ssh-keygen -Ql: avoid repeated SHA256: in output

[Lucas]

Lucas <lucas at sexy.is> wrote:
> If a key is revoked by public key, when printing it, ssh-keygen adds a
> hardcoded "SHA256:" prefix and then prints the result of
> sshkey_fingerprint, which does already include that "SHA256:" prefix.
> Patch below, with extra context lines.
> 
> Also, while verifying this didn't mess up the output for other formats,
> I realized that if you create a KRL with an input file like
> 
> sha256: ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIDx8M2XyhHYN9SS1JzTXNipKfLN2P8ybBqi9S16o1sZs
> 
> or
> 
> hash: SHA256:K2Qu+94rdYJHJ62DTON67p5YejCsW4j3tA6fuTEpMmo
> 
> when printed, it gets printed as
> 
> hash: SHA256:2b642efbde2b75824727ad834ce37aee9e587a30ac5b88f7b40e9fb93129326a
> 
> which is quite unusual. In particular, ssh-keygen itself isn't able to
> parse that kind of lines. I know that ssh-keygen -Ql output aims to be
> *similar*, but not necessarily compatible, with the input format. Also
> I didn't find any "tob64" function to replace "tohex" call.

Bump.

diff c70493106c794797cfb174ed7460065b1a4a89f7 /usr/src
blob - 1975ae6005abdad42436f049c7ed5511c66ea151
file + usr.bin/ssh/krl.c
--- usr.bin/ssh/krl.c
+++ usr.bin/ssh/krl.c
@@ -1382,15 +1382,15 @@ krl_dump(struct ssh_krl *krl, FILE *f)
 		}
 		if ((fp = sshkey_fingerprint(key, SSH_FP_HASH_DEFAULT,
 		    SSH_FP_DEFAULT)) == NULL) {
 			ret = SSH_ERR_INVALID_FORMAT;
 			error("sshkey_fingerprint failed");
 			continue;
 		}
-		fprintf(f, "hash: SHA256:%s # %s\n", fp, sshkey_ssh_name(key));
+		fprintf(f, "hash: %s # %s\n", fp, sshkey_ssh_name(key));
 		free(fp);
 		free(key);
 	}
 	RB_FOREACH(rb, revoked_blob_tree, &krl->revoked_sha256s) {
 		fp = tohex(rb->blob, rb->len);
 		fprintf(f, "hash: SHA256:%s\n", fp);
 		free(fp);