Despite using the Dojo build system, my app is still including a large number of javascript files which I would have hoped to be covered by the build.
Here's my build profile:
var profile = (function(){
return {
basePath: "./",
releaseDir: "release",
action: "release",
selectorEngine: "acme",
cssOptimize: "comments.keepLines",
packages:[{
name: "dojo",
location: "dojo"
},{
name: "dijit",
location: "dijit"
},{
name: "dojox",
location: "dojox"
},{
name: "my",
location: "my"
}],
layers: {
"my/admin": {
include: ['dojo/ready', 'dojo/dom', 'dojo/query', 'dojo/request/xhr', 'my/Form', 'my/Tree/Radio']
}
}
};
})();
The app is still including the following JS files on each request: my/Form.js (even though this is listed in the profile), dojo/fx/Toggler.js, dijit/_base.js, dijit/WidgetSet.js, dijit/_base/focus.js, dijit/_base/place.js, dijit/place.js, dijit/_base/popup.js, dijit/popup.js, dijit/BackgroundIframe.js, dijit/_base/scroll.js, dijit/_base/sniff.js, dijit/_base/typematic.js, dijit/typematic.js, dijit/_base/wai.js, dijit/_base/window.js.
my/Tree/Radio extends dijit/Tree, so I'm assuming a lot of the files above are dijit base files that are being loaded automatically by dijit.Tree. But surely the build tool should resolve dependencies like this and include them in the 'built' file?
I am using Dojo 1.8.3.
In dojo/fx, it dynamically looks up the Toggler with the comment
use indirection so modules not rolled into a build
Not sure why, but if you add dojo/fx/Toggler to the include of your build script, it should not make the additional xhr requests.
EDIT: Apparently dijit/Widget does something similar with dijit/_base, so you will want to add that to the includes as well.
http://trac.dojotoolkit.org/ticket/14262
Related
This is very picky of me, I know, but I have a large Angular application where many files are covered at 100% or pretty darn close to it, and I want to exclude those results, so I can just deal with the ones I need to and don't have all the extra noise.
I don't want to exclude files by name because if they change they may go below the OK threshold.
My karma.conf.js (I'm aware I may not need some of the plugins, this is a shared file by the team):
// Karma configuration file, see link for more information
// https://karma-runner.github.io/1.0/config/configuration-file.html
module.exports = function (config) {
config.set({
basePath: '',
frameworks: ['jasmine', '#angular-devkit/build-angular'],
plugins: [
require('karma-jasmine'),
require('karma-chrome-launcher'),
require('karma-jasmine-html-reporter'),
require('karma-coverage'),
require('#angular-devkit/build-angular/plugins/karma')
],
reporters: ['progress', 'coverage'],
coverageReporter: {
reporters: [
{type: 'text'},
{type: 'text-summary'},
]
},
browsers: ['ChromeHeadless'],
preprocessors: {'**/*.ts': ['coverage']},
restartOnFileChange: true,
});
};
and I'm gonna try to attach a screenshot here of my sample output in the type: 'text' coverageReporter:
I've tried searching SO and karma / karma-coverage documentation, but I only can find excluding files/paths specifically by name, or updating thresholds, but the latter seems to be only for determining what colors show up. Thank you~
EDIT: And since I only have one file per directory, would also be useful to not duplicate the numbers by printing them for the directory AND the one file under it. Thinking I may just have to dig into the weeds and create a pull request or something.
For this you can generate coverage reports in HTML format. There you will get option to sort files based on code-coverage.
To enable such reports you will have to do update coverageReporter property in your karma configurations -
coverageReporter: {
includeAllSources: true,
dir: 'coverage/',
reporters: [
{type: 'text'},
{type: 'text-summary'},
{type: 'html', subdir: 'html/'}
]
},
Reports will be generated within coverage/html folder.
I use 2 entry files, startpage.js and subpage.js, and assign them successfully to their HTML files via the HTMLWebpackPlugins chunks parameter.
