Bug 124483 – Support exporting private WebCrypto RSA keys

• Component: WebCore Misc.
• Status: RESOLVED FIXED

[Alexey Proskuryakov]

The difficulty here is that CommonCrypto functions don't expose all key parts necessary for JWK, so we'll need to compute them from prime factors manually.

[Alexey Proskuryakov]

Created attachment 218322 [details]
proposed patch

[Anders Carlsson]

Comment on attachment 218322 [details]
proposed patch

View in context: https://bugs.webkit.org/attachment.cgi?id=218322&action=review

> Source/WebCore/crypto/CommonCryptoUtilities.cpp:82
> +    RELEASE_ASSERT(!status);

I’d rather assert status == kCCSuccess here (and all the other places).

[WebKit Commit Bot]

Attachment 218322 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/crypto/subtle/rsa-export-private-key-expected.txt', u'LayoutTests/crypto/subtle/rsa-export-private-key.html', u'Source/WebCore/ChangeLog', u'Source/WebCore/crypto/CommonCryptoUtilities.cpp', u'Source/WebCore/crypto/CommonCryptoUtilities.h', u'Source/WebCore/crypto/mac/CryptoKeyRSAMac.cpp']" exit_code: 1
ERROR: Source/WebCore/crypto/CommonCryptoUtilities.cpp:32:  Missing spaces around /  [whitespace/operators] [3]
ERROR: Source/WebCore/crypto/CommonCryptoUtilities.h:100:  Missing spaces around &&  [whitespace/operators] [3]
ERROR: Source/WebCore/crypto/CommonCryptoUtilities.h:102:  Missing spaces around &&  [whitespace/operators] [3]
Total errors found: 3 in 7 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Alexey Proskuryakov]

Committed <http://trac.webkit.org/r160029>.

[Alexey Proskuryakov]

> I’d rather assert status == kCCSuccess here (and all the other places).

These pretty much only fail on out of memory conditions, which is normally a release assert for us.

[Alexey Proskuryakov]

Or did you mean a RELEASE_ASSERT in a different form? I personally prefer "status == kCCSuccess" too, but I don't think we ever do that for other types.

[Anders Carlsson]

(In reply to comment #6)
> Or did you mean a RELEASE_ASSERT in a different form? I personally prefer "status == kCCSuccess" too, but I don't think we ever do that for other types.

I mean i like the condition to be (status == kCCSuccess) as opposed to (!status) since it’s not clear (to me anyway) that kCCSuccess is 0.