I had occasion to check the dgst(1) manpage and found several minor flaws 
(in 1.0.1h, haven't tried 1.0.2 yet and don't do repository), a few of which

might benefit from discussion 

header and synopses: The list of alg names, (correctly) used as both
dash-options 
and (sub)command-names, also used for manpage links, is outdated.
md2 is no longer in the default config, sha224/256/384/512 gost whirlpool
are,
and dss1 is no longer needed (see more on this below).

The second synopsis should start >openssl< [md5|md4...] to be consistent,
and 
I think either it should repeat all the dash-options or (IMO better) add
ellipsis 
[-c] [-d] >...< [file...] to show all options except dash-digestname are the
same.

Description is incomplete. I suggest:
>The digest commands compute and display the digest(s) of a supplied file or
files,
or a Message Authentication Code, or generate and verify a digital signature
on a digest.<
I would also add here, or if not here in Notes:
>dgst handles only an RSA, DSA or ECDSA signature itself, not the related
data to 
identify the signer and algorithm used in formats such as CMS/SMIME and
X.509.<

-r is missing, should be: print in the coreutils format traditional for
separate programs 
like 'md5sum' or 'sha1sum' [only if hex] 

-non-fips-allow is missing, I think should be: enable use of non-FIPS
algorithms 
like MD5 even in FIPS mode 

-fips-fingerprint is missing, I think should be: compute HMAC using a
specific key 
for certain OpenSSL-FIPS operations 

-keyform applies to sign and verify and supports ENGINE PEM >and DER and
P12< 

-macopt on the website the first entry is misformatted (body run-in with
heading, all in bold)
but my man display is good. Maybe because dgst.pod line 120 between =item
B<key:string> 
and Specifies MAC key ... has a tab char whereas other 'blank' (empty) lines
don't.

Notes#1 is IMO stated a bit too strongly, and outdated. Suggest:
When given a choice, for a new or agile application, SHA-1 is the most
widely used digest 
but SHA-224/256/384/512 are also popular and considered more robust
long-term,
and there is still significant usage of other digests, particularly MD5.

Notes#2 is obsolete, since 1.0.0 switched to EVP_DigestSign* etc instead of
EVP_Sign*.

Notes#3 I think should say ECDSA as well as DSA.

END


______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [email protected]
Automated List Manager                           [email protected]

Reply via email to