But since that solution requires to include the CSS files in both the startpage.js and subpage.js, which makes for "double the trouble" (at least during the build process), I decided to create another file, app_head, and put the import 'main.css' statement there. (And I also have a vendor file that should be placed in the header of the HTML, which should happen by adding _head according to the documentation: https://github.com/architgarg/html-webpack-injector - but that does not work either...)
This is the current webpack config (excerpt):
module.exports = {
entry: {
app_head: './src/css/main.css',
vendor_head: './src/scripts/vendor/_all_vendor.js',
startpage: './src/scripts/startpage.js',
subpage: './src/scripts/subpage.js',
},
output: {
filename: '[name].bundle.js',
path: path.resolve(__dirname, '../public'),
publicPath: '/'
},
plugins: [
new HtmlWebpackPlugin({
filename: 'index.html',
template: './src/templates/index.ejs',
chunks: ['app_head', 'vendor_head', 'startpage'],
chunksConfig: {
defer: ['startpage']
},
excludeChunks: ['subpage'],
bodyCss: 'is-startpage',
}),
new HtmlWebpackPlugin({
filename: 'publication.html',
template: './src/templates/publication.ejs',
chunks: ['subpage'],
chunksConfig: {
defer: ['subpage']
},
excludeChunks: ['startpage'],
bodyCss: 'is-subpage',
}),
[...]
The app_head.css is placed properly in the head section of the HTML, but it also generates a useless app_head.js, which only contains webpack code. Unfortunately, I do not know of any way how to exclude that file without also excluding the CSS.
Is there a better way to separate the CSS generation process without producing overhead or garbage?
The problem is you did not initialized the html-webpack-injector plugin.
// import above
const HtmlWebpackInjector = require('html-webpack-injector');
plugins: [
new HtmlWebpackPlugin(...),
new HtmlWebpackPlugin(...),
new HtmlWebpackInjector() // add this in plugins array
Running Karma + Jasmine Tests with RequireJS -- Getting off the ground
Help! . . . _ _ _ . . . SOS!
Currently, I have an exercise project up for getting comfortable with KarmaJS -- and Unit Testing, at large. The broad issue is that I really have no transparent view of what Karma is doing behind the scenes, and I can't seem to find adequate documentation in relevant areas. Without further delay...
Here is my folder structure:
root
|-/lib
|-/[dependencies] (/angular, /angular-mocks, /bootstrap, /etc) # from bower
|-/src
|-/[unreferenced directories] (/js, /css, /views) # not referenced anywhere
|-app.js # sets up angular.module('app', ...)
|-globals.js # may be referenced in RequireJS main file; not used.
|-index.html # loads bootstrap.css and RequireJS main file
|-main.js # .config + require(['app', 'etc'])
|-routeMap.js # sets up a single route
|-test-file.js # *** simple define(function(){ return {...}; })
|-/test
|-/spec
|-test-test-file.js # *** require || define(['test-file'])
|-.bowerrc # { "directory": "lib" }
|-bower.json # standard format
|-karma.conf.js # *** HELP!
|-test-main.js # *** Save Our Souls!!!
