@dr2chase

Note that uintptr is treated specially because reasons, but this "obvious" change to escassign in esc.go

func (e *EscState) escassign(dst, src *Node, step *EscStep) {
	if dst.isBlank() || dst == nil || src == nil || src.Op == ONONAME || src.Op == OXXX {
		return
	}
	if !types.Haspointers(src.Type) && (src.Type.Etype != types.TUINTPTR) {
		if Debug['m'] > 2 {
			Warnl(src.Pos, "Esc-assigned type has no pointers")
		}
		return
	}

produced a build that (1) failed some tests in horrible crashy ways and (2) passed all the escape analysis tests. This was supposed to be easy, and instead it looks interesting.

One of the places where this wrong-fix causes a difference in allocation behavior is test/gogcrt.go, line 174

				a[new(bool)] = false

Putting this off (unless someone else is interested) till after the freeze.

Hello @dr2chase @randall77, shall we punt this to Go1.13? If for Go1.13, @dr2chase let's team up and work on this issue during the cycle.

Definitely punt to 1.13, this is not a crash, and I don't think this is a regression, either.
@mdempsky is working on (mostly done with) a rewrite of the escape analysis code, his new code might handle this more nicely (or at least more understandably).

I don't much like the the bit about my change passing escape analysis tests, yet producing a sometimes-crashing build.

My escape analysis rewrite keep n allocated on the stack.

@mdempsky I'm on version 13b7c04, and it seems fixed. Should I close the issue, or is there something else to do here, like adding test cases?

@ainar-g A regress test case CL would be great!

Change https://golang.org/cl/172358 mentions this issue: test: add regress test for issue 28369

This is my first change to test/, so I'm not entirely sure I've done it correctly.