从gzip文件的开头剥离bash脚本

Question

I have a series of files which are comprised of a bash script, at the end of which a gzip file has been concatenated.

I would like a method of stripping off the leading bash, to leave a pure gzip file.

The method I have come up with is to:

1. Do a hex dump on the file;
2. Use sed to remove everything before the gzip magic number 1f 8b ;
3. Convert the remaining hex dump back to binary.

ie

xxd -c1 -p input | tr "\n" " " | sed 's/^.*?1f 8b/1f 8b' | xxd -r -p > output

This appears to work okay on first glance. However, it would fall apart if the gzip portion of the file happens to contain the byte sequence 1f 8b apart from in the initial header. In these cases it deletes everything before the last occurrence.

Is my initial attempt on the right track, and what can I do to fix it? Or is there a much better way to do this that I have missed?

Answer 1

Perl solution. It sets the record separator to the magic sequence and prints all the records except the first one. The magic sequence must be prepended at the beginning, otherwise, it would be lost together with the bash script, which is the first record.

perl -ne 'BEGIN { $/ = "\x1f\x8b"; print $/; } print if $. != 1' input > output.gz

Answer 2

I would use the sed line range functionality to accomplish this. -n suppresses normal printing, and the range /\\x1f\\x8b/,$ will match every line after and including the first one with \\x1f\\x8b in it and p rint them out.

sed -n '/\x1f\x8b/,$ p'

Alternatively, depending on your tastes, you can add a text marker "### BEGIN GZIP DATA ###" and delete everything before and including it:

sed '1,/### BEGIN GZIP DATA ###/ d'