karma.conf.js
// Karma configuration
// Generated on Wed Nov 19 2014 15:16:56 GMT-0700 (Mountain Standard Time)
module.exports = function(config) {
config.set({
// base path that will be used to resolve all patterns (eg. files, exclude)
basePath: '',
// frameworks to use
// available frameworks: https://npmjs.org/browse/keyword/karma-adapter
frameworks: ['jasmine', 'requirejs'],
// list of files / patterns to load in the browser
files: [
//'test/spec/test-test-file.js',
//'lib/**/*.js',
//'src/**/*.js',
//'test/spec/**/*.js',
'test-main.js',
{pattern: 'lib/**/*.js', included: false},
{pattern: 'src/**/*.js', included: false},
{pattern: 'test/spec/*.js', included: true}
],
// list of files to exclude
exclude: [
'lib/**/!(angular|angular-mocks|angular-resource|angular-route|require|text).js',
'lib/**/**/!(jquery|bootstrap).js',
'src/app.js'
],
// preprocess matching files before serving them to the browser
// available preprocessors: https://npmjs.org/browse/keyword/karma-preprocessor
preprocessors: {
},
// test results reporter to use
// possible values: 'dots', 'progress'
// available reporters: https://npmjs.org/browse/keyword/karma-reporter
reporters: ['progress'],
// web server port
port: 9876,
// enable / disable colors in the output (reporters and logs)
colors: true,
// level of logging
// possible values: config.LOG_DISABLE || config.LOG_ERROR || config.LOG_WARN || config.LOG_INFO || config.LOG_DEBUG
logLevel: config.LOG_INFO,
// enable / disable watching file and executing tests whenever any file changes
autoWatch: true,
// start these browsers
// available browser launchers: https://npmjs.org/browse/keyword/karma-launcher
browsers: ['Chrome'],
// Continuous Integration mode
// if true, Karma captures browsers, runs the tests and exits
singleRun: false
});
};
test-main.js
var allTestFiles = [];
var TEST_REGEXP = /(spec|test)\.js$/i;
var pathToModule = function(path) {
return path.replace(/^\/base\//, '').replace(/\.js$/, '');
};
Object.keys(window.__karma__.files).forEach(function(file) {
if (TEST_REGEXP.test(file)) {
// Normalize paths to RequireJS module names.
allTestFiles.push(pathToModule(file));
}
});
require.config({
// Karma serves files under /base, which is the basePath from your config file
baseUrl: '/base/src',
paths: {
angular: '../lib/angular/angular',
ngRoute: '../lib/angular-route/angular-route',
jquery: '../lib/jQuery/dist/jquery',
bootstrap: '../lib/bootstrap/dist/js/bootstrap',
models: 'models',
controllers: 'controllers',
globals: 'globals',
routeMap: 'routeMap'
},
shim: {
angular: {
exports: 'angular'
},
ngRoute: {
deps: ['angular']
},
jquery: {
exports: '$'
},
bootstrap: {
deps: ['jquery']
}
},
// dynamically load all test files
deps: allTestFiles,
// we have to kickoff jasmine, as it is asynchronous
callback: window.__karma__.start
});
test-test-file.js
console.log('....................');
define(function(){
//console.log('testing test-file', testFile);
describe('Testing testing', function(){
it('should work', function(){
expect(true).toEqual(true);
});
});
});
test-file.js
define('testFile', [], function(){
return function init(sandbox){
var app, application = app = sandbox.app
, globals = sandbox.globals;
return {
some: 'module'
};
};
});
Questions & Descriptions
Key points I would love to hear answers for are
what does { pattern: '...', include: true|false } do?
best way to exclude all the extra stuff inside the bower directories.
what files do I need to include in the test-main.js file?
what files do I need to include in the karma.conf.js file?
what does test-main.js actually do; what's it for?
The times I receive errors & issues is as soon as I wrap my spec in a define(...) call -- event when I give the module an ID -- define('someId', function(){ ... }) -- do I need to return something out of this module, as it is a define call?
Other times, I receive the 'ol ERROR: 'There is no timestamp for /base/src/app.js!'. "Timestamp, of course! How silly of me..." -- what in the world does this mean?! Sometimes I get the infamous "Executed 0 of 0 ERROR" -- I could also use some clarity here, please. Really, I get plenty of ERROR: '...no timestamp...' errors -- and even 404s when it seems I should be pulling that library in with the karma.conf.js files config...???
It even seems that usually when I explicitly tell karma to excludesrc/app.js I still get 404s and errors.
tl;dr
Obviously, I'm a bit of a confused novice about Karma and *DD at large...
I can run test-test-file.js fine when my karma.conf.js files array looks like [ 'test-main.js', 'test/spec/test-test-file.js' ] -- but, still, if I wrap my test in a RequireJS define call I get the "Mismatching anonymous define()" error mentioned above.
It seems that when I add { pattern: '...', include: false } then karma just doesn't add any of my files for the given pattern whatsoever (???).
