Support *git-remote-gcrypt* helper

[kenny-evitt]

In a similar spirit to pass-tomb, I use Pass with the git-remote-gcrypt Git helper to encrypt the file and directory names (and the entire Git history too) of my Pass repos.

git-remote-gcrypt is implemented entirely as 'shell' (sh) scripts but the encrypted repository format seems pretty straightforward.

This isn't critical for me, I'm pretty sure (almost) no one else is using this Git helper with Pass, and I'm also confident that this is a low priority for the contributors to this project, so I'm willing to submit a PR (someday).

[mssun]

I'm pretty sure (almost) no one else is using this Git helper with Pass

Yes, I just heard from you.

About PR, you are welcome.

[neur0manc]

This seems to be an issue that can be safely closed until the PR from @kenny-evitt is in.

[yishilin14]

@ls42 Just marked this as an "enhancement". Hopefully one day, this issue could be closed by an awesome PR. :-)

[frihamn]

This would indeed be a great feature. In particular in light of recent meltdown / spectre vulnerabilities that raises questions regarding security of virtual servers for hosting sensitive git repositories.

[awehrfritz]

This would be a pretty important feature to have!

In fact not having the option to use git-remote-gcrypt or #114 is currently the only think keeping me from using passforios, where I would clearly favour git-remote-gcrypt due to the simplicity and easy of use.

@kenny-evitt: Did you get anywhere with implementing this?

[mssun]

Hi all, note that Pass for iOS is not using the git command line (which is not available in iOS obviously), but using the libgit2's Objective C bindings. Therefore, supporting git-remote-gcrypt is not as simple as in a Linux PC. One has to re-implement the encryption/decryption process himself to have this function. I understand it's a nice feature, but given current unresolved issues/features, I don't think it will be implemented in the near future unless someone familiar with the mechanism can help.

[awehrfritz]

Thanks for your swift reply @mssun!

I guess we are out of luck then. It seems that implementing pass-tomb, as suggested in #114, is even more difficult - if not impossible - since “pass-tomb is based on dm-crypt and therefore it is not compatible with Mac systems”.

[kenny-evitt]

@awehrfritz I haven't even started on implementing this, sadly.