Open mattock opened 7 years ago
I modified the code a bit to get shellcheck and BATS tests passing, could you please re-test?
@mmoll : the logic on/after line if GPG_TTY=$(tty)
seems the exact opposite of what it should. As in: if GPG_TTY is set, then there's no need to export it. I tested this just in case an indeed the old problem is back.
if GPG_TTY=$(tty)
should only guard from aborting when tty
fails. With the original code the BATS tests were broken, probably because the travis envrionment doesn't provide a working tty.
Im puzzeled, why it doesn't work as expected.
However, while testing in my env (debian 9 wityh systemd), I'm getting the following: If I'm exporting GPG_TTY in my shell directly, everything works, but when using freight-cache, even with the original one-liner fix, GPG_TTY is "not on a tty" and gpg fails again... which is bringing me back to the question of GH-72 if the whole siuation can really be work-arounded automatically...
@mmoll : by "when tty fails" you mean "the tty executable is not present/found" or "executable outputs an empty/unexpected string"?
I also had the "not a tty" problem earlier, but I'm no longer sure how to reproduce it.
@mattock "tty fails" -> non-zero exit code
URL: https://github.com/freight-team/freight/issues/72 Signed-off-by: Samuli Seppänen samuli@openvpn.net