If someone can even simply direct me toward how to use RequireJS with Karma -- namely so that I can just wrap my tests in a define/require call and pull in the module I want to test... That would be greatly appreciated.
As its somewhat difficult to keep these types of questions short and still provide adequate information, I hope I didn't make it too long.
EDIT
After reading the answer from glepretre and some fiddling on my own, I reconfigured my project as follows:
Moved test-main.js to test/test-main.js,
renamed test-test-file.js to testFileSpec.js -- moved it from test/spec to test/,
karma.conf.js:
...
// list of files / patterns to load in the browser
files: [
{pattern: 'lib/**/*.js', included: false},
{pattern: 'src/**/*.js', included: false},
{pattern: 'test/**/*Spec.js', included: false},
'test/test-main.js'
],
....
test/test-main.js:
/* **************** HOW COME THE DEFAULT (Karma-generated) CONFIGURATION DOES ***NOT WORK???
var allTestFiles = [];
var TEST_REGEXP = /(spec|test)\.js$/i;
var pathToModule = function(path) {
return path.replace(/^\/base\//, '').replace(/\.js$/, '');
};
Object.keys(window.__karma__.files).forEach(function(file) {
if (TEST_REGEXP.test(file)) {
// Normalize paths to RequireJS module names.
allTestFiles.push(pathToModule(file));
}
});
*/
var tests = [];
for (var file in window.__karma__.files) {
if (/Spec\.js$/.test(file)) {
tests.push(file);
}
}
require.config({
// Karma serves files under /base, which is the basePath from your config file
baseUrl: '/base/src',
paths: {},
shim: {},
// dynamically load all test files
//deps: allTestFiles,
//
deps: tests,
// we have to kickoff jasmine, as it is asynchronous
callback: window.__karma__.start
});
I am now running Unit Tests successfully! Special Thanks to glepretre & all other contributors.
Thanks for any insight at all :)
OK, I will try to address each question at a time:
Question 1
what does { pattern: '...', included: true|false } do?
Karma's default behavior is to:
find all files matching the pattern (mandatory property)
watch them for changes (watched option) in order to restart your unit tests to give you live result when you are editing your code (works if only you leave the default autoWatch default value to true).
serve them using its own webserver (served option)
include them in the browser using <script> (included option)
So, in the files array of the karma config you can use default behavior by adding only string patterns:
files: [
// this will match all your JS files
// in the src/ directory and subdirectories
'src/**/*.js'
]
Or use the full object syntax to customize each option:
files: [
{pattern: 'src/**/*.js', watched: true, served: true, included: false}
]
Using requireJS, you DO NOT want them to be included because it will be in conflict with requireJS behavior!
Included. Description: Should the files be included in the browser using <script> tag? Use false if you want to load them manually, eg. using Require.js.
from karma/config/files docs
NB: Pay attention to the order in which you add the files/patterns in the array. It matters! For more understanding, set logLevel: config.LOG_DEBUG in your karma config.
Question 2
what files do I need to include in the karma.conf.js file?
At least all required files for the proper functioning of your components for unit testing.
Basically, all files listed in your define([]) and require() blocks.
Question 3
best way to exclude all the extra stuff inside the bower directories.
What are you trying to do exactly?
Based on what I wrote before, you can see that you can add selectively the files you will need in your tests.
I use to add the pattern '/bower_components/**/*.js' and even '/bower_components/**/*.html' when my bower packages are using templates. I never noticed any significant performance issue if that's what you are worried about... Up to you to define the file patterns you will need.
Question 4 & 5
what does test-main.js actually do; what's it for?
what files do I need to include in the test-main.js file?
The purpose of the test-main.js file is to find and load your test files before starting Karma. It connects the dots between Karma and requireJS
You must choose a convention to name your test files and then define the TEST_REGEXP to match all of them.
The "official" angular style guide and best practices for app structure recommends using the suffix *_test.js.
