We're evaluating noflo to be executed on an embedded linux box using an simple javascript engine, being an interpreter (no JIT). In our case, the Node.js engine (with embedded V8 engine) might be too resource intensive.
The immediate question is the how to run the noflo runtime in there. Checking out the GitHub repository (https://github.com/noflo/noflo) and using grunt, we have generated the noflo for the browser using grunt build:browser.
As simple example to actually try and run the generated browser/noflo.js file, I used the d8 shell (V8 engine shell) for an isolated Javascript engine outside the Node.js universe, and appended the following code to the noflo.js generated file:
var fbpData = "<some FBP language connections>";
var noflo = require('noflo');
noflo.graph.loadFbp(fbpData, function(graph) {
print("Graph loaded");
});
Then,
d8 noflo.js
on the Linux shell, which reports
rtm.js:9559: TypeError: undefined is not a function
noflo.graph.loadFbp(fbpData, function(graph) {
^
TypeError: undefined is not a function
at rtm.js:9559:13
Without knowing further, leads me to believe that the noflo.js is not self-contained with all core noflo runtime functionality.
What necessary steps are missing here, for me to get noflo library running in an isolated JS engine (V8 is just an example - it could be any engine the is ECMA V5 compliant)
All code examples on the noflo project web site are tailored for Node.js...
PS: I tried as an alternative to build a browser-based noflo from http://noflojs.org/download/, however this always returns "server error".
Best regards
Gunther Strube
The NoFlo-Gnome project contains a browser build of the noflo-runtime-base repository (https://github.com/noflo/noflo-runtime-base) which itself embeds NoFlo.
You might need to add some aliases because the browser build doesn't necessarily fit your engine : https://github.com/noflo/noflo-gnome/blob/master/src/noflo.js#L89
noflo-gnome runs NoFlo in GJS, which is based on Spidermonkey and GLib/GObject.
It has some minimal require() compatibility which allows pulling in NoFlo. There is a checked in build of noflo (+ noflo-runtime-base) in ./src/libs but I did not immediately find how this is created.
If you're considering using a browser build to speed up the startup time, you might also want to look at : https://github.com/djdeath/noflo-iot
At some point I tried to run NoFlo on a board with very slow I/O. It turned out that a single file compacted build of NoFlo (including all the needed components) was significantly faster.
Related
I built an Electron app and I am now looking at how to distribute it.
I went with electron-builder to handle packaging etc.
For a bit of context, as a web developer, I am used to continuously deploy web apps on a web server but I have a hard time figuring out how to distribute a packaged one in Electron.
In electron-builder docs there is a brief mention about testing auto-update:
"Note that in order to develop/test UI/UX of updating without packaging the application you need to have a file named dev-app-update.yml in the root of your project, which matches your publish setting from electron-builder config (but in YAML format)"
But, it's rather vague...
So I actually have two questions:
1. How do I actually test the auto-update flow?
Do I need to actually publish a new version to trigger an update locally? Seems pretty unclear, it would be like developing against the production server.
2. Is it possible to have a fallback for unsigned code?
I don't have yet any certificate for code signing. So the OS/app will block the auto-update. But, I'd still want to tell the user that an update is available so they can go and download the app manually. Can I do that? (going back to point 1, I'd like to be able to test this flow)
I've just finished dealing with this. I also wanted to test against a non-production server and avoid having to package my app each time I iterated. To test downloads I had to sign my app, which slowed things down. But it sounds like you just need to check for updates. Which I think you can do as follows...
I created a dummy github repo, then created a a file dev-app-update.yml containing:
owner: <user or organization name>
repo: dev-auto-update-testing
provider: github
The path where this file is expected to be defaults to a place you can't access. Thankfully, you can override it like so:
if (isDev) {
// Useful for some dev/debugging tasks, but download can
// not be validated becuase dev app is not signed
autoUpdater.updateConfigPath = path.join(__dirname, 'dev-app-update.yml');
}
...that should be enough for your case -- since you don't need downloads.
If not, here are some other tips:
you can change the repo setting in your electron-builder config to point at your dummy repo then package your app. This will give you a packed, production build that points at your dummy repo -- this is how I did my download testing (though I have a cert, and signed my app)
you should be calling autoUpdate's checkForUpdates(), but if checkForUpdatesAndNotify() gives you a useful OS Notification then you should be able to set autoUpdater.autoDownload to false and end up with what you need.
Lastly, it sounds you could skip autoUpdater, since you won't be using the download feature anyway. Instead you could use github's releases api, assuming you use github to host your release. If not then your host should have something similar. Use that to check for updates then tell the user from within your App (could present them with a clickable URL too). If you want OS Notifications electron has a module for that.
We're using electron-updater with GitHub as a provider for auto-updates. Unfortunately, it breaks a lot and the electron-builder team doesn't support these issues well (1, 2, 3) (from my own experience, but you can find more examples on GitHub).
