Placeholder bug for implementing interruptible reflow.

Here are some results. On a 750MHz Win2K machine running iBench and a 2001-06-01 trunk build, Reflow Time count=1976, minimum=0, maximum=170, mean=17.15+/-35.43 0: 1388 10: 174 20: 62 30: 20 40: 12 50: 6 60: 32 70: 32 80: 74 90: 60 100: 31 110: 19 120: 15 130: 21 140: 21 150: 7 160: 1 170: 1 (That's 1388 reflows that took l.t. 10ms, 174 reflows that took 10 to 19ms, etc.)

Here are the results from the same build on a 200MHz machine running WinNT: Reflow Time count=1953, minimum=0, maximum=1240, mean=82.81+/-168.45 0: 683 10: 669 20: 139 30: 20 40: 9 50: 2 60: 7 70: 3 80: 1 90: 3 100: 13 110: 24 120: 9 130: 17 140: 9 150: 11 160: 4 170: 7 180: 2 190: 2 200: 2 210: 1 220: 1 230: 7 240: 9 250: 8 260: 4 270: 16 280: 3 290: 11 300: 11 310: 8 320: 15 330: 18 340: 9 350: 16 360: 7 370: 5 380: 3 390: 1 400: 4 410: 3 420: 7 430: 4 440: 9 450: 9 460: 3 470: 9 480: 8 490: 7 500: 6 520: 2 530: 9 540: 12 550: 6 560: 6 570: 5 580: 4 590: 6 600: 3 610: 3 620: 2 630: 2 640: 7 650: 2 660: 5 670: 1 680: 1 700: 2 710: 3 720: 3 730: 3 740: 4 750: 1 770: 1 1230: 1 1240: 1

We've got a placeholder for this already... consolidating dependency lists and dupping. *** This bug has been marked as a duplicate of 67752 ***