mirror of
https://github.com/golang/go
synced 2024-11-19 03:24:40 -07:00
8bb20b8231
Support for:
(*reflect.rtype).Field
(*reflect.rtype).FieldByName
reflect.MakeSlice
runtime.SetFinalizer
Details:
- analysis locates ssa.Functions for (reflect.Value).Call
and runtime.SetFinalizer during startup to that it can
special-case them during genCall. ('Call' is forthcoming.)
- The callsite.targets mechanism is only used for dynamic
calls now. For static calls we call callEdge during constraint
generation; this is a minor optimisation.
- Static calls to SetFinalizer are inlined so that the call
appears to go direct to the finalizer. (We'll use the same
trick for (reflect.Value).Call.)
- runtime.FuncForPC: treat as a no-op.
- Fixed pointer_test to properly deal with expectations
that are multi-sets.
- Inlined rtypeMethodByNameConstraint.addMethod.
- More tests.
R=crawshaw
CC=golang-dev
https://golang.org/cl/14682045
82 lines
1.8 KiB
Go
82 lines
1.8 KiB
Go
package main
|
|
|
|
import "runtime"
|
|
|
|
func final1a(x *int) int {
|
|
print(x) // @pointsto new@newint:10
|
|
return *x
|
|
}
|
|
|
|
func final1b(x *bool) {
|
|
print(x) // @pointsto
|
|
}
|
|
|
|
func runtimeSetFinalizer1() {
|
|
x := new(int) // @line newint
|
|
runtime.SetFinalizer(x, final1a) // ok: final1a's result is ignored
|
|
runtime.SetFinalizer(x, final1b) // param type mismatch: no effect
|
|
}
|
|
|
|
// @calls main.runtimeSetFinalizer1 -> main.final1a
|
|
// @calls main.runtimeSetFinalizer1 -> main.final1b
|
|
|
|
func final2a(x *bool) {
|
|
print(x) // @pointsto new@newbool1:10 | new@newbool2:10
|
|
}
|
|
|
|
func final2b(x *bool) {
|
|
print(x) // @pointsto new@newbool1:10 | new@newbool2:10
|
|
}
|
|
|
|
func runtimeSetFinalizer2() {
|
|
x := new(bool) // @line newbool1
|
|
f := final2a
|
|
if unknown {
|
|
x = new(bool) // @line newbool2
|
|
f = final2b
|
|
}
|
|
runtime.SetFinalizer(x, f)
|
|
}
|
|
|
|
// @calls main.runtimeSetFinalizer2 -> main.final2a
|
|
// @calls main.runtimeSetFinalizer2 -> main.final2b
|
|
|
|
type T int
|
|
|
|
func (t *T) finalize() {
|
|
print(t) // @pointsto new@final3:10
|
|
}
|
|
|
|
func runtimeSetFinalizer3() {
|
|
x := new(T) // @line final3
|
|
runtime.SetFinalizer(x, (*T).finalize)
|
|
}
|
|
|
|
// @calls main.runtimeSetFinalizer3 -> (*main.T).finalize
|
|
|
|
// I hope I never live to see this code in the wild.
|
|
var setFinalizer = runtime.SetFinalizer
|
|
|
|
func final4(x *int) {
|
|
print(x) // @pointsto new@finalIndirect:10
|
|
}
|
|
|
|
func runtimeSetFinalizerIndirect() {
|
|
// In an indirect call, the shared contour for SetFinalizer is
|
|
// used, i.e. the call is not inlined and appears in the call graph.
|
|
x := new(int) // @line finalIndirect
|
|
setFinalizer(x, final4)
|
|
}
|
|
|
|
// @calls main.runtimeSetFinalizerIndirect -> runtime.SetFinalizer
|
|
// @calls runtime.SetFinalizer -> main.final4
|
|
|
|
func main() {
|
|
runtimeSetFinalizer1()
|
|
runtimeSetFinalizer2()
|
|
runtimeSetFinalizer3()
|
|
runtimeSetFinalizerIndirect()
|
|
}
|
|
|
|
var unknown bool // defeat dead-code elimination
|