Missing DigestEngine operators?

[philipp]

I was thininking that the DigestEngine class needs some additional methods...

const std::string& toString();
bool operator!=(const DigestEngine& other);
bool operator==(const DigestEngine& other);

except of course that the subclasses of the DigestEngines need to be identical...

No point comparing the digest of an MD5 to the digest of an SHA1...

Oh, and was thinking that a SHA2 object might be nice, too.

[guenter]

What would be the point of comparing DigestEngine instances (instead of DigestEngine::Digest values)?

Regarding SHA2, it would be best to implement this in Poco::Crypto, probably in a generic way, using the generic OpenSSL digest functions (EVP_Digest...). Interestingly, I thought this had already been done, but it was wishful thinking on my part...

[philipp]

Err, sorry, yes... comparing Digests, not DigestEngines...

[philipp]

Oh, also comparing to a stringified digest, such that:

digest == digest.toString()

i.e.,

bool operator==(const std::string& otherDigest)
{
return (toString() == otherDigest);
}

[guenter]

I don't think comparison functions belong in the DigestEngine class. Conceptually it just makes no sense, at DigestEngine instances are not value objects, but rather objects creating values (digests), which can then be compared easily.

[philipp]

I don't think comparison functions belong in the DigestEngine class. Conceptually it just makes no sense, at DigestEngine instances are not value objects, but rather objects creating values (digests), which can then be compared easily.

Right, which is why I amended my statement as:

"Err, sorry, yes... comparing Digests, not DigestEngines..."

so it should be possible to compare a Digest and the serialized representation of a Digest (i.e. digest.toString()).

[killerbot]

is there already an idea when sha2 will be supported in poco ?