[webcrypto] Import WebKit's webcrypto tests.

[webcrypto] Import WebKit's webcrypto LayoutTests.

Some modifications necessary to account for differences in version of the spec and implementation.

BUG=377941, 245025
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175833

[eroman, 2014-05-30 20:17:23]

The modifications I made can be seen in difference between patchset 1 and 2.

[jww, 2014-06-03 19:37:47]

lgtm, modulo my one question.

Man, I'm thrilled you did this, but I'm glad I wasn't the one who had to :-)

https://codereview.chromium.org/298023002/diff/40001/LayoutTests/crypto/aes-c...
File LayoutTests/crypto/aes-cbc-invalid-length.html (right):

https://codereview.chromium.org/298023002/diff/40001/LayoutTests/crypto/aes-c...
LayoutTests/crypto/aes-cbc-invalid-length.html:23: debug("importKey() should
have failed, since the key length was 176 bits");
Why drop all the "shouldBe" sanity checks? Are you just assuming that unit
testing elsewhere should catch all that stuff?

Although looking at some of the later tests, you leave in the sanity checks, so
I'm not sure which way you want to go.

[eroman, 2014-06-03 19:52:05]

https://codereview.chromium.org/298023002/diff/40001/LayoutTests/crypto/aes-c...
File LayoutTests/crypto/aes-cbc-invalid-length.html (right):

https://codereview.chromium.org/298023002/diff/40001/LayoutTests/crypto/aes-c...
LayoutTests/crypto/aes-cbc-invalid-length.html:23: debug("importKey() should
have failed, since the key length was 176 bits");
On 2014/06/03 19:37:47, jww wrote:
> Why drop all the "shouldBe" sanity checks? Are you just assuming that unit
> testing elsewhere should catch all that stuff?
> 
> Although looking at some of the later tests, you leave in the sanity checks,
so
> I'm not sure which way you want to go.

Thanks for reading through all of those differences!

The removal of the shouldBe() is intentional.

The reason is as follows:

The original Webkit test does an importKey() for a 176 bit AES key. In WebKit,
this operation succeeds, so it goes on to verify that the key was created with
the expected properties, and goes on to try using it in an encrypt operation.

By contrast in Blink, importKey() for a 176-bit AES key fails. So there are no
properties to check on the key, just the error that was returned by importKey().

The Blink behavior matches the latest version of the spec, which says in section
"18.11.4. Operations" regarding the raw import of AES-CBC keys:

"""If the length in bits of data is not 128, 192 or 256 then return an error
named DataError."""

(The version of the spec which WebKit based this test off of was ambiguous on
how to treat this, but the latest version is pretty clear).

[jww, 2014-06-03 21:00:19]

On 2014/06/03 19:52:05, eroman wrote:
>
https://codereview.chromium.org/298023002/diff/40001/LayoutTests/crypto/aes-c...
> File LayoutTests/crypto/aes-cbc-invalid-length.html (right):
> 
>
https://codereview.chromium.org/298023002/diff/40001/LayoutTests/crypto/aes-c...
> LayoutTests/crypto/aes-cbc-invalid-length.html:23: debug("importKey() should
> have failed, since the key length was 176 bits");
> On 2014/06/03 19:37:47, jww wrote:
> > Why drop all the "shouldBe" sanity checks? Are you just assuming that unit
> > testing elsewhere should catch all that stuff?
> > 
> > Although looking at some of the later tests, you leave in the sanity checks,
> so
> > I'm not sure which way you want to go.
> 
> Thanks for reading through all of those differences!
> 
> The removal of the shouldBe() is intentional.
> 
> The reason is as follows:
> 
> The original Webkit test does an importKey() for a 176 bit AES key. In WebKit,
> this operation succeeds, so it goes on to verify that the key was created with
> the expected properties, and goes on to try using it in an encrypt operation.
> 
> By contrast in Blink, importKey() for a 176-bit AES key fails. So there are no
> properties to check on the key, just the error that was returned by
importKey().
> 
> The Blink behavior matches the latest version of the spec, which says in
section
> "18.11.4. Operations" regarding the raw import of AES-CBC keys:
> 
> """If the length in bits of data is not 128, 192 or 256 then return an error
> named DataError."""
> 
> (The version of the spec which WebKit based this test off of was ambiguous on
> how to treat this, but the latest version is pretty clear).

Got it. Thanks for the explanation!

[eroman, 2014-06-09 22:57:17]

The CQ bit was checked by eroman@chromium.org

[commit-bot: I haz the power, 2014-06-09 22:58:00]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/eroman@chromium.org/298023002/60001

[commit-bot: I haz the power, 2014-06-09 23:10:14]

Change committed as 175833