Am 14.08.2018 um 19:10 schrieb ocs@ocs:
Jochen,
On 14 Aug 2018, at 6:34 PM, Jochen Theodorou <blackd...@gmx.org
<mailto:blackd...@gmx.org>> wrote:
[...]
are you saying x?.foo will NPE if x is null? Or that x?.getFoo() will
NPE in that case? Not sure how to read your comment.
Provided only (a) “the Null.metaclass; returning null from invokeMethod”
is used and no ASTT, then yes, it would NPE. Easiest thing on earth to try:
===
262 */tmp>* <q.groovy
class q {
static main(args) {
// ExpandoMetaClass.enableGlobally() // I thought this is
needed; seems not (though my application use it anyway)
def mc=new OCSNMC(org.codehaus.groovy.runtime.NullObject)
mc.initialize()
org.codehaus.groovy.runtime.NullObject.metaClass=mc
println "null.foo() is OK: ${null.foo()==null}"
println "null.foo we won't see: ${null.foo==null}"
}
}
class OCSNMC extends DelegatingMetaClass {
OCSNMC(Class clazz){
super(clazz)
}
Object invokeMethod(Object object, String methodName, Object[]
arguments) {
null
}
}
null.foo will of course NPE, if you only take care of invokeMethod. have
to do the same treatment for get/setProperty, potentially also for
get/setAttribute
[...]
For all I know, this probably would not work properly
with @CompileStatic (which I do not use at all myself, but others do
frequently).
the result type could be a problem... Worth to check.
Definitely :)
judging by this small snippet:
@groovy.transform.CompileStatic
def m() {
def foo = "s"
def bar = foo?.toUpperCase()
println bar.toUpperCase()
}
and that it compiles (and runs fine) I would say the correct type is
preserved.
Trust me, been there, done that. I am pretty darn sure it would be
/infinitely/ easier and, what's important, more reliable in the core
with an explicit compiler support.
How about making a small github project to dump the current state there?
Not that easy, for my code is mixed up with other ASTT and runtime
stuff; but I'll try to make some simple proof-of-concept ASAP and send here.
would help a lot
bye Jochen
