https://github.com/jamescarr/jasmine-tool
I'm thinking of asking him to rename this to jasmine-tool in npm. Are
people here ok with me doing that?
--
You received this message because you are subscribed to the Google Groups "Jasmine Dev" group.
To post to this group, send email to jasmine...@googlegroups.com.
To unsubscribe from this group, send email to jasmine-js-de...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/jasmine-js-dev?hl=en.
Node, at least at it relates to jasmine, is mostly just a way of
saying "CommonJS compatibility".
> This is the point of Misko's jasmine-node project.
Hrm, Misko just made a jasmine runner for node. That's different from
making this work:
npm install jasmine
...
var jasmine = require("jasmine")
Misko couldn't make that happen unless he forked jasmine and submitted
pull requests back to pivotal/jasmine.
> I'm more inclined to think about making changes that make Jasmine core work
> more easily in Node are great, but are lower priority in the immediate term.
> If standalone usage has to change in order to setup & exec the environment
> then this is great for 2.0.
Davis I don't want to go upsetting some consensus here. Or maybe I
do. I find the talk of "2.0" to be odd. Why can't we just have some
broad ongoing shepherding of pull requests - merging pull requests,
asking for changes then merging, rejecting pull requests - and have
jasmine evolve the way any other github project does?
What I'm driving towards is getting jasmine in shape such that one can
properly declare a dependency into it from a node project. As of
today jasmine must be copied around everywhere:
https://github.com/mhevery/jasmine-node/blob/master/lib/jasmine/jasmine-1.0.1.js
or included via sibling directories:
https://github.com/sconover/knit-js/blob/master/test/helper.js
vs the ideal, which is a simple Gemfile-like dependency declaration:
https://github.com/assaf/zombie/blob/master/package.json
I fully intend to replace all browser testing I do with node/jsdom,
and I'd like to use jasmine as my testing framework, and am attempting
to do what's typically done in git-land to get it there. It won't
take all that much to get jasmine in compliance with CommonJS style
(an initial take is sitting in a pull request).
But if this is moving too fast, or if the aims of pivotal/jasmine are
to stay browser-centric, I have no problem letting this live on my
fork. I could talk to the jasmine-tool guy and make sconover/jasmine
the npm installed version.
https://github.com/pivotal/jasmine/pull/63
https://github.com/pivotal/jasmine/pull/46
Then after a while you might think "well maybe jasmine core should be
able to run its own tests in node out of the box". "You need
something like TrivialReporter, except for the console."
...
The logical conclusion is that jasmine grows its own jasmine-node
capability. I don't think we should think of that as a bad thing (or,
even if you think it's bad, it's at least a forced move). Imagine
jasmine as it stands with the browser-based suite sitting in another
project...it's weird to even consider (or, it is to me).
- browser run
- node run
- jshint check
For a project like jasmine, at HEAD, why is anyone worrying about a CI
green off in the world somewhere? I'm looking through the code and I
see nothing that requires a server process or anything complex...
...pull request: