signing a key on the commandline never completes

[anarcat]

Using the 1.0.1 debian package (self-built because not officially released yet), I cannot seem to sign a key properly. I have tried this:

$ gpg --export C4BC2DDB38CCE96485EBE9C2F20691179038E5C6 > dkg.gpg
$ gnome-keysign-sign-key dkg.gpg 
WARNING:keysign.gpgmeh:INV_SGNR: '9 7D4F37B1948ADAFF'
WARNING:keysign.gpgmeh:INV_SGNR: '9 586073B34023702F'

Then that pops open two terminal windows running the following command:

mutt mailto:dkg@debian.org?&subject=Your%20signed%20key%20C4BC2DDB38CCE96485EBE9C2F20691179038E5C6&body=Hi%20Daniel%20Kahn%20Gillmor%20%3Cdkg%40debian.org%3E%2C%0A%0A%0AI%20have%20just%20signed%20your%20key%0A%0A%20%20%20%20%20%20C4BC2DDB38CCE96485EBE9C2F20691179038E5C6%0A%0A%0AThanks%20for%20letting%20me%20sign%20your%20key%21%0A%0A--%0AGNOME%20Keysign%0A&attachment=/tmp/gnome-keysign-rn1qyx4m.asc

But while mutt normally is able to send email, for some reason, when it's from gnome-keysign, that doesn't work at all: mutt returns (strangely without sending any email) and gnome-keysign just hangs there. Interrupting with control-c yields the following traceback:

Traceback (most recent call last):
  File "/usr/bin/gnome-keysign-sign-key", line 11, in <module>
    load_entry_point('gnome-keysign==1.0.1', 'console_scripts', 'gnome-keysign-sign-key')()
  File "/usr/lib/python3/dist-packages/keysign/SignKey.py", line 45, in main
    input()
KeyboardInterrupt

It seems the program deliberately be waiting for my input, but never tells me so. Looking at the source, it seems that this might be a logging level problem. But looking at the logging tutorial, I would recommend explicitely using print for such prompts instead of the logging system.

[anarcat]

I opened #77 for the problems specific to mutt, but this remains an issue regardless of the mail client used.

[muelli]

oh, interesting.
Yeah, a print sounds more clever there. I wonder why we made it a logging statement in first place. Seems silly now. In fairness, the CLI app is rather auxiliary; but we still ought to make it as nice as possible.
Thanks for the report!

[anarcat]

On 2019-01-21 14:30:01, muelli wrote: oh, interesting. Yeah, a print sounds more clever there. I wonder why we made it a logging statement in first place. Seems silly now.

I've done the same mistake before. :)

In fairness, the CLI app is rather auxiliary; but we still ought to make it as nice as possible.

heh. I'm trying out the monkeysign replacements, and commandline use is one of my requirements. ;) `--debug` would certainly be useful, for example...

[muelli]

Thanks again for the report. Should be fixed with 0fc5abb.