Cadence Encounter detects but does not fix the setup time violations when a multi-cycle path with 0 number of cycles is defined

Hi all,

I am experiencing a small problem where Encounter detects setup violation but doesn't attempt to fix it. Following is the brief description of my problem reproduced on a very simple circuit.

############### SDC file start
create_clock -name "clock1" -add -period 1.5 -waveform {0.0 0.75} [get_ports clk]
set_multicycle_path -from {dff1} -to {dff3} -setup 0
set_multicycle_path -from {dff2} -to {dff3} -setup 0
set_dont_touch [get_cells buf1]
############### SDC file end

Notice that for the "dff1 to dff3" path, setup check must be done on the first (rising) edge of their respective clocks. As such, encounter should advance the clock driving dff3 to meet setup violation at dff3. 

The "skewClock" command prints following warning.
"Analyzing useful skew ...
skewClock did not found any end points to delay or to advance."

The optimization commands before clock design used are as follows --
########################### CTS commands in encounter script
setOptMode -setupTargetSlack 0.05
setOptMode -usefulSkew true
setUsefulSkewMode -maxSkew true -ecoRoute true -noBoundary true -allNegEndPoints true
setCTSMode -opt true -optAddBuffer true
clockDesign
###########################

If there is any way we can get encounter to fix setup violation at dff3, it will be great

Additional points to note. The circuit has plenty of area to add buffers and extra routing. There are no hold violations that restricts encounter from adding a buffer (hold slack is +1700ps whereas setup slack is -343ps)