Problem
I'm testing downloading a file, but when I trigger the download, the "Save as..." prompt appears.
I saw a solution in this SO question but it doesn't seem to work for me.
Config
My protractor config file looks like this (coffeescript):
exports.config =
capabilities:
browserName: "chrome"
shardTestFiles: true
maxInstances: 2
chromeOptions:
args: ['--no-sandbox', '--test-type=browser']
prefs:
download:
prompt_for_download: false
default_directory: '/'
default_content_settings:
popups: 0
More
On chromeOptions.pref webdriver docs states:
See the 'Preferences' file in Chrome's user data directory for examples.
I can't actually see default_directory in my own Chrome preferences file.
"download": {
"directory_upgrade": true,
"prompt_for_download": false
},
System
Protractor: Version 1.5.0 (pretty new)
Node: 0.10.28, 0.11.8 and 0.11.14
Provide an absolute path to an existing directory in default_directory chrome preference.
For me, the Save As prompt was appearing when I mentioned a slash at last in the path for example: A:\Proj\Downloads\ & didn't see any prompt when mentioned without a slash at the end of the path i.e A:\Proj\Downloads
Related
I am using ESLint in my Vue(Nuxt) project in VSCode. When I save I would like my ESLint to run automatically and fix all the warnings for me automatically.
This is my settings.json file in VSCode:
{
"editor.codeActionsOnSave": {
"source.fixAll.eslint": true
},
"vetur.format.defaultFormatter.js": "vscode-typescript",
"vetur.validation.template": false,
"vetur.completion.scaffoldSnippetSources": {},
"vetur.completion.useScaffoldSnippets": false,
"vetur.format.defaultFormatter.html": "none",
"workbench.iconTheme": "material-icon-theme",
"git.autofetch": true,
"git.defaultCloneDirectory": "",
"gitlens.views.repositories.files.layout": "list",
"editor.tabSize": 2,
"editor.detectIndentation": false,
}
And this is my .eslintrc.js file:
module.exports = {
root: true,
env: {
browser: true,
node: true,
},
extends: [
"#nuxtjs",
"plugin:nuxt/recommended",
"../.eslintrc.js"
],
rules: {
//Add custom rules for the frontend here.
//Rules that are common for shared, frontend and backend should go in the parent file
"nuxt/no-cjs-in-config": "off",
},
}
The linked ../.eslintrc.js file contains the following:
module.exports = {
parserOptions: {
parser: 'babel-eslint',
},
plugins: ['jest'],
rules: {
'prefer-const': 'off',
'comma-dangle': ['error', 'always-multiline'],
'prefer-template': 'error',
},
env: {
'jest/globals': true
}
}
Whenever I save the file the warnings just show up and will not automatically fix themselves.
EDIT:
I've turned on verbose output and i'm seeing this in the output:
(node:6832) UnhandledPromiseRejectionWarning: Error: Failed to load plugin 'import' declared in 'frontend\.eslintrc.js » #nuxtjs/eslint-config': Cannot find module 'eslint-plugin-import'
Require stack:
I've then ran yarn add eslint-plugin-import and tried again, it still returns the same error.
Get eslint plugin, add this code to your settings.json
{
"editor.codeActionsOnSave": {
"source.fixAll.eslint": true
},
"eslint.validate": ["javascript"]
}
source
Launch VSCode,
Command + Shift + P, type settings and hit enter, paste and save the following:
{
"editor.codeActionsOnSave": {
"source.fixAll.eslint": true,
}
}
You're good to go!
I've managed to fix the issue.
The problem was that there were multiple working directories in my solution, which all have their own eslint config.
Putting the following line in the settings.json file of VSCode solved my issue:
"eslint.workingDirectories": [{ "mode": "auto" }]
I tried those solutions and others, and it still didn't work. Actually it was just that ESLint's use had to be approved for use in VSCode. That is, I clicked on the ESLint item on the editor's bottom bar:
Which opened a popup asking me to approve ESLint. After approval autocorrect was running as expected.
Install ESLint extension from the VSCode marketplace.
Once the ESLint extension has installed you may use CTRL + SHIFT + P to open the Command Palette. Search “ESLint fix all auto-fixable Problems” and press enter.
This command would enable eslint to fix the file on save.
