- From: Mark Watson <watsonm@netflix.com>
- Date: Fri, 22 Jan 2016 11:04:09 -0800
- To: Boris Zbarsky <bzbarsky@mit.edu>
- Cc: "public-webcrypto@w3.org" <public-webcrypto@w3.org>
- Message-ID: <CAEnTvdDD+QzdPANMjTn65P0_ifO=PzQCMSvJFwbnc46Pp7-sHA@mail.gmail.com>
On Fri, Jan 22, 2016 at 10:26 AM, Boris Zbarsky <bzbarsky@mit.edu> wrote: > On 1/22/16 11:51 AM, Mark Watson wrote: > >> No idea. Point me at web-platform-tests 101 and I'll tell you :-) >> > > http://testthewebforward.org/docs/ is sort of a place to start, kinda. > > But in terms of evaluating what you can do with them, the upshot is that > you have a web server running on a localhost port, you have HTML, JS, etc > on that server. You use > http://testthewebforward.org/docs/testharness-library.html to write the > actual tests. > > There are some tests for getRandomValues at > https://github.com/w3c/web-platform-tests/tree/master/WebCryptoAPI that > you can examine if that helps. > > You can also run the tests directly on the web at > https://w3c-test.org/WebCryptoAPI/test_getRandomValues.html (or the same > with http:// if desired). > > One benefit of using web platform tests for this is that I know for a fact > that both Chrome and Firefox are running them in continuous integration, > with periodic syncs from the upstream repo, and I belive Microsoft is as > well (e.g. they have a fork of that repo at < > https://github.com/MicrosoftEdge/web-platform-tests>). > > Another benefit is that it should be pretty easy to use the existing > automation bits it has to generate reports on which browsers pass which > tests. > > A third benefit is that anyone can contribute tests and all that. ;) Well, those all sound like quite compelling benefits. Unless anyone has an opinion to the contrary I suggest we work on putting a full suite of WebCrypto tests there. I can start by migrating the Netflix test suites. ...Mark > > > -Boris >
Received on Friday, 22 January 2016 19:04:41 UTC