I would like to do an MD5 checksum on a directory in linux. Problem is that the checksum says that whitespace is makes the check fail. How can I ignore white space when doing a md5sum check.


I generate chksum file like so

find /tmp/ -type f -name * -exec md5sum {} >> tmp.chksum \;


then check other side on different server

md5sum -c tmp.chksum

There is a space which is being counted as difference.

How can I do this. Should I use another tool

where is this space located exactly?

The original file would be

(test.file)
hello

THe file im testing against will be
(test.file)
hello
#new line here

so the files are not identical, but one of them contains an additional new line. You need to remove it before calculating md5sum. I can't see other ways, but probably someone has an idea.
you can use sed/awk/perl/python or something else to remove that char. Using perl or python you may also try to read the file, cut the last char and calculate md5sum "in one".

Huh ?.
They are different - why do you expect them to resolve to the same ?. Doesn't make any sense. Nor does "munging" the file before md5 - ask an auditor what s?he thinks about that. Whoa.

Hi andrew6166 and welcome to LQ.

As others are saying, there is no way to ask an MD5 checksum to ignore whitespace, newlines (which is not whitespace by the way), or other characters.

The long winded solution would be to download the MD5Sum source and write your custom form of it.

I do not recommend doing a special, one-of solution for this.

Please realize that the whole point of the MD5 checksum is to guarantee that both files are absolutely identical in all respects.

You may wish to consider diff(1), please check out the -b and -w options, as well as explore the wide variety of available options for that command.

1 members found this post helpful.