Re: [CommonJS] Recount on Unit Testing API (was: Vote: use QUnit API or Unit Testing/A)

43 views
Skip to first unread message

Kris Kowal

unread,
Oct 9, 2009, 4:24:51 PM10/9/09
to comm...@googlegroups.com
Proposal: we specify in CommonJS that layer 1 be the modules "test"
and "assert" as in #2. We converge with QUnit on standard names for
assertions. Then, we proceed to standardize loaders so that we can do
module dependency injection. From there, we can build a CommonJS
system that uses module dependency injection to load QUnit DSL modules
that call into the layer 1 interface. We'll probably also eventually
standardize an API layer for connecting the test.run method to call
into a logging layer, e.g., back into QUnit.

With this proposal, we only need to agree on names for the assertion
module's methods. If this proposal is reasonable, please contribute
nominations for complete API's in the following form, and then we can
do run offs.

A (ok, equiv, notEquiv, equal, notEqual)
B (ok, eq, ne, is, isnt)
C (ok, eq, uneq, is, isnt)

Kris Kowal

Dean Landolt

unread,
Oct 9, 2009, 4:45:41 PM10/9/09
to comm...@googlegroups.com

 A, B, C -- in that order.

Until, of course, someone proposes something else. This is gonna be hard to keep track of :/

Kris Kowal

unread,
Oct 9, 2009, 4:50:44 PM10/9/09
to comm...@googlegroups.com
On Fri, Oct 9, 2009 at 1:45 PM, Dean Landolt <de...@deanlandolt.com> wrote:
> Until, of course, someone proposes something else. This is gonna be hard to
> keep track of :/

Yes, please hold off showing your hand until all nominations are in.

Dean Landolt

unread,
Oct 9, 2009, 4:51:06 PM10/9/09
to comm...@googlegroups.com

D (ok, equal notEqual, deepEqual, notDeepEqual), A, B, C

Kris Kowal

unread,
Oct 9, 2009, 4:52:46 PM10/9/09
to comm...@googlegroups.com
New nomination:

In positions:
1 is the plain assertion method
2 is the deep equivalence comparison method
3 is the opposite thereof
4 is the shallow equivalence comparison (==)
5 is the opposite thereof
6 is the shallow strict equivalence comparison (===)
7 is the opposite thereof

A (ok, equiv, notEquiv, equal, notEqual)
B (ok, eq, ne, is, isnt)
C (ok, eq, uneq, is, isnt)

D (ok, deepEqual, notDeepEqual, equal, notEqual, same, notSame)

Kris Kowal

John Resig

unread,
Oct 9, 2009, 4:53:04 PM10/9/09
to comm...@googlegroups.com
> D (ok, equal notEqual, deepEqual, notDeepEqual), A, B, C

Dean, Ash, Hannes, Nathan, Kris, and I discussed this on IRC and
decided that the following was acceptable:


D (ok, equal notEqual, deepEqual, notDeepEqual)

equal/notEqual uses either == or === (to be decided)
deepEqual/notDeepEqual does a recursive comparison on the object properties.

--John

Kris Kowal

unread,
Oct 9, 2009, 4:55:40 PM10/9/09
to comm...@googlegroups.com
Scaling back scope:

In positions:
1 is the plain assertion method
2 is the deep equivalence comparison method
3 is the opposite thereof

4 is the shallow equivalence comparison (either == or == to be decided later)


5 is the opposite thereof

A (ok, equiv, notEquiv, equal, notEqual)


B (ok, eq, ne, is, isnt)
C (ok, eq, uneq, is, isnt)

D (ok, deepEqual, notDeepEqual, equal, notEqual)

Kris Kowal

Kevin Dangoor

unread,
Oct 12, 2009, 11:20:00 AM10/12/09
to comm...@googlegroups.com
I'm fine with the idea you're proposing Kris.

Go with D, which is the one John said that several folks were okay with already, stick mostly with the existing Unit_Testing/A proposal and eventually converge on a system that can readily handle both QUnit style tests and Unit Testing/A tests.

I would still like to get something into the proposal allowing failed assertions to log rather than throw an exception...

Kevin

--
Kevin Dangoor

work: http://labs.mozilla.com/
email: k...@blazingthings.com
blog: http://www.BlueSkyOnMars.com

Kris Kowal

unread,
Oct 12, 2009, 2:28:19 PM10/12/09
to comm...@googlegroups.com
On Mon, Oct 12, 2009 at 8:20 AM, Kevin Dangoor <dan...@gmail.com> wrote:
>> D (ok, deepEqual, notDeepEqual, equal, notEqual)
> Go with D, which is the one John said that several folks were okay with
> already, stick mostly with the existing Unit_Testing/A proposal and
> eventually converge on a system that can readily handle both QUnit style
> tests and Unit Testing/A tests.

Ok.

> I would still like to get something into the proposal allowing failed
> assertions to log rather than throw an exception...

I'll write some verbiage about that. Probably something like
exports.assertions = "log"; or by default "throw".

Kris Kowal

George Moschovitis

unread,
Oct 13, 2009, 6:39:51 AM10/13/09
to CommonJS
> A (ok, equiv, notEquiv, equal, notEqual)
> B (ok, eq, ne, is, isnt)
> C (ok, eq, uneq, is, isnt)
> D (ok, deepEqual, notDeepEqual, equal, notEqual)


D, A

-g.
Reply all
Reply to author
Forward
0 new messages