Page MenuHomePhabricator

Make the Conduit auth error for an unrecognized public key a little more useful
ClosedPublic

Authored by epriestley on Jul 17 2018, 11:06 PM.
Tags
None
Referenced Files
Unknown Object (File)
Tue, Nov 19, 9:47 PM
Unknown Object (File)
Sat, Nov 16, 1:38 PM
Unknown Object (File)
Tue, Nov 12, 3:29 AM
Unknown Object (File)
Fri, Nov 8, 5:37 AM
Unknown Object (File)
Oct 22 2024, 8:07 AM
Unknown Object (File)
Oct 11 2024, 2:46 AM
Unknown Object (File)
Oct 7 2024, 9:22 PM
Unknown Object (File)
Oct 7 2024, 9:22 PM
Subscribers

Details

Summary

Ref T13168. This is just a small quality-of-life fix: we can disclose which public key we're talking about because public keys are public.

Test Plan
  • Hit public key error (through my own bumbling / not reading or following instructions). Specifically, I haven't associated the key with a device in Almanac.
  • Before: vague error.
  • After: more specific error with enough key material that I could grep for it.

Diff Detail

Repository
rP Phabricator
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

Isn't the end of the public key way more useful for human readability in general? Most of my keys generated via ssh-keygen (which I assume is the practically universal method?) end in an identifier that I can plausibly use to identify the key, like this one:

ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQDzzrWAoWk0QFKtGz3860178qb5ue1cMSqrwYnBOJah7SXEUfwCOjXtvfvE3hyShS3SBdhm6qJZfSMngq7jO84Je1T7HU9FFjLg9HUPsfXaPtsojg4HhhN2n1dm+SysdmXScqIDb0g5/JWHenHTRh146t7ptP6+JzSHQGCWBzXsgXFCdlRrcrM+cUIGz9A0AIr9rZBgcwRtp8qK+TG9geA+Vcxf2Lw2T/GUS4nAeuT8KslAY6oC7ge2zT33VUhkvHyWV4iNlTdDHFneMvtdyXiCFa0oZwaQ7PDNw6Hrde7TDOkTi9Pppxawi8UYotsL7CzkeNIRWDNGXGTLyVsxHYBX mriley@mriley-C02TW3N0HTDD

(C02TW3N0HTDD is the very useful totally human interpretable name of my laptop)

Isn't the end of the public key way more useful for human readability in general?

I was going to say the same thing:

~ $ grep -r AAAAB3NzaC1 .ssh/ | wc -l
      18

64 leading characters is definitely unique enough though. Accepting as-is, but I agree with @yelirekim that the last 64 chars would be more useful.

This revision is now accepted and ready to land.Jul 19 2018, 3:58 AM

I think structuring it as The public key ending in "...<xyz>". is at least a little weird, and potentially confusing even if it's technically better in most cases.

The comment may also be arbitrarily long and non-unique. (When we generate keys, I think we currently just use the comment "Generated", which in a German localization would of course be "Gensherkeymatthaffenuntermazerhosenauomattermachinen", making the last 64 characters nonunique and nonidentifying.)

Perhaps best would be "aaa ... bbb" but we don't have a convenient UTF8-aware truncation process for that today.

(Practically, we don't have a "...xxx" truncator either, and we can't just strrev() the input and output because we don't have a UTF8-aware string reverser.)

I suppose reversing phutil_utf8v_combined() is a reasonably good approximation of a UTF8 reverser. It's probably not perfect (e.g., RTL marks should be inverted, I guess?) but if we were to implement phutil_utf8_strrev(), reversing phutil_utf8v_combined() would probably be a reasonable v0.

This revision was automatically updated to reflect the committed changes.