Further test shows that :
Sun Jan 26 14:24:07 2014,MF Test 2.12 Log
a2200,100d,CHDK,1.3.0-3337
Mode:P,Continuous_AF:0,Servo_AF:0
tests : set_focus:0, set_aflock:1, levents:0, event_procs:1, bypass:1
test,set,actual,result,prop.FOCUS_STATE,prop.AF_LOCK,prop.FOCUS_MODE,prop.REAL_FOCUS_MODE
1) Testing initial conditions
...interlock bypass function available
...start, , , ,0,0,0,0
...shot1,1020,595,failed,16,0,0,0
2) Testing set_focus() only
..disabled
3) Testing set_aflock()
...start, , , ,16,0,0,0
...set_aflock(1), , , ,8,0,0,0
set_prop.AF_LOCK(1)
...start, , , ,8,1,0,0
...shot1,400,403,passed,8,1,0,0
...shot2,800,797,passed,8,1,0,0
...set_prop.AF_LOCK(0), , , ,8,0,0,0
...set_aflock(0), , , ,8,0,0,0
4) Testing levent PressSw1AndMF
..disabled
5) Testing SS.Create
...start, , , ,8,0,0,0
...SS.Create success
...>testing SS.MFOn
...SS.MFOn failed
...>testing PT_MFOn
...PT_MFOn success
...start, , , ,0,0,1,0
...shot1,600,595,passed,0,0,1,0
...shot2,1000,1020,passed,0,0,1,0
...disable_PT.MFOn, , , ,0,0,0,0
6) Testing RegisterShootSeqEvent
...RegisterShootSeqEvent failed
7) Done
tests set_focus() and levent PressSw1AndMF fails , although set_aflock() passes
which explains the weird previous results