Incorrect sha512 values vs node 'crypto' and golang sha512

[tobowers]

Hi,

I'm having trouble getting a correct sha512 out of forge. I keep thinking it must be encoding issues, but I have some pretty straightforward code generating different results.

Data used (hex) 0a102f696e736161736974792f616c6963651203616c6c1a05616c6963652209696e73616173697479320c08bfe2d7ca0510e197be9301

Forge generated sha512 (hex): 6a09c3a667c3b3c2bcc38908c2bb67c2aec285c284c38ac2a73b3c6ec3b372c3bec294c3b82bc2a54fc3b53a5f1d36c3b1510e527fc2adc3a6c282c391c29b0568c28c2b3e6c1f1fc283c399c2abc3bb41c2bd6b5bc3a0c38d19137e2179

Expected sha512 (hex): a9a49744122da1eca02334b2e37a7230864cafebd02479eb230d57e3caa1220a67d52316f1bfef68bcabb12b960dddc4ad104480428a9349b0e279dc02207030

nodejs crypto library producing expected: https://runkit.com/tobowers/5955f7b1fef5ec00129c7dd2
golang cyrpto library producing expected: https://play.golang.org/p/g3P9h9Wjo_

forge producing unexpected result: https://runkit.com/tobowers/5955f80ef9fe430012c7986c

PS: Amazing library and I'm super excited to use it. All of the more advanced features are working perfectly for me (rsa keypairs, certs, PSS, etc).

[tobowers]

Interestingly, the 'abc' from the tests seems to be the same in both langs: https://play.golang.org/p/BOlFmb-UU3
https://github.com/QuorumControl/forge/blob/e3c68e9695607702587583cda291d74e5369f21c/tests/unit/sha512.js#L18-L24

[tobowers]

ok... I think I got it: https://runkit.com/tobowers/5955faae1cba570012814f47

Which I kind of gleamed from here: #128

That interface is a bit confusing though... if the function can take a buffer it should probably do that conversion for you? Or raise an error when it's not a string?

[dlongley]

That interface is a bit confusing though... if the function can take a buffer it should probably do that conversion for you? Or raise an error when it's not a string?

Yes, you're right. Dealing with encoding issues and improving the API (throwing better errors, requiring buffers, forcing string encodings to be specified and so forth) is on the roadmap, with some preliminary work done on another branch here:

https://github.com/digitalbazaar/forge/tree/0.8.x

That branch is needs to be rebased on top of the recent build fixes that were performed to go from forge 0.6.x to 0.7.x before further work can occur.

[dlongley]

Closing, resolving this issue is on the roadmap.