Crypto Hmac SH256 returning incorrect values?

[cjc]

G'day,

I've created a test script based on RFC4231 -"Identifiers and Test Vectors for HMAC-SHA-224, HMAC-SHA-256, HMAC-SHA-384, and HMAC-SHA-512". Of the 7 test vectors provided, crypto.createHmac("sha256",.. appears to be returning incorrect results for 3,5,6 and 7.

http://gist.github.com/608586

[bartt]

Same thing happens in https://gist.github.com/663052

[bkw]

Test 5 should use truncation. If applied, the result looks valid to me.
Test 6 and 7 should hash the key since it is too long (See RFC2104).
Did not check Test 3.

We should fix these tests and add them to the test suite, though.

[stbuehler]

it doesn't like binary salts at all... probably as it converts the buffer to strings before decoding it as binary.
for the data it skips the toString part and uses the buffer directly.

[cjc]

I'm not sure when this changed, but a retest in 0.4.9 and 0.5.0 shows these tests working in both.

[stbuehler]

I just tried my test cases without my patch, and they still don't work. with my path, they do.

[bnoordhuis]

Okay, reopening. cjc's test case fails on master, will have to check what's at fault here.

[stbuehler]

well, just accept my pull request for it. includes test cases :) #1027

[koichik]

@cjc - After the patch was applied, your test1-4 passed successfully (before the patch was applied, test3 failed). Thanks for the report!

@bartt - Your test still failed. I think your problem is the same as #749.