Extract CPIO as a reusable module

[alexeymorozov]

@jordansissel

Would allow to use the CPIO parser outside of the project. For example, we need it in fontist.

What do you think? What can I do?

[jordansissel]

I'm open to having the CPIO parser available to other libraries. +1

[jordansissel]

That said, I don't think arr-pm uses the CPIO code anywhere. There's a few different CPIO formats (ascii, binary, etc) and I think I only implemented the ASCII one (used by RPM?). The arr-pm code extracting CPIO uses IO#popen to call the cpio program ;)

ruby-arr-pm/lib/arr-pm/file.rb

Line 101 in 5e35996

extractor = IO.popen("#{tags[:payloadcompressor]} -d | (cd #{target}; cpio -i --quiet --make-directories)", "w")

I'm thinking about a few different ideas here:

1. You copy the cpio parser code into your repo?
2. Instead, maybe use libarchive which supports different CPIO encodings? (https://github.com/chef/ffi-libarchive has ruby bindings)
3. Merge this PR and push a new arr-pm

[alexeymorozov]

Agree, my research shown the same: arr-pm doesn't currently use this implementation of CPIO extraction.

Considering of the ideas:

1. Made exactly this way, copied to https://github.com/fontist/excavate/tree/main/lib/excavate/extractors/cpio as a new format (the ASCII one)
2. Our library strives to be cross-platform, butlibarchive requires to have compiled binary in system, and there is no such library on Windows by default. So it's a hard option, currently we skipped it.
3. If there is anyone else who needs the feature, it can be helpful.
4. Also we may extract the CPIO implementation as a new gem? 🤷‍♂️