Bug 1485414 - Make lazy_static a dev dependency. r=whimboo This silences an unused macro_use warning at build time.

Bug 1484941 - Change ErrorStatus::{ScriptTimeout,Timeout} HTTP codes from 408 to 500. r=whimboo Whilst it is logically correct to use 408 for the ErrorStatus::ScriptTimeout and ::Timeout errors, it causes a collision with HTTP semantics implement in HTTP clients. To support Keep-Alive we allow retries in HTTP clients and if a client sees code 408 it thinks that the server has gone away and retries the request. This causes WebDriverCommand::ExecuteScript, ::Get, and ::Refresh commands to be sent twice with some HTTP clients. This is a backwards incompatible change to WebDriver in order to not break HTTP/1.1 Keep-Alive.

Bug 1484157 - Add bounds check tests for WebDriver pause action. r=whimboo The pause action takes an unsigned integer as input, but it was discovered in https://github.com/mozilla/geckodriver/issues/1355 that chromedriver accepts floating point values. To achieve conformance in this area, this will test various allowed and disallowed values so that the pause action's bounds are properly checked.

Bug 1465046 - Remove ProcessHandler.waitForFinish(). r=whimboo

Bug 1484153 - Drop duplicate outerWindowID logging in listener.js. r=whimboo When receiving observer notifications during navigation, we log the outerWindowID: logger.debug(`Received observer notification ${topic} for ${winID}`); This is no longer necessary as https://bugzilla.mozilla.org/show_bug.cgi?id=1464469 introduced the outerWindowID as a prefix to all log messages originating from this frame script.

Bug 1482829 - Track Marionette logger verbosity with Log#manageLevelFromPref. r=whimboo This patch adopts Logger#managerLevelFromPref from Log.jsm to set and keep track of the Marionette logger's verbosity. This has the advantage that we do not have to roll separate implementations of Log for the child- and parent processes. It also has the upside that the log level will be reflected when changed at runtime through the use of an observer.

Bug 1482829 - Convert logging::Level into mozprofile::preferences::Pref. r=whimboo In order to facilitate the use of Log.jsm's Logger#manageLevelFromPref, geckodriver needs to ensure that the input value for the marionette.log.level preference conforms to the variants in the Log.Level enum. This patch implements the Into<T> conversion from geckodriver's logging::Level into mozprofile::preferences::Pref by way of a new function to_gecko(), that ensures the preference value is correctly formatted. Logger#manageLevelFromPref expects a string value such as "Info", which exactly matches Log.Level's own properties. It is in other words case sensitive, and this ensures that Marionette no longer has to case convert the input data.

Bug 1484161 - Replace nsITimer with Sleep for dispatchPause. r=whimboo

Bug 1483288 - Avoid "as i64" coercion of marionette.port pref value. r=whimboo

Bug 1483288 - Provide more number conversions for mozprofile::preferences::PrefValue. r=whimboo Allow i8, u8, i16, u64, i32, and u32 to be implicitly converted into PrefValue::Int. u64 is not supported because it would overflow, so this still needs to be handled manually. geckodriver stores the port number as u8 and this will allow it to implicitly convert it to PrefValue::Int without using the unsafe "as i64" coercion.

Bug 1478558 - Remove usage of http://mozqa.com/ in services/sync/tests/tps/test_tabs.js r=whimboo,tcsc MozReview-Commit-ID: Gd8qmfBWf8u Differential Revision: https://phabricator.services.mozilla.com/D2547

Bug 1443572 - Disable Photon onboarding experience for Marionette and geckodriver r?whimboo MozReview-Commit-ID: HiT4QaV6FPg

Bug 1480612: Test Register Application Restart more thoroughly. r=whimboo

Bug 1478094 - [marionette] Update crash unit tests to make use of about:crashparent and about:crashcontent r=whimboo MozReview-Commit-ID: H9LBBD7t9vL

Bug 1434281 - also delete local/cache profile directory from refresh test, r=whimboo Ensure local profile directory is also deleted in test teardown. Differential Revision: https://phabricator.services.mozilla.com/D2671

Bug 1480105 - Using WebDriver:AcceptAlert instead of WebDriver:AcceptDialog in Marionette. r=whimboo MozReview-Commit-ID: 1lZYKHlhhpn

Bug 1470659 - Parametrize capabilities tests. r=whimboo To be able to ignore tests for individual capabilities we need to parametrize these tests. geckodriver now supports setWindowRect, but fails the proxy capability test because it is for some reason not propagated back.

Bug 1470659 - Move timeouts test to separate file. r=automatedtester,whimboo For similar reasons as for platformName, tests for configuring the timeouts object do not belong in the same parent test as those for response body structure.

