I haven't been able to verify the fix for #3591 for some time now. Unfortunately however, commit 822c7b0341064c3d0d9b142dc6ce61c66797159a breaks muxcover, at least for some non-trivial cases. I currently have a design where muxcover after 822c7b0341064c3d0d9b142dc6ce61c66797159a generates x states on the S input of several $_MUX8_ cells. There are no x states present in the design before executing muxcover. Reverting muxcover.cc to before 822c7b0341064c3d0d9b142dc6ce61c66797159a does not generate x states (but still has the cost estimation issue mentioned in #3591). At this point, I cannot provide the actual design that reproduces the issue. I am unsure though, how to create a minimal example.
Expected Behavior
design with correctly connected $_MUX8_ cells
Actual Behavior
design with x states on (at least) S input of $_MUX8_ cells
Version
0.47
On which OS did this happen?
Linux, Windows
Reproduction Steps
I haven't been able to verify the fix for #3591 for some time now. Unfortunately however, commit 822c7b0341064c3d0d9b142dc6ce61c66797159a breaks muxcover, at least for some non-trivial cases. I currently have a design where muxcover after 822c7b0341064c3d0d9b142dc6ce61c66797159a generates x states on the S input of several
$_MUX8_cells. There are no x states present in the design before executing muxcover. Reverting muxcover.cc to before 822c7b0341064c3d0d9b142dc6ce61c66797159a does not generate x states (but still has the cost estimation issue mentioned in #3591). At this point, I cannot provide the actual design that reproduces the issue. I am unsure though, how to create a minimal example.Expected Behavior
design with correctly connected
$_MUX8_cellsActual Behavior
design with x states on (at least) S input of
$_MUX8_cells