WebCrypto: HDKF and PBKDF2 return an empty ArrayBuffer when length is zero #49048
Conversation
javifernandez
changed the title
javifernandez
force-pushed
the
hkdf-and-pbkdf2-zero-length
branch
from
980f750 to
45204d9
Compare
The PR#380 [1] changed the PBKDF2 deriveBits operation to allow zero length and introduced an additional step to return an empty string in that case. It also reversted the PR#275 [2] so that HKDF also handles the zero length in the same way. This PR updates the tests cases affecting this 2 algorithms on the cases where zero was passed in the length parameter. [1] w3c/webcrypto#380 [2] w3c/webcrypto#275
javifernandez
force-pushed
the
hkdf-and-pbkdf2-zero-length
branch
from
45204d9 to
194773e
Compare
There was a problem hiding this comment.
LGTM 👍
However, @javifernandez do you know why the pipelines failed, or who to ask? Seems like installing Safari Technology Preview hit a 403 here: https://dev.azure.com/web-platform-tests/wpt/_build/results?buildId=129776&view=logs&j=0a19af81-f788-555d-f214-a5e35905eaf5&t=84b058e1-3bc1-5dbb-06da-ea0f1018bbe7
I'll try to get some information about it. |
I have contacted some folks at Apple involving on the WPT support; it seems there was a problem with the STP downloads, but the infra issue should be solved now. I think we can merge this now, but let me now if you think we should re-trigger the checks. |
twiss
changed the title
The PR#380 [1] changed the PBKDF2 deriveBits operation to allow zero length and introduced an additional step to return an empty string in that case. It also reversted the PR#275 [2] so that HKDF also handles the zero length in the same way.
This PR updates the tests cases affecting this 2 algorithms on the cases where zero was passed in the length parameter.
[1] w3c/webcrypto#380
[2] w3c/webcrypto#275