#227 Erroneous messages for duplicate cells. Issue created by d-m-bailey During caravel

make ship

, cells are read from multiple locations.

load user_project_wrapper; \
		property LEFview true; \
		property GDS_FILE $(UPRJ_ROOT)/gds/user_project_wrapper.gds; \
		property GDS_START 0; \
		load $(UPRJ_ROOT)/mag/user_id_programming; \
		load $(UPRJ_ROOT)/mag/user_id_textblock; \
		load $(CARAVEL_ROOT)/macros/simple_por/maglef/simple_por; \
		load $(UPRJ_ROOT)/mag/caravel_core; \
		load $(CARAVEL_ROOT)/mag/caravel -dereference; \

The mag files contain an explicit reference for the first cell loaded, but implicit references for subsequent cells.

use sky130_fd_sc_hd__diode_2  ANTENNA__061__A0 $PDKPATH/libs.ref/sky130_fd_sc_hd/mag
timestamp 1665323087
transform -1 0 6716 0 1 2720
box -38 -48 222 592
use sky130_fd_sc_hd__diode_2  ANTENNA__062__B
timestamp 1665323087
transform -1 0 6072 0 1 1632
box -38 -48 222 592

It appears that these implicit references are not recorded properly and cause a false mismatch when subsequent cells are read which results in a cell rename.

Duplicate cell in housekeeping:  Instance of cell sky130_fd_sc_hd__nand2_1.mag is from path /home/user/mpw-9/dependencies/pdks/sky130A/libs.ref/sky130_fd_sc_hd/mag but cell was previously read from /home/user/mpw-9/caravel_user_project3/mag/dependencies/pdks/sky130A/libs.ref/sky130_fd_sc_hd/mag.
Cell name conflict:  Renaming original cell to sky130_fd_sc_hd__nand2_1#0.

The previously read location does not exist - I think, it has been recorded incorrectly.

/home/user/mpw-9/dependencies/pdks/sky130A/libs.ref/sky130_fd_sc_hd/mag

is the correct PDK location. The recorded previous location

/home/user/mpw-9/caravel_user_project3/mag

is the location of the block that was opened and has no sub-directory

dependencies

. This error only occurs if there are implicit instances in the mag file. If there is only one instance of a cell in the previously loaded block, there do not appear to be any errors. Running

make ship

with the

caravel/Makefile

substitutions at the top, should recreate the problem. (There are other problems with the

Makefile

though). RTimothyEdwards/magic