When I start a new Spectron application (for each test suite) I would like to fix the size of the app (for consistency across all machines and reloads).
Commands like setBounds and maximise do change dimensions, but they only do so after the app has started (which means that some components have already assumed certain dimensions, which then changes test results).
In the Spectron docs various launch options are available. I've tired to use webdriver and chromedriver options but they don't seem to work. Here is an example:
app = new Application({
path: kElectronPath,
webdriverOptions: {
width:1368,
height:769,
},
I just assumed that the webdriver options came from the browserwindow class. How are webdriver and chromedriver options accessed in Spectron?
This should help
this.app = new Application({
path: './ac.exe',
args: ['app'],
webdriverOptions: ({deprecationWarnings : false})
});
pass the parameters that need to be changed during app launch
and have your new application config in before hook
Related
I have a few tests that download files and assert data inside. The problem I am facing is that the tests run in parallel so I can't delete the download directory after each test or else they delete each others files. The issue with not doing so however is that the filename includes timestamp(unique identifier) which is not known to the test so not possible to know which file to open. Is there a way to change default_directory for a given test in the middle of test run? The idea is to be able to tell capybara that for certain tests override the download path to be another path temporarily? I guess there could be cross wiring here too if it was possible as other tests could still be running expecting the original path to be set?
Alternatively, any suggestions on handling this?
My capybara config looks like this
options = Selenium::WebDriver::Chrome::Options.new
preferences = {
prompt_for_download: false,
credentials_enable_service: false,
default_directory: DownloadUtil::PATH
}
options.add_preference(:download, preferences)
options.add_argument('--disable-infobars')
options.add_argument('--headless')
Capybara::Selenium::Driver.new(app, browser: :chrome, options: options)
end
Most multi-process parallel test setups provide you with an environment variable you can use to configure things that need to be different between each instance of test runner (DB name, ports, etc). In the case of parallel_rspec that is TEST_ENV_NUMBER. Using that you can configure the selenium/chrome instance in each test runner to use a different download directory - something like
preferences = {
prompt_for_download: false,
credentials_enable_service: false,
default_directory: DownloadUtil::PATH + ENV['TEST_ENV_NUMBER']
}
This is the same question as this but for the recent version of application insights 2.2.1
Since updating to the 2.2 version the debug output is filled with AI data even if it is disabled the way it used to be done.
Previously AI was enabled in startup and I could do something like this:
services.AddApplicationInsightsTelemetry(options =>
{
options.EnableDebugLogger = false;
options.InstrumentationKey = new ConnectionStringGenerator().GetAITelemetryKey();
});
The new method of adding application insights, per the new VS templates, is to add it in Program.cs like this:
public static IWebHost BuildWebHost(string[] args) =>
WebHost.CreateDefaultBuilder(args)
.UseStartup<Startup>()
.UseApplicationInsights(connectionStringGenerator.GetAITelemetryKey())
.UseSerilog()
.Build();
In this case there is no construction that takes any options and if I remove the 'UseApplicationInsights' and revert to the original method it makes no difference. Either way I get the output debug window filled wit AI logs.
In fact, even if there is no method to load AI (i.e. I remove both the 'UseApplicationInsights' and 'AddApplicationInsightsTelemetry' I get the logs.
Thanks for any help.
You can opt out of telemetry (for debug, for example) by setting a DOTNET_CLI_TELEMETRY_OPTOUT environment variable to 1.
Visual Studio is lighting up Application Insights even if you have no code to enable it. You can create an environment variable, ASPNETCORE_PREVENTHOSTINGSTARTUP = True, to prevent Visual Studio from lighting up Application Insights.
How to do this?
Right click the project in VS, Select Properties.In Debug options add environment variable as shown in below screenshot.
I'm working on a Browser extension/add-on. We have it working in Chrome, so I'm trying to get it working in Firefox.
I've gotten my add-on to load in Firefox Developer Edition 49.0a2 (2016-07-25).
My extension involves a content_script set to run_at: document_start, so it can inject a script tag before other page scripts run, so it can make an object globally available to websites.
This has seemed to work fine in Chrome, but in Firefox it has proven to be a bit of a race condition, with other page resources loading first most of the time.
Is there a strategy to load a content script in a way that it can inject & load a script before any other page scripts run?
When I add logs, I can isolate what is happening pretty nicely. In this example content-script:
// inject in-page script
console.log('STEP 1, this always happens first')
var scriptTag = document.createElement('script')
scriptTag.src = chrome.extension.getURL('scripts/inpage.js')
scriptTag.onload = function () { this.parentNode.removeChild(this) }
var container = document.head || document.documentElement
// append as first child
container.insertBefore(scriptTag, container.children[0])
Now if the file scripts/inpage.js simply runs a log, like
console.log('STEP 2, this should always run second')
And I visit a page with a script like this:
console.log('Step 3, the page itself, should run last')
In practice, Step 2 and Step 3 run in a non-deterministic order.
Thanks a lot!
I have Firefox-compatible version of the script in a public repository on a special branch if you dare to try it yourself: https://github.com/MetaMask/metamask-plugin/tree/FirefoxCompatibility
An dynamically inserted script with an external source (<script src>) does not block the execution of scripts, so there is no guarantee that your script would load. If your extension worked in Chrome, it was just by sheer luck.
