WebCrypto: Fix a crash when passing invalid ArrayBufferView to importKey.

WebCrypto: Fix a crash when passing invalid ArrayBufferView to importKey.

BUG=245025
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=154965

[eroman, 2013-07-25 23:29:38]

Adam, is this a bug in the IDL generator?

In the SubtleCrypto.idl [1] I specify "keyData" as "ArrayBuferView keyData", so
I figured the bindings generator would make sure it was the right type before
calling my function.

Obviously it was my bug regardless for not testing it, but wondering if I should
follow-up with the bindings generator.

[1]
https://chromium.googlesource.com/chromium/blink/+/master/Source/modules/cryp...

[abarth-chromium, 2013-07-25 23:31:30]

LGTM

https://codereview.chromium.org/20582003/diff/1/LayoutTests/crypto/importKey....
File LayoutTests/crypto/importKey.html (right):

https://codereview.chromium.org/20582003/diff/1/LayoutTests/crypto/importKey....
LayoutTests/crypto/importKey.html:210:
shouldThrow("crypto.subtle.importKey(keyFormat, null, algorithm, extractable,
keyUsages)");
null is a perfectly fine ArrayBufferView.  What happens if you pass in something
like 5 or document.body ?

[eroman, 2013-07-25 23:36:41]

I tested with:

  []
  null
  ''
  document.body

These all are accepted by the generated bindings, and pass NULL as the
ArrayBufferView.

[abarth-chromium, 2013-07-25 23:37:18]

That sounds like a bug in the code generator.  At least some of those should be
rejected as having the wrong type.

[eroman, 2013-07-25 23:38:06]

Thanks, I will follow-up with the IDL generator then!

[commit-bot: I haz the power, 2013-07-25 23:45:02]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/eroman@chromium.org/20582003/1

[commit-bot: I haz the power, 2013-07-26 02:46:55]

Change committed as 154965