In the snap above as you can see that I am getting eslint errors and just to inform you all that despite saving the file, all auto-fixable problems are were not getting fixed by eslint/prettier setup.
So I tried pressing ctrl+shift+p and selecting prettier as default formatter and also tried doing eslint restart server but that didn't worked.
I noticed that vscode was giving me some notifications at the bottom right corner (bell icon). I clicked on that and some list of pop up came up stating that there are multiple formatters installed for the same extension file. Like for example in the below snap there is .jsx file(it had two formatters one was prettier and other was vscodes inbuilt formatter). I clicked on configure button and selected prettier as default and when I saved the file it worked!
If this doesn't works for you then I think this all worked for me because I had eslint npm packages installed in my project that works with prettier to enforce the prettier rules. (these packages are eslint-config-prettier and eslint-plugin-prettier)
I ran into a similar problem-- ESLint was not properly formatting only certain, seemingly random, files on save. Running ESLint --fix would fix the formatting errors, but saving would not. Adding this line to our workspace settings.json fixed the problem:
"eslint.format.enable": true,
Making all our formatter settings look like this:
"editor.formatOnSave": true,
"editor.codeActionsOnSave": {
"source.fixAll": true
},
"eslint.format.enable": true,
You can also go into the ESLint extension settings and check off the checkbox labeled ESLint > Format:Enable. Worked like a charm!
What fixed it for me was adding this to settings.json, because VSCode didn't know what formatter I wanted to be used on save:
"[javascript]": {
"editor.defaultFormatter": "dbaeumer.vscode-eslint"
},
Check if in the settings.json there are other formatters enabled, in my case I had this by mistake.
"[javascript]": {
"editor.defaultFormatter": "vscode.typescript-language-features"}
After getting the Eslint plugin you need to add this line to the settings.json:
{
"editor.codeActionsOnSave": {
"source.fixAll.eslint": true
},
}
Still not working? check if your eslint works fine by running this in the terminal:
eslint --ext \".js,.vue\" --ignore-path .gitignore .
If it failed with exit code 2 try removing node modules and install again.
After running this command you should see the eslint errors.
I was dealing with the same issue, and nothing seemed to help, even though I did all the configurations correctly, and ESLint was marking the problems in my files correctly.
For me the solution was to move the .vscode folder to the project root.
Now everything works as intended.
I'm trying to get trace logging information out of the gecko driver, in order to debug some issues we're having with moving to newer versions of Python, Robot Framework, Selenium, Firefox.
See in the DEBUG robot output below, I'm setting the capabilities to include log level arguments as part of moz:firefoxOptions.
When the capabilities are printed out for Open Browser, the moz:firefoxOptions value has been replaced with the profile info, so I never get tracing set.
20190219 13:07:22.461 - INFO - {'browserName': 'firefox', 'marionette': True, 'acceptInsecureCerts': True, 'moz:firefoxOptions': {'prefs': {'log': {'level': 'trace'}}, 'args': ['-headless', '-profile', '/tmp/tmpimda__f2']}, 'binary': '/usr/bin/firefox'}
20190219 13:07:22.461 - INFO - +--- END KW: BuiltIn.Log (1)
20190219 13:07:22.461 - INFO - +--- START KW: CustomLib.AdminClass.Open Browser [ ${LOGIN URL} | ${BROWSER} | desired_capabilities=${dc} | ff_profile_dir=${profile} ]
20190219 13:07:22.462 - INFO - Opening browser 'Firefox' to base url 'https://HOST:PORT/bmmadmin/login.jsp'.
20190219 13:07:22.480 - DEBUG - POST http://127.0.0.1:38132/session {"capabilities": {"firstMatch": [{}], "alwaysMatch": {"browserName": "firefox", "acceptInsecureCerts": true, "moz:firefoxOptions": {"profile": "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"}}}, "desiredCapabilities": {"browserName": "firefox", "acceptInsecureCerts": true, "moz:firefoxOptions": {"profile": "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"}, "binary": "/usr/bin/firefox", "marionette": true}}
Anyone have any ideas what I may be doing wrong here? And/OR some alternative methods of getting trace logging from geckodriver?
BTW: the moz:firefoxOptions I'm setting may not be correct, they're just the latest combination I have tried, but I can't tell what is going to work until the options I set are actually being sent to firefox.
