doc: Fix 'pbkdf2Sync' encoding problem

[ghost]

Ref: #1796.

Since the '/newUser' request is encoded by 'sha512', and your
corresponding '/auth' must also use 'sha512' as the digest encoding
as well. Otherwise your decoding isn't right.

[ghost]

/cc：@trevorah

[vsemozhetbyt]

cc @nodejs/crypto

[trevorah]

😳 whoops, sorry I didnt notice that in my original PR. Good spot @Maledong!

[tniessen]

Unrelated: 512 is an unusual key size (note that the key size parameter is in bytes, not bits). This might have originated from the misconception that it must match the output size of the hash function?

[trevorah]

@tniessen I picked sha512 for this guide because its used in the crypto examples (https://nodejs.org/api/crypto.html#crypto_crypto_pbkdf2_password_salt_iterations_keylen_digest_callback) which this guide appears to mirror.

[tniessen]

@trevorah SHA512 is a good choice, just the output key size is unusual (512 bytes!). This works equally well:

crypto.pbkdf2Sync('password', 'salt', 10000, 32, 'sha512')

PBKDF2 is usually used to derive symmetric keys or to store passwords, neither of these applications requires keys with a size of 512 bytes.

[trevorah]

@tniessen hmm... weird. The only crypto.pbkdf2* examples that use a 512 key size are the ones that set crypto.DEFAULT_ENCODING = 'hex'. All the others use a 64 byte key size. Maybe the original author wanted to show some difference when setting DEFAULT_ENCODING?

[tniessen]

Uhh DEFAULT_ENCODING... What a design mistake. It still doesn't really matter, if you set the DEFAULT_ENCODING to hex, the output will still be a 512-byte (1024 hex digits) sequence, just encoded. But again, it is unrelated, so doesn't need to be changed in this PR. Also, it's not wrong, just unusual.

[ghost]

@trevorah：No problem. Just found that by accident when reading lines :)

[vsemozhetbyt]

Doc format LGTM)

[ghost]

Thanks all!