- From: Alexei Barantsev <barancev@gmail.com>
- Date: Mon, 16 Jan 2017 23:44:41 +0300
- To: Simon Stewart <simon.m.stewart@gmail.com>
- Cc: David Burns <dburns@mozilla.com>, public-browser-tools-testing <public-browser-tools-testing@w3.org>
- Message-ID: <CAF5xrgkuki8wAhniJuYa4pT0t8ACR6kwcV5xqs2C-mCRUJqsJw@mail.gmail.com>
Legacy FirefoxDriver have similar code too: https://github.com/SeleniumHQ/selenium/blob/master/javascript/firefox-driver/js/syntheticMouse.js#L112-L210 Regards, -- Alexei Barantsev Software-Testing.Ru Selenium2.Ru 2017-01-16 23:02 GMT+03:00 Simon Stewart <simon.m.stewart@gmail.com>: > If we can get it landed, that'd be great. Perhaps someone from the chrome > team can pick this up? They have working code already…. Presumably this is > hooking into the hit testing somehow. > > Simon > > > On Mon, Jan 16, 2017 at 7:58 PM, David Burns <dburns@mozilla.com> wrote: > >> The last one. >> >> Looking at the current prose in Element Click, adding the necessary prose >> shouldn't be that difficult. I am happy to take that on. >> >> David >> >> On 16 January 2017 at 14:12, Simon Stewart <simon.m.stewart@gmail.com> >> wrote: >> >>> Hi, >>> >>> The chromedriver, at least, will notify users if they attempt to click >>> on an element that isn't actually clickable (eg. because it's covered by >>> another) There's an open PR (#299 >>> <https://github.com/w3c/webdriver/pull/299>) that adds a new "element >>> intercepting click" for this, but there's some discussion on that diff >>> concerning how this should be implemented. >>> >>> Rather than leaving the PR languishing in purgatory, we should have a >>> discussion here. The choices that I see are: >>> >>> * Ignore this case, and leave it unspecified. >>> >>> * Optionally allow drivers to raise an "invalid element state" error if >>> the user would not actually be able to click on the element (which may make >>> testing for clickjacking impossible since we don't have any language to say >>> whether or not an element would be user visible any more) >>> >>> * Add a new error message, and work out the spec-ese to make this happen. >>> >>> I'm assuming that this only applies to "Element Click", since the >>> advanced user interactions are working on coordinates rather than elements, >>> and so should be figuring out the element which should receive events. >>> >>> If we had the time, the last option feels like the best, with users >>> being forced to use the Action commands to test clickjacking. Thoughts? >>> >>> Simon >>> >> >> > ᐧ
Received on Monday, 16 January 2017 20:45:15 UTC