Environment:
Python 3.6.3
geckodriver 0.24.0 ( 2019-01-28)
Mozilla Firefox 60.4.0
PIP:
requests (2.21.0)
robotframework (3.1.1)
robotframework-requests (0.5.0)
robotframework-selenium2library (3.0.0)
robotframework-seleniumlibrary (3.3.1)
selenium (3.141.0)
setuptools (40.8.0)
urllib3 (1.24.1)
ROBOT:
Open Browser To Login Page
${dc}= CustomLib.Get Desired Capabilities ${BROWSER}
Set To Dictionary ${dc}
... binary ${PATH TO MARIONETTE}
Log ${dc}
${profile}= CustomLib.Create FF Profile
Open Browser ${LOGIN URL} ${BROWSER} desired_capabilities=${dc} ff_profile_dir=${profile}
PYTHON:
#keyword
def get_desired_capabilities(browser):
_capabilities = {'ff' : DesiredCapabilities.FIREFOX,
'firefox' : DesiredCapabilities.FIREFOX
}
browserKey = str(browser).lower()
desired = _capabilities[browserKey]
BuiltIn().log("get_desired_capabilities B4 :" + str(desired))
#THIS IS ASSUMING FIREFOX
desired['marionette'] = True
desired['acceptSslCerts'] = True
desired['ignoreProtectedModeSettings'] = True
desired['headless'] = True
desired['moz:webdriverClick'] = True
desired['moz:firefoxOptions'] = {'log':{'level':'trace'}, 'args':['-headless']}
#desired['moz:firefoxOptions']['log'] = {'level':'trace'}
#desired['moz:firefoxOptions']['args'] = ['-headless']
BuiltIn().log("get_desired_capabilities returning :" + str(desired))
return desired
#keyword
def create_ff_download_profile(path):
from selenium import webdriver
BuiltIn().log('PATH: ' + path)
fp = webdriver.FirefoxProfile()
fp.set_preference("browser.download.folderList", 2)
fp.set_preference("browser.download.manager.showWhenStarting", False)
fp.set_preference("browser.helperApps.alwaysAsk.force", False);
fp.set_preference("browser.helperApps.neverAsk.saveToDisk", 'text/xml,text/plain,application/octet-stream')
fp.set_preference("browser.download.dir", path)
fp.update_preferences()
return fp.path
thanks,
Tom
This is how I set the log level to be trace in Firefox:
*** Settings ***
Library SeleniumLibrary
*** Keywords ***
${options}= Evaluate sys.modules['selenium.webdriver.firefox.options'].Options() sys
${options.log.level}= Set Variable trace
${driver}= Create Webdriver Firefox options=${options}
Go To https://www.google.com
It's basically following this python pattern from the firefox docs and translating it into Robotframework
Hope this helps!
A protractor spec running from a Jenkins job, that connects to SauceLabs. It clicks a button to download a PDF, and checks that the file successfully downloaded. I am unable to get the chrome browser to NOT open a "Save As" prompt using an absolute path. I AM able to avoid the "Save As" prompt if I use '~/Downloads' as filename, but then my browser.wait that waits for the file to exist hangs forever.
// spec.js
import fs from 'fs'
import path from 'path'
fs.mkdirSync('./downloads')
describe('Clicking DOWNLOAD button', () => {
it('should download a proposal', () => {
const filename = path.resolve(__dirname, './downloads/proposal.pdf')
if (fs.existsSync(filename)) {
fs.unlinkSync(filename)
}
page.downloadProposalBtn.click()
browser.wait(() => fs.existsSync(filename), 180000)
.then(() => {
expect(fs.existsSync(filename)).toBe(true)
})
}, 180000)
})
Below is the relevant portion of my conf file. I would expect the prompt_for_download setting to make the prompt not show, but it does..
// conf.js
capabilities: {
platform: 'macOS 10.12',
browserName: 'chrome',
version: '59.0',
screenResolution: '1400x1050',
chromeOptions: {
args: ['--no-sandbox', '--test-type=browser', 'disable-infobars=true'],
prefs: {
download: {
prompt_for_download: false,
directory_upgrade: true,
default_directory: path.resolve(__dirname, './downloads'),
},
credentials_enable_service: false,
},
},
},
Am I missing something? I feel like I might be misunderstanding where Saucelabs is running these tests from, but it would seem that given I fs.mkdir the ./downloads folder, then when I path.resolve it, that should work.