Bug 1470659 - Move platformName test to separate file. r=whimboo Specific tests for the platformName capability do not belong in the test for the response body structure.

Bug 1470659 - Add setWindowRect capability to Marionette. r=whimboo The setWindowRect capability is mandated by the WebDriver standard and is an indication whether the driver supports manipulating the window dimensions and position. This will always be true for Firefox and always false for Fennec.

Bug 1470659 - Simplify Marionette capability parsing. r=whimboo Change the switch-statement to only do assertions, and replace the v variable with any parsed values before writing it to the matched set of capabilities in one location.

Bug 1451725 - Remove deprecated WebDriver Commands r=whimboo MozReview-Commit-ID: 7PXrlwyiFvN

Bug 1319793 - [marionette harness tests] Default to not swallow stdout from pytest. r=whimboo Added the flag 'capture --no' to the harness_unit tests. Stdout is now displayed by default. MozReview-Commit-ID: HZj3vntLcKI

Bug 1480105 - Using WebDriver:AcceptAlert instead of WebDriver:AcceptDialog in Marionette. r?whimboo MozReview-Commit-ID: 1lZYKHlhhpn

Bug 1480105 - Using WebDriver:AcceptAlert instead of WebDriver:AcceptDialog in Marionette. r?whimboo MozReview-Commit-ID: 1lZYKHlhhpn

Bug 1478979 - Disallow pageLoadStrategy to be null. r=whimboo Capabilities matching is now done in geckodriver and Marionette receives the negotiated set of capabilities, so there is no need to derive default values in case the value is null.

Bug 1400233 - Drop ContentWebElement.LegacyIdentifier key from Marionette. r=ato,whimboo Remove the legacy key that Marionette uses to identify web elements.

Bug 1477475 - Drop legacy Marionette capabilities duplication. r=whimboo geckodriver currently writes a legacy capabilities structure to the WebDriver:NewSession command it uses on creating a new WebDriver session which duplicates the top-level object: {capabilities: {<caps>, desiredCapabilities: <caps>}} Where caps can look something like: {acceptInsecureCerts: true, pageLoadStrategy: "normal", ...} Marionette has picked up the top-level "capabilities" key for a very long time now, and there is no longer any need to duplicate the body.

Bug 1319793 - [marionette harness tests] Default to not swallow stdout from pytest. r?whimboo Added the flag 'capture --no' to the harness_unit tests. Stdout is now displayed by default. MozReview-Commit-ID: HZj3vntLcKI

Bug 1319793 -[marionette harness tests] Default to not swallow stdout from pytest. r?whimboo Added the flag 'capture --no' to the harness_unit tests. Stdout is now displayed by default. MozReview-Commit-ID: HZj3vntLcKI

Bug 1477978 - Swallow SessionNotCreatedException on session.end(). r=whimboo Subsequent calls to session.end() should not result in SessionNotCreatedException being raised.

Bug 1452179: [mozharness] Don't set virtualenv_modules in talos, since it is unused; r=whimboo Talos uses an in-tree module that it installs in code, ignoring the config value. Differential Revision: https://phabricator.services.mozilla.com/D986

Bug 1452179: [mozharness] Don't special case the path to mozinstall on windows; r=whimboo New versions of pip don't install entrypoint scripts with a `-script.py` on windows. Instead, just call the bare script name everywhere. On unix-like systems, this uses the shebang; on Windows, this executes the `.exe` wrapper. Differential Revision: https://phabricator.services.mozilla.com/D967

Bug 1452490 - Remove marionetteScriptFinished. r=whimboo MozReview-Commit-ID: CdQCmtaodww

Bug 1121705 - Look at window handles decrement when closing tab. r=whimboo When Puppeteer opens a new tab using various strategies it relies on the list of window handles increasing. When performing the reverse operation of closing a tab, it looks at the length of <tab> chrome elements in the UI. This changes the close operation to use the same mechanism as opening a new tab to determine if the tab has been closed. This seems to be as reliable as looking at the number of <tab> elements. As part of a forthcoming window tracking refactoring of Marionette (https://bugzilla.mozilla.org/show_bug.cgi?id=marionette-window-tracking), the list of window handles will be made even more reliable: a content browser will not appear in the window handle list until both the tab and it's linked content browser has been created and properly initialised. MozReview-Commit-ID: FY5vGBpn64R

Bug 1121705 - Look at window handles decrement when closing tab. r?whimboo When Puppeteer opens a new tab using various strategies it relies on the list of window handles increasing. When performing the reverse operation of closing a tab, it looks at the length of <tab> chrome elements in the UI. This changes the close operation to use the same mechanism as opening a new tab to determine if the tab has been closed. This seems to be as reliable as looking at the number of <tab> elements. As part of a forthcoming window tracking refactoring of Marionette (https://bugzilla.mozilla.org/show_bug.cgi?id=marionette-window-tracking), the list of window handles will be made even more reliable: a content browser will not appear in the window handle list until both the tab and it's linked content browser has been created and properly initialised. MozReview-Commit-ID: FY5vGBpn64R

