Bug#1100485: hopenpgp-tools: hokey canonicalize damages signature

[Clint Adams]

> 	+    Bad Signature:
> 	+                   Version: 4
> 	+                   Type: SubkeyBinding
> 	+                   Pk algo: RSA
> 	+                   Hash algo: SHA256
> 	+                   Hashed area:
> 	+                     Signature creation time: 2025-02-25 05:18:24 UTC (critical)
> 	+                     Issuer: 39CB544D6527CF60
> 	+                       Nicolas Pitre <nico at fluxnic.net> (UNAUTHENTICATED)
> 	+                     Notation: salt at notations.sequoia-pgp.org
> 	+                       00000000  1a 30 59 f3 ea fd 72 88  a3 2b 5e a5 1b e2 43 bd
> 	+                       00000010  89 d8 f6 37 92 11 28 a5  50 8d b1 af c8 e9 16 48
> 	+                     Key flags: S
> 	+                     Embedded signature:  (critical)
> 	+                                                Version: 4
> 	+                         Type: PrimaryKeyBinding
> 	+                         Pk algo: RSA
> 	+                         Hash algo: SHA256
> 	+                         Hashed area:
> 	+                           Signature creation time: 2025-02-25 05:18:24 UTC (critical)
> 	+                           Issuer: 7F4A62820BF463B7
> 	+                             Nicolas Pitre <nico at fluxnic.net> (UNAUTHENTICATED)
> 	+                           Notation: salt at notations.sequoia-pgp.org
> 	+                             00000000  d8 bd 36 7c ef bd c5 da  85 b8 f7 02 5d 3b 81 28
> 	+                             00000010  1b b8 e1 68 40 15 89 ec  b5 8b f0 eb d4 bb b0 f4
> 	+                           Issuer Fingerprint: E582CAEAF7CBA7AA04344A927F4A62820BF463B7
> 	+                             Nicolas Pitre <nico at fluxnic.net> (UNAUTHENTICATED)
> 	+                         Digest prefix: 4CA6
> 	+                         Level: 0 (signature over data)
> 	+                   Digest prefix: DB75
> 	+                   Level: 0 (signature over data)

It does look like only this signature is getting corrupted; replacing
that one in the hokey output with the original packet makes
`sq inspect` happier.

My suspicion is that something is horribly wrong with the
reserialization of the hashed embedded signature subpacket.