LinuxQuestions.org - GPG - encode file while manipulating output file name

- Linux - Newbie (https://www.linuxquestions.org/questions/linux-newbie-8/)

- - GPG - encode file while manipulating output file name (https://www.linuxquestions.org/questions/linux-newbie-8/gpg-encode-file-while-manipulating-output-file-name-754839/)

GPG - encode file while manipulating output file name

I'd want to encrypt some files with GPG, named, as example "1.txt" and "2.txt".

So far so good, I can use something like:

Code:

find /tmp -name '*.txt' | xargs -I {} gpg -r test@test.com --output {}.enc --encrypt {}

Problem is, that I want to add the date to the filename.

Now, with a quite awkward stat+sed I can generate the filename (excerpt):

Code:

stat -c%n%y {} | sed -r 's/(.*).txt(\w+)-(\w+)-(\w+).*/\1_\2\3\4.txt/'

(note: the '.' in '.txt' is inappropriate but acceptable)

How can I incorporate this last statement as output file name in the find/xargs/gpg output parameter?

If I do something like:

Code:

find /tmp -name '*.txt' | xargs -I {} gpg -r test@test.com --output $(stat -c%n%y {} | sed -r 's/(.*).txt(\w+)-(\w+)-(\w+).*/\1_\2\3\4.txt/') --encrypt {}

The $() inputs the {} literally.

This won't work either:

find /tmp -name '*.txt' | xargs -I {} sh -c "cat {} | gpg -r test@test.com > $(stat -c%n%y {} | sed -r 's/(.*).txt(\w+)-(\w+)-(\w+).*/\1_\2\3\4.txt/')"

Is there a simple[r] oneliner for doing this operation?

Thanks!
Saverio

Quote:

Originally Posted by itmozart (Post 3681105)

I'd want to encrypt some files with GPG, named, as example "1.txt" and "2.txt".

So far so good, I can use something like:

Code:

find /tmp -name '*.txt' | xargs -I {} gpg -r test@test.com --output {}.enc --encrypt {}

Problem is, that I want to add the date to the filename.

Now, with a quite awkward stat+sed I can generate the filename (excerpt):

Code:

stat -c%n%y {} | sed -r 's/(.*).txt(\w+)-(\w+)-(\w+).*/\1_\2\3\4.txt/'

(note: the '.' in '.txt' is inappropriate but acceptable)

How can I incorporate this last statement as output file name in the find/xargs/gpg output parameter?

If I do something like:

Code:

find /tmp -name '*.txt' | xargs -I {} gpg -r test@test.com --output $(stat -c%n%y {} | sed -r 's/(.*).txt(\w+)-(\w+)-(\w+).*/\1_\2\3\4.txt/') --encrypt {}

You know, writing scripts should not be an obfucscation contest. Maybe it would be better to stop trying to see whether any problem can be solved using a long, incomprehensible command line. Please reconsider, open a script file and type some clear, comprehensible instructions:

Code:

path="/path/of/interest"



find $path | egrep "\.txt$" | while read inpath

do

  date=`date`

  date=${date// /_}

  outpath="${inpath//.txt}_${date}.enc"

  gpg --encrypt-files $inpath --recipient test@test.com --output $outpath 

done

Notice that your one-liner, once it actually worked, would be longer than this formal, transparent, easily understood, solution to your problem.

And please understand that the ultimate victim of miraculous one-liners is ... you, six months from now, when you have to try to decipher what the one-liner actually does.

I will never understand why people insist on using "xargs" and then can't understand why it won't give them what they want. The method I posted above takes longer to run (because it creates a separate process for each file), but overall it's faster because it saves development time.

One more comment. In the future, when parallel processing is fully mature, methods like this (that spawn one process per file) will actually run faster than the alternatives.

Quote:

Originally Posted by lutusp (Post 3682002)

Well, ok, I give up the obfuscation contest ;-)
One lines had some advantages, but in this case you're right, it's pushing too further.

Thanks!
Saverio