It's free at the moment, so feel free to try it. We will announce pricing and licensing in advance, so there'll be enough time to decide whether or not to keep using the tool.
Sorry, but I've been there done that.
"Here, just try it for free", they say. "What harm can it do?", they say. "Licensing costs will be basically nothing", they say. "Trust us", they say.
So I try, and I like and I get used to it and then they start with "Oh you wanted to use that one really useful feature, That will be $5000 extra. Oh, you're running on Solaris, that will be $10000 extra. Oh you have more than 2 CPUs, that will be $20000 extra. Oh you want to use it Thursdays as well, that will be $30000 extra".
So now I never ever try software until I've seen a complete price list.
He said " less expensive than some similar" software, not all similar software. So for any given price point all they need is someone somewhere selling something similar for more.
I also missed that at first, and was disappointed. Hopefully the pricing will be reasonable enough because I'm very excited about this - just waiting on VS integration
We will most likely have a subscription model, so the tool will be very affordable to start using without any significant capital investments. We didn't put the pricing section on top because there's no pricing yet. The section is right next to the download one, so hopefully those who download - see it.
There's very little done for code analyze in JS, I can't even find an editor that gives me function parameter hints And show me a list of functions.
I have even considered making my own editor.
Coding without such tools force me to write code that is easy to manage and review though.
Both Visual Studio (from Microsoft) and WebStorm (from JetBrains) give you JS intellisense. Granted, JS intellesense is never going to give you as much information as a statically typed language, but Typescript helps the editor as does comment type-hinting in VS (using something along the lines of JSDoc)
You can annotate types within JSDoc, which can help somewhat. But granted, compared to statically typed languages it's still a clusterfuck. Currently writing documentation comment translation for a C#→Java/JS compiler and I'm injecting as much type-system metadata as I can, but IDE and editors still disagree to what extent they're actually using that.
I've used NCrunch when writing C# code and it makes life tremendously better. I've been longing for a JavaScript equivalent ever since, looks like it might finally be here! Hooray!!
I write a lot of tests and I've tried this a lot of times (since it can be done with any test framework plus a file watcher on any language) but it doesn't help on my workflow.
The integration with the code editor looks nice as long as it has a way to "jump to the first broken test", and show all the details of the error. I remember when I used to work on c# the test runner of resharper integrated and it was great. My current workflow is, I run the tests in the console and if some test is broken, I search the name of the test across the project to inspect.
Parallel test execution in javascript hasn't worked for me, most of the time I depend on some "state" call it DOM (browser), or call it a test database (node). The fact that I can assume that the runner run one test at the time makes my tests and the before/after code simpler.
Wallaby.js is very different from a file watcher. File watcher would run all of your JavaScript tests and it can be very time consuming. Wallaby uses sophisticated dependency analysis and lots of other tricks to only run tests affected by your code changes, no more no less, so even in large codebases with lots of tests the feedback is delivered very fast.
Yep, wallaby can jump to failing test(s), error sources, show error details inline and much more.
Parallel execution in wallaby works fine with browser "state". Imagine running a half of your tests in one browser tab and another one in a second tab - that's what wallaby does but in 6 or 8 or N processes. As long as your tests set up context for themselves and don't depend on each other (and they shouldn't) - it all works.
Okay this looks pretty interesting. I'll give it a go. Right now my primary editor is Sublime 3 (occasionally atom.io but it's kinda slow) so support for at least Sublime would be awesome. I've been meaning to try IntelliJ so maybe this will give me that excuse.
This looks well-done. One thing that wasn't clear to me was whether Wallaby supports simultaneous cross-browser testing a la Karma. Ideally, I'd want it to just work with Karma. Wallaby folks in this thread, can you clarify this for me?
So how does Wallaby run your front-end tests, if not in a real browser?
Even if it does run in a real browser, cross-browser issues are common enough in my work that my TDD loop always includes multiple browsers. I wouldn't be willing to give that up for what Wallaby offers.
Part of the reason I'm interested in something like Wallaby is that running in multiple browsers is slow--but if I could run just the right subset, it would be fast enough for my needs. (And Karma still runs about 200 tests / second on my mid-2002 MacBook Pro, across 10 browsers, including IE VMs, so it's not like Karma is that slow.)
Wallaby.js supports PhantomJs. It also supports node.js/io.js for node tests, as well as for front-end tests with something like jsdom.
Karma is not slow, but some of our tests may be. Wallaby just allows not to run them if we are not editing anything that affects them.
Cross browser issues are always a pain, I know what you mean. However, I have realised that I can write the bulk of my code without worrying too much about them.
When I'm writing the bulk of logic, especially with TDD, I found that it's more efficient to run tests as fast as possible in the fastest engine and care about running across all browsers much less frequently. Especially if you run tests as you type.
Later on, closer to the end of the iteration, when preparing to commit my changes, I surely run all of the tests in all browsers/smartphones/tablets that I support. Typically it just works, especially when your codebase is mature, you already know the majority of cross-browser issue sources and have something in place to avoid them. Even if it doesn't work, the issues are normally easy to solve. Overall, less context switching, more focussed work.
Anyway, it's just my experience (you can read more about the motivation behind wallaby.js in my blog post - http://dm.gl/2015/01/30/wallaby/) and we are planning to support more browser runners, or even just karma as a runner for wallaby, in future. Please feel free to vote (https://github.com/wallabyjs/public/issues/29) or create new feature requests.
That's a reasonable perspective. For me, my thorniest issues have come from browser incompatibilities, so it's more important to catch and fix issues at the moment they're introduced rather than later. This isn't necessarily testing all browsers; just the ones that I know are likely to trigger differences. For example, I might run IE 8 + IE 11 + Firefox + Chrome + Mobile Safari.
Fast tests (< 1-5s) have a lot of value to me (although "test when I press save" is as good or even better than "test as I type," as far as I'm concerned), so Wallaby is something I'm interested in. I'm more likely to do manual test sub-setting than give up my cross-browser tests, though.
One potential customer's perspective; you can take it or leave it. :-) I'm not necessarily a good example of the mainstream. I've voted on the issue.
Running affected tests as fast as possible and running all of them in a variety of different browsers to make sure your code works everywhere - are two different tasks. Not like they are completely mutually exclusive, but it's hard to run something very fast, as you type your code, if you have to do it in multiple versions of IE/Chrome/FF and perhaps a smartphone/tablet at the same time. Easier to just use different tools at different stages - wallaby when you write code and tests, karma when you need to run them all in many browsers.
Can I ask what is the point of having tests that take that long to run? Being notified 3 weeks after I made a change that I broke something is pretty useless to me...
And what language/framework or other reason could possibly contribute to that long a test suite?
We will at some point. Right now working on module loaders support (browserify, webpack), then will be adding more testing frameworks support. TypeScript/CoffeeScript next.
Feel free to add feature requests/vote - https://github.com/wallabyjs/public/issues
Once more information is available I'll probably be interested.