If you really want to run some script before the rest, you have to run it inline:
var actualCode = `
// Content of scripts/inpage.js here
`;
var s = document.createElement('script');
s.textContent = actualCode;
(document.head || document.documentElement).appendChild(s);
s.remove();
Ideally, your build script would read scripts/inpage.js, serialize it to a string and put it in the actualCode variable. But if inpage.js is just a few lines of code, then the above can be used.
Note that you should not inject code in the web page unless it is absolutely necessary. The reason for that is that the execution environment of the page is untrusted. If you inject at document_start, then you can save functions and (prototype) methods that use for later (in a closure), but very careful coding is required.
If your content script is not generated by a build script and you still want to keep the scripts separate, then you can also use synchronous XMLHttpRequest to fetch the script. Synchronous XHR is deprecated for performance reasons, so use it at your own risk. Extension code is typically bundled with your extension, so the use of sync xhr should be low-risk:
// Note: do not use synchronous XHR in production!
var x = new XMLHttpRequest();
x.open('GET', chrome.runtime.getURL('scripts/inpage.js'), false);
x.send();
var actualCode = x.responseText;
var s = document.createElement('script');
s.textContent = actualCode;
(document.head || document.documentElement).appendChild(s);
s.remove();
If you are using a bootstrap.js based addon you can use a framescript and DOMWindowCreated to work with the document before even the HTML DOM (past basics of document.body etc) renders - https://developer.mozilla.org/en-US/Firefox/Multiprocess_Firefox/Frame_script_environment#Events - the innerHTML will be available but no script would have executed. You can put your inline script at the top as #Rob mentioned.
Using Intern JS with WebDirver to communicate with Appium 1.4.1, I have a simple functional test, part of which should find a text input box on the login screen of our iOS app, tap into it and type the user's login name:
define([
'intern!object',
'intern/chai!assert',
'require'
], function (registerSuite, assert, require) {
registerSuite({
name: 'Suite Name',
'Login Screen': function () {
return this.remote
.setFindTimeout(50000)
.findByXpath("//UIAApplication[1]/UIAWindow[1]/UIAScrollView[1]/UIATextField[1]/UIATextField[1]")
.tap()
.type('student1#demo.com')
.end()
.then(function (pagetitle) {
assert.strictEqual(pagetitle, 'DEMO COLLEGE',
'Should land on app dashboard with school title displayed');
});
}
});
});
When executed using intern-runner everything appears to go well, Appium launches our app and awaits my input - but no matter what I try I can't find the element I need to type into using Leadfoot's findByXpath:
$ intern-runner config=tests/appium
Listening on 0.0.0.0:9000
Starting tunnel...
Initialised iOS on MAC 8.1
Test main - Suite Name - Login Screen FAILED on iOS on MAC 8.1:
NoSuchElement: [POST http://[appium-server-address]:4723/wd/hub/session/80e20453-452e-4181-8713-4f9e0cfa427f/element / {"using":"xpath","value":"//UIAApplication[1]/UIAWindow[1]/UIAScrollView[1]/UIATextField[1]/UIATextField[1]"}] An element could not be located on the page using the given search parameters.
at Server._post <../../.nvm/v0.10.35/lib/node_modules/intern/node_modules/leadfoot/Server.js:68:9>
Using Appium's GUI "inspector" I've confirmed the Xpath to the text-input-box element is:
"//UIAApplication[1]/UIAWindow[1]/UIAScrollView[1]/UIATextField[1]/UIATextField[1]"
Can anyone suggest where I'm going wrong please?
Looks like the Appium Inspector tool on my Mac was duplicating part of the path?
.findByXPath("//UIAApplication[1]/UIAWindow[1]/UIAScrollView[1]/UIATextField[1]/**UIATextField[1]**")
.type("wibble")
Remove the duplicated element in bold (the final UIATextField[1]) ....
.findByXPath("//UIAApplication[1]/UIAWindow[1]/UIAScrollView[1]/UIATextField[1]/**UIATextField[1]**")
.type("wibble")
And it works.
Thanks Appium Inspector. Thanks a lot.
Are there any ideas how I can setup my Dart app to use different settings in debug mode (running in Dartium) and production mode?
For example, I'm using PouchDb in my app, that replicates the database to a particular CouchDb instance, given by an url: db.replicateTo(url);
In debug mode, I would like to use another CouchDb instance (another url) than in production mode.
So, are there any ideas or approaches, to use different setups in both modes?
this works since a short while:
transformers: # or dev_transformers
- $dart2js:
environment: { PROD: "true" }
access it from the code like
String.fromEnvironment()
main() {
print('PROD: ${const String.fromEnvironment('PROD')}');
// works in the browser
// prints 'PROD: null' in Dartium
// prints 'PROD: true' in Chrome
}
see also
Configuring the Built-in dart2js Transformer
How to achieve precompiler directive like functionality
https://github.com/dart-lang/pub/issues/798
http://blog.sethladd.com/2013/12/compile-time-dead-code-elimination-with.html
https://api.dartlang.org/stable/1.22.0/dart-core/bool/bool.fromEnvironment.html
dart check if is building