Due to disabling Selenium's control flow in the future, I'm trying to write new specs using async/await.
To do this, I have to use SELENIUM_PROMISE_MANAGER: false config option, which is unfortunately making my old specs that use Promise Manager fail.
I've tried setting this option in beforeAllhook, but then I get the following error:
TypeError: Unable to create a managed promise instance: the promise manager has been disabled by the SELENIUM_PROMISE_MANAGER environment variable: undefined
If it's possible, I'd like to avoid having separate protractor conf files for async/non-async specs - it would make my grunt exec command even more complicated that it is now.
Is there any other option to set this flag only for one or few specs/suites?
Analysis:
The 'SELENIUM_PROMISE_MANAGER: false' in protractor config file should be global level, find a way to disable on spec level.
Possbile Solution:
add 'promise.USE_PROMISE_MANAGER = false' before spec use await, and revert to true at the end of spec. I'm not try as that, if it workable, please tell me.
Related
I am trying to implement raygun for a react-native app. I am running into problems with jest and I am not sure of the best approach given the window does not exist. In my package.json I have already added the following:
"globals": {
"window": {}
}
I get the following error though and I am unable to resolve it as of yet:
ReferenceError: raygunUtilityFactory is not defined
This is a very basic implementation as per the docs:
https://raygun.com/docs/languages/javascript#reactnative
I presume this is something with jest but having searched on github, here and google I still have not been able to resolve this.
Raygun4JS does not currently work correctly in a non-browser environment. Your best bet would be to not include the Raygun4JS code inside of your test environment (which is what we do at Raygun) but that is a bit more complicated using the UMD module from NPM (we use the snippet approach to inject it dynamically).
I think the problem stems from window not being considered the global object in Jest vs the browser, we have an indirect call to the window object to call the raygunUtilityFactory. Are you using the jsdom test environment? We also have the same window definition in our jestconfig.json and export some variables onto window in our setupframework.js file that are accessed indirectly through the global object and it appears to work correctly.
I am new to webpack and attempting to use it together with the Babel loader to write a library in ES6. I also want to use Karma/PhantomJS for a testing pipeline.
I have hit an odd issue where PhantomJS indicates that ES6 code didn't get converted when I run the tests:
SyntaxError: Use of reserved word 'let' in strict mode
at webpack:///say/hello.js:2:0 <- say/hello.spec.js:22929
But this only happens when I export my webpack config as a function (to take advantage of environment as per https://webpack.js.org/configuration/configuration-types/#exporting-a-function-to-use-env). Exporting config as an object is fine.
I have made a public repo to help others reproduce this at https://github.com/agentreno/es6-library-example with the npm 'test' and 'test:broken' tasks to demonstrate. I would greatly appreciate any assistance.
The problem is not the webpack config, but how you use it in your karma.config on line 37. The webpack option in the karma config expects an object, but you're giving it a function. You actually need to call the function to get the resulting object, which webpack usually does for you when you run webpack from the command line.
As you're not currently using env (the input parameter to the webpack function), you can simply call it:
webpack: webpackConfig(),
But when you start having conditions based on this, you'd need to pass the correct values. Let's say you're using the example mentioned in the docs: --env.production and --env.platform=web. Webpack will call the function as follows:
webpackConfig({ production: true, target: 'web' })
I am running Jasmine tests on Karma (latest, 0.12.14) from IntelliJ IDEA (latest, 13.1.2) on OS X.
IntelliJ just uses Karma plugin that points to my karma installation: /usr/local/lib/node_modules/karma
The problem is that Karma watcher doesn't pick new changes from JS files. When I run tests again, everything is fine. It's not a huge issue (they run pretty fast), but it's annyoing to always run tests twice after you do a change.
Anybody experienced similar issue?
Add the following property to your karma.conf.js file:
usePolling: true
I put it after the autoWatch flag.
Reference: AutoWatch doesn't work #895
Finally got this to work. This is a known issue when using Karma in IntelliJ with build tools like Webpack / Gulp / etc... and that the post-processed files are outside of the files that Karma is directly serving.
