Sink regression on devel: cannot compile the generated C code anymore

[mratsim]

In the last 3 days there was a regression in devel on the codegen for sink proc:

I don't have a full self-contained reproducible example at the moment but the colontmpD_ and blitTmp variable are not declared before use.

proc nextTask*(childTask: bool): Task {.inline.} =

  profile(enq_deq_task):
    if childTask:
      result = myWorker().deque.popFirstIfChild(myTask())
    else:
      result = myWorker().deque.popFirst()

  # TODO: steal early

  shareWork()

  # Check if someone requested to steal from us
  var req: StealRequest
  while recv(req):
    # If we just popped a loop task, we may split it here
    # It makes dispatching tasks simpler
    if myWorker().deque.isEmpty() and result.isSplittable():
      if req.thiefID != myID():
        splitAndSend(result, req)
      else:
        forget(req)
    else:
      dispatchTasks(req)

C code (with focus on related paragraph)

static N_INLINE(tyObject_TaskcolonObjectType___1Z1XSKW5drk9cOAphS9cyhbg*, nextTask__UE7GE7k3y9crMoWUP5T9boBQscheduler)(NIM_BOOL childTask) {	

       ///////////////////////////////////////
       ...
       ///////////////////////////////////////

	req = (tyObject_StealRequestcolonObjectType___LTgtnfBhL5TdUm2ePZ4M6Q*)0;
	{
		nimln_(68, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
		while (1) {
			NIM_BOOL T8_;
			T8_ = (NIM_BOOL)0;
			T8_ = recv__rh6W1iOJIA0BqMW3EaQ72gvictims(&req);
			if (!T8_) goto LA7;
			nimln_(71, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
			{
				NIM_BOOL T11_;
				T11_ = (NIM_BOOL)0;
				T11_ = isEmpty__zTHBqKFOE8nGHllwGgDJrgprell_deques(localCtx__xr8jkofR2rSp69czCGO79bqQ.worker.deque);
				if (!(T11_)) goto LA12_;
				T11_ = isSplittable__N8IdLdi5T7pGxk1tzH003g(result);
				LA12_: ;
				if (!T11_) goto LA13_;
				nimln_(72, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
				{
					tyObject_StealRequestcolonObjectType___LTgtnfBhL5TdUm2ePZ4M6Q* blitTmp;
					if (!!(((*req).thiefID == localCtx__xr8jkofR2rSp69czCGO79bqQ.worker.ID))) goto LA17_;
					nimln_(73, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
					colontmpD_ = 0;
					colontmpD_ = req;
					blitTmp = colontmpD_;
					colontmpD_ = 0;
					splitAndSend__YufgLt6fiFyPy0arOhE8tA(result, blitTmp);
				}
				goto LA15_;
				LA17_: ;
				{
					tyObject_StealRequestcolonObjectType___LTgtnfBhL5TdUm2ePZ4M6Q* blitTmp_2;
					nimln_(75, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
					colontmpD__2 = 0;
					colontmpD__2 = req;
					blitTmp_2 = colontmpD__2;
					colontmpD__2 = 0;
					forget__FoJB8hpGPM8HP8Am6rgcTQ_2(blitTmp_2);
				}
				LA15_: ;
			}
			goto LA9_;
			LA13_: ;
			{
				tyObject_StealRequestcolonObjectType___LTgtnfBhL5TdUm2ePZ4M6Q* blitTmp_3;
				nimln_(77, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
				colontmpD__3 = 0;
				colontmpD__3 = req;
				blitTmp_3 = colontmpD__3;
				colontmpD__3 = 0;
				dispatchTasks__bgTUscn2shIVJXNChVXG9bg(blitTmp_3);
			}
			LA9_: ;
		} LA7: ;
	}
	popFrame();
	return result;
}

Commit 1b2c1bc from Thursday was working fine and generated the following instead

static N_INLINE(tyObject_TaskcolonObjectType___1Z1XSKW5drk9cOAphS9cyhbg*, nextTask__UE7GE7k3y9crMoWUP5T9boBQscheduler)(NIM_BOOL childTask) {	

       ///////////////////////////////////////
       ...
       ///////////////////////////////////////

	req = (tyObject_StealRequestcolonObjectType___LTgtnfBhL5TdUm2ePZ4M6Q*)0;
	{
		nimln_(68, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
		while (1) {
			NIM_BOOL T8_;
			T8_ = (NIM_BOOL)0;
			T8_ = recv__rh6W1iOJIA0BqMW3EaQ72gvictims(&req);
			if (!T8_) goto LA7;
			nimln_(71, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
			{
				NIM_BOOL T11_;
				T11_ = (NIM_BOOL)0;
				T11_ = isEmpty__zTHBqKFOE8nGHllwGgDJrgprell_deques(localCtx__xr8jkofR2rSp69czCGO79bqQ.worker.deque);
				if (!(T11_)) goto LA12_;
				T11_ = isSplittable__N8IdLdi5T7pGxk1tzH003g(result);
				LA12_: ;
				if (!T11_) goto LA13_;
				nimln_(72, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
				{
					if (!!(((*req).thiefID == localCtx__xr8jkofR2rSp69czCGO79bqQ.worker.ID))) goto LA17_;
					nimln_(73, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
					splitAndSend__YufgLt6fiFyPy0arOhE8tA(result, req);
				}
				goto LA15_;
				LA17_: ;
				{
					nimln_(75, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
					forget__FoJB8hpGPM8HP8Am6rgcTQ_2(req);
				}
				LA15_: ;
			}
			goto LA9_;
			LA13_: ;
			{
				nimln_(77, "/home/beta/Programming/Nim/weave/weave/scheduler.nim");
				dispatchTasks__bgTUscn2shIVJXNChVXG9bg(req);
			}
			LA9_: ;
		} LA7: ;
	}
	popFrame();
	return result;
}

Source: https://github.com/mratsim/weave/blob/e3a1f2944a1bf3bdd25155a4c220b7aabdf38a93/weave/scheduler.nim#L54-L77

Repro:

git clone https://github.com/mratsim/weave
cd weave
mkdir build
# in weave/async.nim at the bottom comment out main2() and comment in main()
# unless your PC is fast you want to heat your home by computing fib(40) the naive way in debug mode
nim c --threads:on -o:build/async weave/async.nim

[mratsim]

After bisecting, it seems like the culprit is this "bugfix" that now injects destructors ;) c85e266

[Araq]

Some snippet would really be nice in order to tackle this...

[mratsim]

Dropping it to lower priority, seemingly unrelated changes (no change in type, proc or even file) removed the issue in later commits in my repo.

It seems like tough bug to extract.

[mratsim]

Apparently removing inline from the proc that have issues worked so maybe related to #12770