| Thread Previous • Date Previous • Date Next • Thread Next |
On 2010-11-18 23:48, Jonathan Lange wrote:
AIUI, we need it for some tests, and we need it for deployment. Why not get the tests to generate the wadl just-in-time, and have a separate make target for things that need to be done on deployment?
Did you really just juxtapose the phrases "generate the wadl" and "just in time" in a proper sentence? My compliments. ;)
But that aside, much as I like the thought of not building the wadl during development, wouldn't it invite hiccups in deployment?
Also note that there's a 'make compile' step which builds practically everything you need for most tests, and does *not* build wadl.
That does sound like it justifies what you said earlier. Developers shouldn't have to predict such subtleties. Is the WADL secretly a test resource that we set up too aggressively?
Treating it as a test resource would solve both problems: avoid building where possible, yet exercise on every test run.
(On a loosely related note, without knowing much about it, I'm not even sure I like the type system that WADL seems to be a part of. The classes we get on the API feel more like rigid Modula/Pascal/Java than modern Smalltalk/Python/Go to me, if you get my drift.)
Jeroen
| Thread Previous • Date Previous • Date Next • Thread Next |
