cburgmer

Author of another CSS regression testing tool here which doesn't depend on your full web app being up and running (which is the default for WebDriver and friends). I'd be seriously interested in feedback from people who have run a CSS regression test suite against a deployed web app, and especially how maintainable their testing setup was a year in and how much defects it found in that process. My personal experience with WebDriver for acceptance tests have made me approach such things very carefully. I am very conservative in how much I want to leverage a full stack setup.

"github-backup [...] backs up everything GitHub publishes about the repository, including branches, tags, other forks, issues, comments, wikis, milestones, pull requests, watchers, and stars."

https://github.com/joeyh/github-backup

Great to see more data-driven approaches for optimising builds.

Erik has written a bit about his approaches in the past [1] and Graham [2] and I [3] are both currently trying different ways of visualising pains the teams have we are working in.

[1] http://erik.doernenburg.com/2009/11/making-build-pain-visibl... [2] https://github.com/kuona/kuona [3] https://github.com/cburgmer/buildviz

https://www.npmjs.com/package/rasterizehtml renders HTML into the browser's canvas.

<rant>Yet another grid implementation that requires jQuery.</rant>

My last search on responsive grids (June 2015) has not turned up anything without jQuery. That would be a cool new "feature", that none of the existing solutions so far offer.

Alternatively for RSpec and friends there's https://github.com/grosser/parallel_tests. Works without a centralised server.

PhantomJS, SlimerJS or Selenium?

Selenium if you want to support all browsers, including mobile.

PhantomJS if you want a performant and easy solution.

SlimerJS if you want PhantomJS but want HTML5/CSS3, too (or just need Firefox).

My 2 cents

One quick tip if you're torn between SlimeJS and PhantomJS, do either of them or both at the same time, they expose the same* API.

* with minor changes, as documented with SlimerJS

A few years back I started to work on a Python library around Chinese called cjklib (https://pypi.python.org/pypi/cjklib). The intention was to facilitate the creation of Open Source tools for language learners. I even built my own application on that, sadly only for KDE (https://code.google.com/p/eclectus/wiki/About).

To be fair I've done a few things wrong (up-front design, no feedback process). However, I still found the general response less then I had hoped for.

I hope there are still a bunch of hackers around that like to solve similar problems. One is certainly https://github.com/nieldlr/Hanzi and that would be something I hacked on if I was coming back to the world.

I disagree with what I think this picture implies. That is shielding the developer from interaction makes him/her more productive. In contrary, interruption and conversation belong to development. The more my fellow developer colleagues talk, the better the code base is.

It looks like it's fixed to QUnit for now (configuration allows for a test.html file). No Jasmine then.

There is configuration for scripts, pre/post build and test phases. Unsure how this works together with having a set of browsers in the background.

In theory it makes sense to extend command line testing (maybe against PhantomJS) with integration tests against actual browsers. Let's see how this works out with BrowserSwarm.

I am using impress.js for all my presentations: http://bartaz.github.io/impress.js/

aemkei delivered an awesome presentation yesterday at JSConf EU.

This was his final slide.

Here is a project that gathers all available solutions around CSS testing: http://csste.st

Starts with a good set of things everybody might need, and then quickly trails off into an opinionated set of rather random things

Why does Object.prototype keep making it on those lists? Hands up please if you have used inheritance in JS (or any language) lately.

Sourcing the files from other servers solved that for me. I am still thinking in terms of a locked down jsfiddle I guess.

Copy pasting huge amount of JS libraries does not seem to work for me. Is there a way of uploading files from my local hard drive?

I am currently trying to set up a demo for a project, and a simple 'plug and play' would be great.

All the examples in the README point to the use around using this and new. I try following the Crockford school and thus hardly use any of them. So I ask myself whether this plugin is adding any benefit to me.

I probably cba to add [String] annotations.

Nice work. I'd like to see more projects around public data. Especially as mashups could potentially add much benefit on a otherwise narrow source.

Also kudos to TfL for providing that data. Back in Germany it's very hard to come by such sources.

I started using it last week to calculate a fingerprint of a set of files (http://cburgmer.github.com/csscritic/)

Seriously, PHP? Integrating that into your favorite editor?

Anyways. A tool to help you have clean CSS sounds cool.