EDIT: Your regexp is not working because it is defined to catch "spec.js" || "test.js" at the end or your spec file name is ending by "file.js" ;) Please see http://regex101.com/r/bE9tV9/1
One more thing
I hope I was clear enough. You can have a look at our starter app structure for our projects using Angular + Require: angular-requirejs-ready. It's already set up and tested with both Karma and Protractor.
I'm hopelessly trying to make the Dijit template inlining functionality of Dojo builds for my AMD project work with no luck yet ...
The particular issue isn't the inlining of the HTML templates themselves, but the fact that they are still requested with Ajax (XHR) after being successfully inlined.
Templates are inlined the following way :
"url:app/widgets/Example/templates/example.html": '<div>\n\tHello World!</div>'
The Dijit widget itself, after building, defines templates like this :
define("dojo/_base/declare,dijit/_Widget,dojo/text!./templates/example.html".split(","), function (f, g, d) {
return f("MyApp.Example", [g], {
templateString: d,
});
});
I tried to build with :
the shrinksafe / closure optimiser
relative / absolute paths
using the old cache() method
using the templatePath property
but even after having run a successful build (0 errrors and a few warnings) where the templates were inlined, Dojo / Dijit still makes Ajax requests to these resources.
Here is my build profile :
var profile = {
basePath: '../src/',
action: 'release',
cssOptimize: 'comments',
mini: true,
optimize: 'closure',
layerOptimize: 'closure',
stripConsole: 'all',
selectorEngine: 'acme',
internStrings: true,
internStringsSkipList: false,
packages: [
'dojo',
'dijit',
'dojox',
'app'
],
layers: {
'dojo/dojo': {
include: [
'app/run'
],
boot: true,
customBase: true
},
},
staticHasFeatures: {
'dojo-trace-api': 0,
'dojo-log-api': 0,
'dojo-publish-privates': 0,
'dojo-sync-loader': 0,
'dojo-xhr-factory': 0,
'dojo-test-sniff': 0
}
};
Due to this issue my application is completely unusable because there are so many files to download separately (browsers have a limit on the number of parallel connections).
Thank you very much in advance !
UPDATE :
The two lines loading dojo.js and the run.js in my index.html :
<script data-dojo-config='async: 1, tlmSiblingOfDojo: 0, isDebug: 1' src='/public/dojo/dojo.js'></script>
<script src='/public/app-desktop/run.js'></script>
Here is the new build-profile :
var profile = {
basePath: '../src/',
action: 'release',
cssOptimize: 'comments',
mini: true,
internStrings: true,
optimize: 'closure',
layerOptimize: 'closure',
stripConsole: 'all',
selectorEngine: 'acme',
packages : [
'dojo',
'dijit',
'app-desktop'
],
layers: {
'dojo/dojo': {
include: [
'dojo/request/xhr',
'dojo/i18n',
'dojo/domReady',
'app-desktop/main'
],
boot: true,
customBase: true
}
},
staticHasFeatures: {
'dojo-trace-api': 0,
'dojo-log-api': 0,
'dojo-publish-privates': 0,
'dojo-sync-loader': 0,
'dojo-xhr-factory': 0,
'dojo-test-sniff': 0
}
};
My new run.js file :
require({
async: 1,
isDebug: 1,
baseUrl: '/public',
packages: [
'dojo',
'dijit',
'dojox',
'saga',
'historyjs',
'wysihtml5',
'app-shared',
'jquery',
'jcrop',
'introjs',
'app-desktop'
],
deps: [
'app-desktop/main',
'dojo/domReady!'
],
callback: function (Main) {
debugger;
var main = new Main();
debugger;
main.init();
}
});
and my main.js file looks like this :
define([
'dojo/_base/declare',
'app-desktop/widgets/Application',
'app-desktop/config/Config',
'saga/utils/Prototyping',
'dojo/window',
'dojo/domReady!'
], function (declare, Application, ConfigClass, Prototyping, win) {
return declare([], {
init: function() {
// ... other stuff
application = new Application();
application.placeAt(document.body);
// ... some more stuff
}
});
});
In build-mode, I get the following error :
GET http://localhost:4000/app-desktop/run.js 404 (Not Found)
which is weird because it means that the build process made it so that dojo has an external dependency rather than an a already inlined dojoConfig variable in the builded file.