The workaround is to manually modify the IntelliJ config in this file: ...\plugins\js-karma\js_reporter\karma-intellij\lib\intellijRunner.js
At or near line #75 in the runWithConfig() method, change the refresh property --> true.
See this comment for details
Had this issue in both IntelliJ and Webstorm and fixed both the same way.
In the case of using IntelliJ IDEA, maybe try
Preferences > General > Uncheck "safe write"
I am not sure if question is still actual, but i found a solution (or just a workaround).
I changed this file C:\Users\MyUser\.IntelliJIdea14\config\plugins\js-karma\js_reporter\karma-intellij\lib\intellij.conf.js a little bit here:
config.singleRun = false;
var originalAutoWatch = config.autoWatch;
//config.autoWatch = false; <-- this line
config.autoWatchBatchDelay = 0;
For sure the plugin maintainers had some reason to add this line, but for me this worked fine enough (there were isolated crashes I can't reproduce).
I am not sure about older versions of intelij but I am using 2016.3 (the latest version at the time of writing this) and it all works OK assuming karma config has autoWatch: true.
Having exactly the same problem all I have to do was to switch on Toggle auto-test in intelij and all started working fine without any other change.
Please check autoWatch property in karma config file set to TRUE.
Also if you are using Sublime Text 3 then set "atomic_save": false in your user settings as it also causes issues with watches. see here
Macros worked for me. After trying usePolling: true, modifying the intellijRunner.js, and unchecking Safe Write I still had to save or run tests twice to catch the changes.
Made a compound macro to both Save and Run File and Save and Re-Run Test. Then I overwrote each commands' shortcut via Keymap.
How to bind multiple actions to a shortcut
screenshot of macro: Save and Run test at cursor
screenshot of macro: Save and Re-run last test
screenshot of Keymap for macros
i have a problem to run test. My model use extension Yii mail and then i run test its fail with wrong assert path. Another test runs finaly (model dont use any extensions). Preloading is only log.
I had a similar error and I explicitly set the basePath in config/test.php.
'components'=>array(
...
'assetManager'=>array(
'basePath'=>dirname(__FILE__).'/../../assets',
)
)
Im solved problem
public function setUp(){
Yii::app()->assetManager->basePath = '../../asserts';
}
Im dont know why this error throw only in one model...
PhpUnit runs primary in CLI mode and therefore some of environmental variables are missing. Yii's AssetManager uses one of such variable to determine webroot and since the variable does not exist, it will either throw error or set up invalid assets path on first attempt.
In my opinion, this issue is (indirectly) caused by PHPUnit because it only supports CLI testing mode, which makes some things really more difficult to test than it would be in HTTP request mode. Some guys therefore wrote tools to run unit tests via standard web GUI with whole native HTTP environment (e.g. https://github.com/NSinopoli/VisualPHPUnit). Eventually, you may use HTTP clients like Selenium to run your tests as if clicking over the page (see http://phpunit.de/manual/3.7/en/selenium.html).
Nevertheless, it's a matter of subjective opinion - somebody may argue, that testing in CLI mode has advantages, some guys will hate it. But the fact is, that one has to bear in mind differences between HTTP and CLI mode.
Anyone know if and how it's possible to run casperjs from within the phantomjs shell (a.k.a InteractiveModeREPL )?
I've also tried passing the direct path to the casper.js module and that has not worked either.
Progress/Update:
Tried phantomjs.injectJs('C:/casperjs/module/casper.js'); but got Error: Cannot find module 'colorizer' I guess I'm getting close.
This gets me closer but still missing path:
phantomjs.injectJs('C:/casperjs/module/bin/bootstrap.js')
errors out with Cannot find package.json at C:/package.json
OK, looks like I can pass the --casper-path option when starting phantomjs (see - casper/bin/bootstrap.js: line 189).
OK that worked. (passing the option did not work but setting the path inside of phantom did).
So to get this stuff to run inside the phantomjs shell first you need to set a casperPath variable in the phantom global object.
phantom.casperPath = "C:/casper";
Then you need to inJect caspers's bootstrap.js file.
phantom.injectJs("C:/casper/bin/bootstrap.js");
Now you can instantiate a casper object and play with it in the shell.
var casper = require("casper").create();
enjoy.