One way to test updates in dev mode:
Create a build of your app with an arbitrarily high version number
Create a public repo and publish the above build
Create a dev-app-update.yml next to your main entry point and configure it for the repo above (see)
In your main entry point:
import { autoUpdater } from "electron-updater";
...
if (process.env.NODE_ENV === "development") {
// Customize the test by toggling these lines
// autoUpdater.autoDownload = false
// autoUpdater.autoInstallOnAppQuit = false;
autoUpdater.checkForUpdates();
}
Then when running yarn dev you should see something like:
Checking for update
...
Found version 100.0.0 (url: <>.exe)
Downloading update from <>.exe
updaterCacheDirName is not specified in app-update.yml Was app build using at least electron-builder 20.34.0?
updater cache dir: C:\Users\<>\AppData\Local\Electron
New version 100.0.0 has been downloaded to C:\Users\<>\AppData\Local\Electron\pending\<>.exe
And it should install when you close the dev app.
This should give you some certainty but we still ran into issues in production. If you want to be sure, play through the full update flow with a test repo but packaged production apps just as you would do with the live one.
There is a javascript library, pre-built and available on npm, that I wish to develop with/debug. In my case, it is openlayers.
In the classic way of requiring a javascript file and wanting to debug, one would just switch the script url from the production version to the debug version, ie:
to
However, when using webpack and then importing via npm:
import openlayers from 'openlayers'
Gets you the production distribution of the library, the same as the ol.js script from above.
On a side note, to stop webpack trying to parse a prebuilt library and throw a warning about that you must include something like this:
// Squash OL whinging
webpackConfig.module.noParse = [
/\/dist\/ol.*\.js/, // openlayers is pre-built
]
Back to the problem at hand: how can I conditionally load a different entry-point for a module prebuilt and imported like this?
Of course, I can do it in a hacky way. By going into the node_modules/openlayers/package.json and switching the browser field from
"browser": "dist/ol.js",
to
"browser": "dist/ol-debug.js",
Is there a way I can request a different entry point via webpack or by using a different import syntax? Do I first have to petition the library maintainers to update the browser field to allow different entry point hints to browsers, according to the spec? https://github.com/defunctzombie/package-browser-field-spec
Thoughts on a more effective way to make this happen? Yearning to be able to programmatically switch loading of the production and debug versions of a library based on env variables.
Webpack has configuration options for replacing a module into a different path: https://webpack.github.io/docs/configuration.html#resolve-alias
This resolves the openlayers import to use the debug version:
webpackConfig.resolve.alias: {
openlayers: 'openlayers/dist/ol-debug.js'
}
In my build system I have a function that takes the environment type and returns the matching webpackConfig. Based on the parameter I include the above snippet or not.
Full code: webpack-multi-config.js
I have two different (gulp-) tasks for development and production. For example the production task: webpackProduction.js
Line 1 imports the config script with production as type.
My build system is based on gulp starter.
Trying to achieve this proposal to make a div size based on its background image I'm using Less 2.5.1 compiled by Web Essentials 2013 for Update 4 version 2.5.4 and some functions do not resolve:
My Less File:
.myClass{
height:image-height("myUrl.png"); //<-- Do not resolve
width:percentage(1/2); //<--Resolves
}
Resolved Css File:
.myClass{
height:image-height("myUrl.png"); //Not good
width:50%; //Good
}
I'm having this problem with image-height, image-width and image-size. Not with the rest of Misc Functions
There's no problem resolving the image by the compiler, because it's working fine when I used it like:
background-image:url("myUrl.png");
In the documentation there's a note for these methods saying:
Note: this function needs to be implemented by each environment. It is
currently only available in the node environment.
In the WebEssentials documentation it says:
Web Essentials uses the node-less compiler and it always uses the
latest version
And
NodeJS compilation - Web Essentials uses NodeJS to run the compiler.
It's the fastest and most accurate compiler for LESS available.
I'm missing something that I don't know. What does that note mean? What
should I do?
I've create a jsfiddle to share with you my goal.
I'm trying to write tests for my addon, but encountering some weird behaviour.
I have created a service (via ember-cli generate), which is used inside a component.
When an actual application is running everything is working fine.
However, when testing the component I get an error saying that the service is undefined when trying to access any of its properties/methods.
In the test i've put the service in "needs" like so:
needs: ['service:my-service']
"Needing" other components (e.g. child ones used inside) works as expected, services strangely fail.
Are there any additional steps that need to be done?
Running ember-cli 0.1.12.
When you generate a service, it also generates an initializer whose job it is to inject the service into the various places that you need it.
So, when you run acceptance tests your app will have booted and initializers will have been run, therefore the services will be available.
However, when unit testing components you get a clean container (better for testing). You just need to inject what you need:
moduleForComponent('foo-bar', null, {
setup: function(container) {
container.register('service:foo', FooService);
container.injection('component', 'fooService', 'service:foo');
}
});
I managed to get this working by using the new Ember.inject API available in the latest (as of writing) 1.10 release.
Apparently the new inject API is intended to replace needs in the future, it also works great with unit tests.
We just managed to get one working using needs: ['service:myService'] instead of needs: ['service:my-service'].
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.