Bug 1121705 - Look at window handles decrement when closing tab. r?whimboo When Puppeteer opens a new tab using various strategies it relies on the list of window handles increasing. When performing the reverse operation of closing a tab, it looks at the length of <tab> chrome elements in the UI. This changes the close operation to use the same mechanism as opening a new tab to determine if the tab has been closed. This seems to be as reliable as looking at the number of <tab> elements. As part of a forthcoming window tracking refactoring of Marionette (https://bugzilla.mozilla.org/show_bug.cgi?id=marionette-window-tracking), the list of window handles will be made even more reliable: a content browser will not appear in the window handle list until both the tab and it's linked content browser has been created and properly initialised. MozReview-Commit-ID: FY5vGBpn64R

Bug 1121705 - Look at window handles decrement when closing tab. r?whimboo When Puppeteer opens a new tab using various strategies it relies on the list of window handles increasing. When performing the reverse operation of closing a tab, it looks at the length of <tab> chrome elements in the UI. This changes the close operation to use the same mechanism as opening a new tab to determine if the tab has been closed. This seems to be as reliable as looking at the number of <tab> elements. As part of a forthcoming window tracking refactoring of Marionette (https://bugzilla.mozilla.org/show_bug.cgi?id=marionette-window-tracking), the list of window handles will be made even more reliable: a content browser will not appear in the window handle list until both the tab and it's linked content browser has been created and properly initialised. MozReview-Commit-ID: FY5vGBpn64R

Bug 1121705 - Look at window handles decrement when closing tab. r?whimboo When Puppeteer opens a new tab using various strategies it relies on the list of window handles increasing. When performing the reverse operation of closing a tab, it looks at the length of <tab> chrome elements in the UI. This changes the close operation to use the same mechanism as opening a new tab to determine if the tab has been closed. This seems to be as reliable as looking at the number of <tab> elements. As part of a forthcoming window tracking refactoring of Marionette (https://bugzilla.mozilla.org/show_bug.cgi?id=marionette-window-tracking), the list of window handles will be made even more reliable: a content browser will not appear in the window handle list until both the tab and it's linked content browser has been created and properly initialised. MozReview-Commit-ID: FY5vGBpn64R

Bug 1470646 - Update Wd tests to use WebDriver conforming platformName. r=whimboo MozReview-Commit-ID: LCi2BgdMQLb

Bug 1470646 - Update Fxfn tests to use WebDriver conforming platformName. r=whimboo MozReview-Commit-ID: 1GYS453pA5D

Bug 1470646 - Update Mn and Fxfn tests to use WebDriver conforming platformName. r=whimboo MozReview-Commit-ID: 387jlJCbV59

Bug 1470646 - Return platformName as recommended by WebDriver. r=whimboo The WebDriver standard recommends that we return one of "windows", "mac", or "linux". Additionally Marionette supports Fennec, for which we return "android": this is fine because the specification does not yet cover mobile browsers. MozReview-Commit-ID: EgZ9UKrbsrd

Bug 1470646 - Modularise capabilities module. r=whimboo This removes the "session." prefixed global export type in favour of individually exported types. MozReview-Commit-ID: 3DHTrJsy2IN

Bug 1470646 - Rename session module to capabilities. r=whimboo MozReview-Commit-ID: 4hQN2gv3HDd

Bug 1470646 - Increase wdspec long timeout to three minutes. r=whimboo The long timeout for WPT wdspec tests is currently two minutes, but with modifications to make more session tests run in Firefox, it has been discovered that the long timeout is not sufficient time for Firefox to consistently run the New Session command tests in automation without timing out. This increases the long timeout to three minutes (180 seconds). MozReview-Commit-ID: A8A4zecOwES

Bug 1470646 - Silence missing nodeType property warning. r=whimboo If val does not have a property nodeType, a warning is emitted to the browser console. This is observable when running the xpcshell unit tests because we mock out val without a nodeType property. MozReview-Commit-ID: GzqMoJQQdF8

Bug 1470646 - Update Wd tests to use WebDriver conforming platformName. r?whimboo MozReview-Commit-ID: LCi2BgdMQLb

Bug 1470646 - Update Fxfn tests to use WebDriver conforming platformName. r?whimboo MozReview-Commit-ID: 1GYS453pA5D

Bug 1470646 - Update Mn and Fxfn tests to use WebDriver conforming platformName. r?whimboo MozReview-Commit-ID: 387jlJCbV59