Bug 227731 – PCM: Add error logging for CryptoKit operations

• Status: RESOLVED
• Resolution: FIXED
• Component: WebKit Misc.

[John Wilander]

There are three FIXMEs for error logging in PCM's CryptoKit operations. We should add the logging.

[John Wilander]

<rdar://80221057>

[John Wilander]

Created attachment 432994 [details]
Patch

[John Wilander]

Comment on attachment 432994 [details]
Patch

Marking as obsolete pending updates to API tests.

[John Wilander]

A failing API test was addressed in a separate patch (Bugzilla related to this one).

[John Wilander]

Created attachment 433106 [details]
Patch

[Brent Fulgham]

Comment on attachment 433106 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=433106&action=review

r=me, but consider my suggestion about the test case.

> Tools/TestWebKitAPI/Tests/WebCore/PrivateClickMeasurement.cpp:323
> +    EXPECT_FALSE(pcm.calculateAndUpdateSourceUnlinkableToken(serverPublicKeyBase64URL));

I think these tests would look better if you asserted EXPECT_FALSE(pcm.calculateAndUpdateSourceUnlinkableToken(serverPublicKeyBase64URL).has_value());

The fact that the return value is an error message (or no-value if successful) isn't obvious from the name, so it looks like we are expecting the calculateAndUpdate to fail.

[John Wilander]

Created attachment 433115 [details]
Patch for landing

[John Wilander]

(In reply to Brent Fulgham from comment #6)
> Comment on attachment 433106 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=433106&action=review
> 
> r=me, but consider my suggestion about the test case.
> 
> > Tools/TestWebKitAPI/Tests/WebCore/PrivateClickMeasurement.cpp:323
> > +    EXPECT_FALSE(pcm.calculateAndUpdateSourceUnlinkableToken(serverPublicKeyBase64URL));
> 
> I think these tests would look better if you asserted
> EXPECT_FALSE(pcm.
> calculateAndUpdateSourceUnlinkableToken(serverPublicKeyBase64URL).
> has_value());
> 
> The fact that the return value is an error message (or no-value if
> successful) isn't obvious from the name, so it looks like we are expecting
> the calculateAndUpdate to fail.

You're absolutely right. I made the test look like the actually call sites with a local variable errorMessage. Should make it clear.

Thanks for the review!

[EWS]

Committed r279710 (239502@main): <https://commits.webkit.org/239502@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 433115 [details].