On Fri, 2011-02-25 at 11:08 -0800, Peter Niederwieser wrote:
> Real-life examples are important, but we can't possibly have an example for
> every method and property in Gradle's APIs (and bootstrapClasspath is a
> rarely needed one). Hence we also need to empower users to find the answers
> themselves, provided that they have studied the central APIs and principles
> behind Gradle (which is their responsibility).

I agree bootstrapClasspath is "advanced stuff", so less pressure for
examples in the API documentation.  But the Gradle API documentation
does need to have more explanatory material that it does now, at least
before 1.0 is released.

The principles, and indeed the overall architecture and computational
model, probably need a little more documentation that there is
currently.  Or I have missed stuff in skimming the guide -- which is
entirely possible.

> One such principle (idiom) is that collections can be overwritten with "coll
> = [1,2,3]", and added to with "coll 1,2,3". I had specifically pointed you
> to the latter variant of bootstrapClasspath in the DSL guide (was there a
> problem with the anchor?). From the documentation: "Adds the given values to
> the end of the bootstrap classpath for the process.". 

Your link didn't work, it just put Epiphany into a hang.  I had tried
coll << [ 1 , 2 , 3 ] but that was problematic.  I also tried coll.add
( 1 ) but that gave me an error that coll could not be updated, but then
coll isn't user defined it is system defined so that might be
reasonable.  I didn't think of just coll ( 1 , 2 , 3) 

The problem is I need to prepend not append.

> Hope this helps.

Progressing, but not there yet.

I think this JSR 166 build is going to see if Gradle really can do stuff
that is trivial, albeit extremely verbose, in Ant.

The core issue is that the jsr166.jar needs to be first in the bootstrap
class path for compiling and running all tests for all the other
projects.  I am trying to decide if it is necessary for the compilation
of jsr166x, jsr166y, and extra166y as well, I think it is.  I think I
have got away so far simply because the API is still teh same as my
OpenJDK 6 one.  i.e. pure fluke!

-- 
Russel.
=============================================================================
Dr Russel Winder      t: +44 20 7585 2200   voip: sip:[email protected]
41 Buckmaster Road    m: +44 7770 465 077   xmpp: [email protected]
London SW11 1EN, UK   w: www.russel.org.uk  skype: russel_winder

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to