CircleCI uses cached FMS image even when Dockerfile changes for FMS stage #30
Comments
nbren12
pushed a commit
that referenced
this issue
Sep 24, 2021
* add new suite definition files for GFSv15p2, GFSv16beta * update existing suite definition files because of the interstitial code rearrangement (NCAR/ccpp-physics#372) * clean up CCPP compiler flags, add -Wall for GNU in DEBUG mode * add satmedmfvifq, shalcnv, sascnvn, Ferrier-Aligo microphysics and dependencies to CCPP prebuild config * bug fixes for ugwp and noahmp * move previous-timestep precipitation variables from Tbd to Sfcprop (as recommended by @HelinWei-NOAA, see NCAR/ccpp-physics#319) * fix compiler warnings about non-existent include directories * cleanup of old comments in GFS_physics_driver.F90 (see NCAR#4, NCAR#16, and NCAR/ccpp-physics#319 * CCPP annotations in GFS_driver.F90, GFS_radiation_driver.F90, GFS_physics_driver.F90 (comments that describe where code blocks ended up in the CCPP interstitial code) * capability to coldstart FV3 with GSD physics from RAP/HRRR initial conditions (required for FV3 SAR) * new suite definition file for coupled model
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
In #29 I changed the build instructions for the FMS stage of the Dockerfile (8f5a3af) and yet CircleCI still pulled a pre-built FMS docker image instead of rebuilding it. This mean the regression tests passed when they shouldn't have.
The text was updated successfully, but these errors were encountered: