Read concatenated output of tgz archive (with plain text files), with filenames, without unpacking in bash?

Question

I have a bunch of text files/logs, which I'd like to pack into a gzipped tar archive, say:

cd /tmp
echo "My Content One" > aaa.log
echo "My Content OneA" >> aaa.log
echo "My Content Two" > bbb.log
echo "My Content TwoB" >> bbb.log
tar czvf test.tgz *.log

Now, if I open this with less test.tgz, I get a listing of files same as tar tzvf test.tgz would give me:

-rw-rw-r-- user/user 31 2017-10-29 03:10 aaa.log
-rw-rw-r-- user/user 31 2017-10-29 03:10 bbb.log

First, I found https://www.serverwatch.com/tutorials/article.php/3798511/Reading-Compressed-Files-With-less.htm, and tried setting up in my ~/.bashrc:

 alias tgzless='LESSOPEN="|tar --to-stdout -zxf %s" less'

Now I can do tgzless test.tgz, and this prints the concatenated contents of all files:

My Content One
My Content OneA
My Content Two
My Content TwoB

... but unfortunately doesn't print the filenames, so I don't really know which part belongs where.

Then, I found How can i view the contents of a tar.gz file (filenames + filesize), and tried this command:

gzip -dc test.tgz | less

Now, this will print both a file heading, and the content of the file - unfortunately the file heading is binary and difficult to read - it looks something like this inside less:

aaa.log^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@0000664^@0001750^@0001750^@00000000037^@131752
34246^@014636^@ 0^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ustar  
^@user^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@user^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@My Content One
My Content OneA
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
...

... so it is hard to read (note that gzip -dc test.tgz | tar -tf - | less gives just a listing of filenames, though less verbose one than tar tzvf)

So can I set up an alias somehow using these standard commands, that would print all filename contents in a tar.gz archive, such that first the filename of the file is printed (preferably in tar tzvf format, with username, permission and timestamp), and then its contents?

Answer 1

Assuming you’re using GNU tar, create a file named tarcat on your path with the following contents:

#!/bin/sh

# Permissions user/group size date time name
printf "%s %s/%s %7d %s %s\n" "${TAR_MODE}" "${TAR_UNAME}" "${TAR_GNAME}" "${TAR_SIZE}" "$(date -d @${TAR_ATIME})" "${TAR_FILENAME}"

exec cat

Then run tar x --to-command=tarcat -f test.tgz | less:

0664 user/user      31 Sun 29 Oct 03:10:00 UTC 2017 aaa.log
My Content One
My Content OneA
0664 user/user      31 Sun 29 Oct 03:10:00 UTC 2017 bbb.log
My Content Two
My Content TwoB

Decoding the octal file mode and adapting the date/time output is left as an exercise for the reader.