0.13.0 Final
Changes from beta:
- Fail when removed option is detected and warn if removed env var is detected
(@SevInf). - Reports now always will be generated, even when there is a critical error
(@zumra6a). - Correctly cancel queued tests after critical error (@SevInf).
- Produce more debug information in various modules (@SevInf).
- In case of a critical error, try to print what test was executed when it
happened (@SevInf). - Corrected updates (@maximerassi).
Changes from 0.12.x:
-
BREAKING CHANGE: new config format. See docs.
New config allows to set many previously global options on per-browser
basis. The following things have changed in this config:-
rootUrl
,gridUrl
,tolerance
,strictComparison
,screenshotsDir
,
windowSize
options can now be set separately for each browser:gridUrl: http://grid.example.com browsers: firefox: # Uses gridUrl from top level desiredCapabilities: ... chrome: # Overrides top-level value gridUrl: http://chrome.example.com desiredCapabilities: ...
-
projectRoot
,sourceRoot
,plugins
,debug
,parallelLimit
,
diffColor
,referenceImageAbsence
are moved intosystem
section and
can not be set per-browser.system: debug: true diffColor: #ff0000 plugins: teamcity: true ...
-
browsers
are no longer default tophantomjs
. If you've used this
default, set up the browser explicitly:browsers: phantomjs: desiredCapabilities: browserName: phantomjs
-
coverage settings are now grouped under
system.coverage
section:- to enable coverage, set
system.coverage.enabled
to true. coverageExclude
is moved tosystem.coverage.exclude
.coverageNoHtml
is replaced bysystem.coverage.html
. Set it to false
to disable html report generaion.
- to enable coverage, set
-
--noCalibrate
custom capability is replaced bycalibrate
option which
can be set for every or any particular browser. -
browser capabilites are set in
desiredCapabilites
option.browsers: chrome: desiredCapabilities: browserName: 'chrome', version: '45'
-
top-level
capabilities
option is replaced bydesiredCapabilites
option. -
http
settings are removed. Use new optionhttpTimeout
to set timeout.
Setting retires is no longer possible. -
sessionMode
is replaced by more flexiblesuitesPerSession
setting (see
below).
-
-
The way config options are overriden by CLI flags and environment variables
are now unified (option path is converted to--option-path
for cli and
gemini_option_path
for environment variables). Due to this change, some
old flags and environment variables won't work:--sorce-root
andGEMINI_SOURCE_ROOT
becomes--system-source-root
and
gemini_system_source_root
respectively.--debug
andGEMINI_DEBUG
becomes--system-debug=true
and
gemini_system_debug
respectively.--coverage
andGEMINI_COVERAGE
becomes
--system-coverage-enabled=true
andgemini_system_coverage_enabled
respecitvely.--coverage-no-html
andGEMINI_COVERAGE_NO_HTML
becomes
--system-coverage-html=false
andgemini_system_coverage_html
respectively.GEMINI_ROOT_URL
becomesgemini_root_url
.GEMINI_GRID_URL
becomesgemini_grid_url
.GEMINI_SCREENSHOTS_DIR
becomesgemini_screenshots_dir
.GEMINI_WINDOW_SIZE
becomesgemini_window_size
.
-
API:
Gemini
constuctor does not accepts overrides object anymore. -
sessionMode
is replaced bysuitesPerSession
option which specifies
number of test suites to run in a single WebDriver session. Value of.inf
is equivalent toperBrowser
session mode and value of1
is equivalent to
perSuite
. This option can be set globally or separately for each browser. -
New option
sessionsPerBrowser
allows to launch multiple session for each
browser and run tests in parallel.