questions about the gds generated by "make test"

[stefanottili]

I just installed OpenLane and ran "make test".

Looking thru the directories, if found that the 3 gds written by magic are identical:
signoff/spm.gds
signoff/spm.magic.gds
final/gds/spm.gds

signoff/spm.klayout.gds has the same geometry, but different layer mapping

a) why are there 3 redundant gds saved in the flow ?
b) wouldn't it be better to have these 4 gds gzipped ?
c) why would signoff/spm.klayout.gds not use the same layer mapping as the other gds ?
d) is there any plan to use the klayout lefdef map file for klayout ?
see: The-OpenROAD-Project/OpenROAD#1919
Using the map file allows to put def PIN on both metal and pin layer:datatype.
e) klayout also has "strm2gds/oas" buddy tools that can be used for def2gds/oas conversion.

[kareefardi]

So the idea was to create gds files using multiple signoff tools (klayout and magic) and then based on the configuration select on of them hence the copy signoff/spm.gds and then the final directory will always have copies from other parts of the flow.
At the time of implementation magic didn't support .gds.gz (it should now) so perhaps we should move to .gds.gz
Finally, yes there is a plan to use klayout lefdef map file see: efabless/sky130_klayout_pdk#2
As for oas that might be an enhancement we can look towards adding it