Select OTP slots by name about nitrocli HOT 10 CLOSED

I think the first version should be minimal and include neither the scdaemon hack nor the clipboard access. (We already have the scdaemon problem with the otp subcommand. I don’t think we should treat the otpcache extension differently.)

I’ll personally use the extension with a wrapper script that opens dmenu, like passmenu. This wrapper script will probably contain the scdaemon hack and the clipboard handling. We could add this script to the repository too.

Anyway, I’ll continue to work on the draft and open a PR once it’s ready.

from nitrocli.

#146 has been merged (thank you!) so we can close this issue, right?

from nitrocli.

I do see why remembering slot numbers would be cumbersome. But truth be told, I haven't used the otp functionality much, so I can't really judge how bad it is to work with what we have.
That being said, I am not a fan of overlaying arguments with functionality as you are proposing. There rarely ever comes anything good out of that and it is pretty much always a source of bugs/unexpected behavior.

So my tendency would be to get rid of the means for addressing slots by number altogether (although off hand I can't say whether that is possible; there may be some bit that I am missing) and only work with names (well, perhaps add a way to disambiguate in case a name is ambiguous).

@robinkrahl do you have an opinion on this request by any chance?

from nitrocli.

We can only query the name for one slot at a time. So if we would remove access by number, we would have to execute n + 1 commands instead of only one command to retrieve an OTP code (where n is the number of used OTP slots). Therefore I don’t want to change the behaviour of the otp command.

As discussed in #50, I’d prefer an extension, let’s call it otpcache, that caches the names of the OTP slots and also provides an option to get an OTP code by slot name. So the first invocation of the otpcache subcommand would initialize the cache (per device). It could be updated with nitrocli otpcache update, and queried using nitrocli otpcache status or nitrocli otpcache get <name>.

from nitrocli.

Ah, thanks for refreshing my memory here, Robin. :-)

I agree with your inclination of making this an extension. Yet another reason to finally prioritize this work...

from nitrocli.

As the v0.4.0 release includes support for extensions (thanks!), we should be able to implement this feature. There is a prototype on the topic/otp-cache branch. Do you plan to continue working on this branch? Should I have a look at it?

from nitrocli.

Feel free to give it a go, Robin. It was your original proposal. It's working great for me, but it requires the killing of scdaemon to prevent smartcard lock ups as I am sure you recall. I am not sure we want to keep that, but at the same time without that or a different workaround the functionality basically has negative value for me (and I suspect for others).
Clipboard management is another issue: not sure how many ways there are to set the clipboard on Linux alone. But I suppose that does not have to be included in a first version.

from nitrocli.

I realized that we have the same problem with PWS, so I’d suggest to have a cache extension with otp and pws subcommands instead.

from nitrocli.

Sounds good!

from nitrocli.

Yep.

from nitrocli.