,
 - assaf