Johan Euphrosine

I'd argue that it would be best for `make` to always surface build error messages (and exit with the corresponding exit code) as it makes it challenging to debug in...

+1, something like: ``` diff --git a/Makefile b/Makefile index e1d61fb..dc0efc4 100644 --- a/Makefile +++ b/Makefile @@ -18,12 +18,10 @@ INSTALL_CAD_DIRS = windows doc ${TECH} all: $(ALL_TARGET) standard: - @echo ---...

I think in that case you'd want to either: - add `-o pipefail` to `SHELL` - add `|| exit ${PIPESTATUS[0]}` to make sure you propagate the `make` exit code back...

maybe it's best to keep the `Makefile` simple and document in the `README` how to get it to log a file if needed with `make 2>&1 | tee make.log`, wdyt?

instead the `Makefile` should probably do something like `(cd $$dir && ${MAKE} module) || exit 1` to ensure that the build fails early and terminate with the appropriate error code.

Possibly related to https://github.com/The-OpenROAD-Project/OpenLane/issues/1380 ?

@vijayank88 can you point to the commits that fixed this issue? (I still get it on OpenLane master)

Since this script is triggered by `DIODE_INSERTION_STRATEGY==3`: https://github.com/The-OpenROAD-Project/OpenLane/blob/ebad315d1def25d9d253eb2ec1c56d7b4e59d7ca/scripts/tcl_commands/routing.tcl#L67 setting it to another value could be a temporary workaround to skip that step.

however setting `DIODE_INSERTION_STRATEGY=0` ends up triggering the following error later in the flow: ``` [INFO]: Running OpenROAD Antenna Rule Checker... [ERROR]: during executing: "openroad -python /OpenLane/scripts/extract_antenna_violators.py --output /OpenLane/designs/inverter/runs/RUN_2022.07.12_15.44.38/reports/signoff/34-antenna_violators.rpt /OpenLane/designs/inverter/runs/RUN_2022.07.12_15.44.38/reports/signoff/34-antenna.rpt" [ERROR]:...

> Would you describe how this would be advantageous over "openroad -python"? It does make it easier to use openroad from other python projects or from a jupyter notebook. >...