In normal-mode, files get requested, but the app is never created.
In both cases none of the two debuggers set in the run.js file were run which means that the callback method was never called for some reason.
Thank you for your help !
I've printed values of requireCacheUrl and require.cache to console in the method load() of dojo/text.js. At least in my case, keys of my templates in the cache differs from lookup keys on one leading slash.
For example, I have "dojo/text!./templates/Address.html" in my widget. It present with key url:/app/view/templates/Address.html in the cache but is searched like url:app/view/templates/Address.html, causing cache miss and xhr request.
With additional slash in dojo/text.js (line 183 for version 1.9.1) it seemingly works (line would looks like requireCacheUrl = "url:/" + url).
Not sure what kind of bugs this "fix" could introduce. So, it probably worth to report this issue to dojo folks.
UPD: Well, I see you've already reported this issue. Here is the link: https://bugs.dojotoolkit.org/ticket/17458.
UPD: Don't use hack described above. It was only attempt to narrow issue. Real issue in my project was with packages and baseUrl settings. Initially I created my project based on https://github.com/csnover/dojo-boilerplate. Then fixed it as in neonstalwart's sample.
This sounds like https://bugs.dojotoolkit.org/ticket/17141. If it is, you just need to update to Dojo 1.9.1.
I am new to Intern and struggling with trying to get a simple test to run in my environment. I was able to get the tutorial test to run but I've tried to set up a test where the test file is located inside my app directory hierarchy. The module being tested is located here:
sandbox/web/libs/ev/grids/FilterGrid.js
The test file is located here:
sandbox/web/libs/ev/tests/FilterGrid.js
My intern config file is located here:
sandbox/tests/intern.js
My loader and suites objects looks like this:
loader: {
packages: [
{ name: 'dojo', location: 'web/libs/dojo' },
{ name: 'dijit', location: 'web/libs/dijit },
{ name: 'dgrid', location: 'web/libs/dgrid' },
{ name: 'put-selector', location: 'web/libs/put-selector' },
{ name: 'xstyle', location: 'web/libs/xstyle' },
{ name: 'ev', location: 'web/libs/ev' }
]
},
suites: ['ev/tests/FilterGrid'],
When the loader tries to load this, I get:
Defaulting to "console" reporter
ReferenceError: document is not defined
at /home/bholm/Projects/src/sandbox/web/libs/dojo/selector/_loader.js:5:15
at execModule (/home/bholm/Projects/src/sandbox/node_modules/intern/node_module
/dojo/dojo.js:512:54)
at /home/bholm/Projects/src/sandbox/node_modules/intern/node_modules/dojo/dojo.js:579:7
at guardCheckComplete (/home/bholm/Projects/src/sandbox/node_modules/intern/node_modules/dojo/dojo.js:563:4)
at checkComplete (/home/bholm/Projects/src/sandbox/node_modules/intern/node_modules/dojo/dojo.js:571:27)
at onLoadCallback (/home/bholm/Projects/src/sandbox/node_modules/intern/node_modules/dojo/dojo.js:653:7)
at /home/bholm/Projects/src/sandbox/node_modules/intern/node_modules/dojo/dojo.js:746:5
at fs.js:266:14
at Object.oncomplete (fs.js:107:15)
Does the unit tests using Intern need a DOM document defined??
I also notice that Intern lists dojo2_core as it's dependency. So it's using unreleased code?
Any help with this would be appreciated!
It looks like you’re trying to load some code using the Node.js client that requires a browser environment. This won’t work. You should only load the ev/tests/FilterGrid test suite in a browser. You can do this by modifying your Intern configuration file to look something like this:
define([ 'intern/node_modules/dojo/has' ], function (has) {
var suites = [];
if (has('host-browser')) {
suites.push('ev/tests/FilterGrid');
}
return {
// ...your existing configuration...
suites: suites,
// ...
};
});