Unsure if this constitutes an "answer" but this is how I'm proceeding. After a lot of research I think that testing if a file downloaded on a remote VM (like SauceLabs) isn't possible. What I am doing instead is breaking the test into two parts:
Test the download button: click the button and assert that no error occurred
Make GET request to the underlying endpoint api/download that the button is using, write the response to a folder, and from there assert (using nodejs) that the file exists in my Jenkins project's workspace. This feels hacky, but given SauceLabs doesn't seem to give much access to the VM that the webdriver is running on, I don't see another way.
You will get the prompt, even though the option is set to "false" if the default download directory doesn't exist. "./downloads" is OSX syntax, which the chrome driver probably doesn't understand.
Try this instead:
default_directory: __dirname + '/downloads'
I'm trying to make test for AngularJS web page in WebStorm (using Jasmine and Protractor frameworks), I'm using JHint for code inspection...
All code is OK except one command: "browser", example of code:
describe('Test',function(){
it('Open page',function(){
browser.get('https://www.angularjs.org');
browser.sleep(2000);
});
});
JSHint is still highlighting errors with browser:
Problem synopsis JSHint: 'browser' is not defined. (W117)
Unresolved function or method sleep() at line 20
In JHint Environment I have enabled:
Jasmine
Node.js
In JavaScript Libraries I have enabled:
Node.js Core
angular-protractor-DefinitelyTyped
jasmine-DefinitelyTyped
selenium-webdriver-DefinitelyTyped
Does anybody know what do I have to enable or which Library do I have to download to make JSHint understand the "browser" command please?
Do you have a .jshintrc file ?
{
"globals": {
"browser": false,
},
"jasmine": true
}
You can add browser as a global.
JSHint works on per-file basis and doesn't 'see' global variables defined in other files unless they are added to 'global' list. This can be done by either adding the corresponding comments (/* global browser*/) to your files - see http://www.jshint.com/docs/, or by adding variables/functions you'd like to use globally to the 'Predefined' list in WebStorm Preferences -> Languages & Frameworks -> Javascript -> Code Quality Tool -> JSHint -> Predefined (,separated).
Using CMD from within Sencha Architect I've been able to build a production build of my application. However I can not seem to figure out how to exclude a js file from the build process. I don't want it compiled in with app.js I want it as a separate script include in index.html - so cmd shouldn't touch it basically.
Sencha Arhitech generates and calls build.xml which calls build-impl.xml which calls init-impl.xml
Everywhere I've read, they say to include the following;
<target name="-before-init">
<echo>Setting build.operations...</echo>
<echo>app.dir=${app.dir}</echo>
<property name="build.operations">
exclude
-file=\resources\js\version.js
</property>
</target>
However it refuses to exclude the file...I can see the echos so I know it's hitting the target..
Any ideas? Is this how I am supposed to exclude files?
app.framework.version=4.2.1.883
app.cmd.version=4.0.4.84
Turns out this won't be possible to do until Sencha Architect 3.1
Steps by which i was able to exclude AppConfig file in production build.
Here file exclude means it will not be compressed/bundled and variable/properties of this file could be used any where in the app.
1. Config file(AppConfig.js in our case) MUST be inside resources fodler. Below are the contents of our AppConfig file
/////////////IxDetect is my Application Namespace///////////////////
var IxDetect = IxDetect || {};
IxDetect.AppConfig = {
logoPath: '',
logoTitle: 'Internal',
pentahoUrl: 'http://107.20.104.150/pentaho',
pentahoRptCube: 'TrafficWithFraudIndex'
};
////////////////////////////////
2. Link this file in index.html page like below
<script src="resources/AppConfig.js"></script>
3. Add one more item in "js" array in "app.json" file
"js": [
{
"path": "resources/AppConfig.js", // This is my file. Also make a sure you do not miss bundle and includeInBundle property
"bundle": false,
"includeInBundle": true
},
{
"path": "app.js",
"bundle": true
}
],
4. Try development and production build all should work file
Note: All above changes are done and tested on 6.2(Framework/CMD)