.murdock: drop Kconfig compile check

[benpicco]

Contribution description

We want to remove Kconfig, so don't fail CI when the Kconfig build no longer aligns with the classic make based build.

Testing procedure

CI should continue functioning for non-Kconfig builds.

Issues/PRs references

https://forum.riot-os.org/t/last-chance-for-kconfig-dependency-modelling/4068

[riot-ci]

Murdock results

✔️ PASSED

ad70211 .murdock: drop Kconfig compile check

Success	Failures	Total	Runtime
8101	0	8101	17m:43s

Artifacts

• Documentation preview

[MrKevinWeiss]

I think there are a bunch more things that can be removed here but lets wait until the specified date (2024) though from the complete silence of the thread I would not expect much.

[benpicco]

I was just wondering whether we should still put work into resolving Kconfig issues in open PRs (#19540, #19786) or if those should just wait until the new year.

[benpicco]

@MrKevinWeiss it's 2024

[MrKevinWeiss]

@MrKevinWeiss it's 2024

Wow, you really don't waste time. OK I will have a look :)

[MrKevinWeiss]

Probably we should get rid of everything?

If you want I can directly push?

diff --git a/.murdock b/.murdock
index e0edec251a..5a5f8d8ee1 100755
--- a/.murdock
+++ b/.murdock
@@ -40,125 +40,6 @@ esac
 # is in
 : ${TEST_BOARDS_LLVM_COMPILE:="iotlab-m3 native nrf52dk mulle nucleo-f401re samr21-xpro slstk3402a"}
 
-: ${TEST_KCONFIG_BOARDS_AVAILABLE:="
-adafruit-itsybitsy-m4
-arduino-due
-arduino-leonardo
-arduino-mkrzero
-arduino-mega2560
-arduino-nano
-atxmega-a1-xplained
-atxmega-a3bu-xplained
-avr-rss2
-avsextrem
-bastwan
-bluepill
-b-l475e-iot01a
-cc1352-launchpad
-cc2650-launchpad
-derfmega128
-dwm1001
-esp32-ci
-esp32-heltec-lora32-v2
-esp32-olimex-evb
-esp32c3-ci
-esp32c3-devkit
-esp32s3-ci
-esp32s3-devkit
-esp32s2-ci
-esp32s2-devkit
-esp8266-ci
-esp8266-esp-12x
-hamilton
-hifive1
-mbed_lpc1768
-mega-xplained
-microbit
-mulle
-native
-nrf52840dk
-nrf9160dk
-nucleo-f072rb
-nucleo-f103rb
-nucleo-f207zg
-nucleo-f334r8
-nucleo-f429zi
-nucleo-f767zi
-nucleo-g071rb
-nucleo-g474re
-nucleo-l011k4
-nucleo-l073rz
-nucleo-l152re
-nucleo-l433rc
-nucleo-l552ze-q
-pba-d-01-kw2x
-p-nucleo-wb55
-qn9080dk
-remote-revb
-rpi-pico
-same54-xpro
-samr21-xpro
-seeedstudio-gd32
-slstk3400a
-sltb001a
-slwstk6220a
-stm32f723e-disco
-stm32mp157c-dk2
-waspmote-pro
-waveshare-nrf52840-eval-kit
-weact-f401ce
-z1
-"}
-
-: ${TEST_KCONFIG_ENFORCE_APP_GROUPS:="
-examples/hello-world
-tests/build_system/external_board_dirs
-tests/drivers/b*
-tests/drivers/f*
-tests/drivers/g*
-tests/drivers/h*
-tests/drivers/i*
-tests/drivers/j*
-tests/drivers/l*
-tests/drivers/o*
-tests/drivers/p*
-tests/drivers/q*
-tests/drivers/r*
-tests/drivers/s*
-tests/drivers/t*
-tests/drivers/u*
-tests/drivers/v*
-tests/net/ieee802154_hal
-tests/periph/*
-tests/pkg/elk
-tests/pkg/uzlib
-tests/sys/cb_mux*
-tests/sys/congure_*
-tests/sys/prng_*
-tests/sys/trace
-tests/sys/xtimer_*
-tests/sys/ztimer_*
-"}
-
-# This list prevents boards from being tested by the TEST_KCONFIG_TEST_ALLOWLIST
-# This should only be used for boards that require a lot of work, such as,
-# requiring modelling of a clock tree.
-# As a rule of thumb, only add boards here if there are not differences in
-# modules or packages.
-# Eventually this list will be removed...
-#
-: ${TEST_KCONFIG_BOARD_BLOCKLIST:="
-"}
-
-# This list will force all boards that are not in the TEST_KCONFIG_BOARD_BLOCKLIST
-# to be tested on the apps. The purpose is to prevent new boards from skipping
-# the kconfig implementations.
-: ${TEST_KCONFIG_TEST_ALLOWLIST:="
-examples/hello-world
-tests/drivers/saul
-tests/drivers/mtd_mapper
-"}
-
 : ${TEST_WITH_CONFIG_SUPPORTED:="examples/suit_update tests/drivers/at86rf2xx_aes"}
 
 export RIOT_CI_BUILD=1
@@ -251,54 +132,6 @@ fi
 
 CFCR_ARGS="--upstreambranch ${CI_BASE_COMMIT}"
 
-get_supported_kconfig_board_app() {
-    local board=$1
-    local appdir=$2
-    if is_in_list "${board}" "${TEST_KCONFIG_BOARD_BLOCKLIST}"; then
-        return 1
-    fi
-
-    # On nightlies run all possible kconfig tests on all boards
-    # Normally we don't want to do this as it adds quite a bit to build time
-    # and the subset of boards we are testing for are pretty good at catching
-    # any bugs...
-    # This will be over one day, I promise.
-    if [ ${NIGHTLY} -eq 1 ]; then
-        if is_in_list "${appdir}" "${TEST_KCONFIG_ENFORCE_APP_GROUPS}"; then
-            return 0
-        fi
-        if [ -f "${appdir}/app.config.test" ]; then
-            return 0
-        fi
-    fi
-    if is_in_list "${appdir}" "${TEST_KCONFIG_TEST_ALLOWLIST}"; then
-        return 0
-    fi
-    if is_in_list "${board}" "${TEST_KCONFIG_BOARDS_AVAILABLE}"; then
-        if is_in_list "${appdir}" "${TEST_KCONFIG_ENFORCE_APP_GROUPS}"; then
-            return 0
-        fi
-        if [ -f "${appdir}/app.config.test" ]; then
-            return 0
-        fi
-    fi
-    return 1
-}
-
-kconfig_module_packages_diff() {
-    local board=$1
-    local appdir=$2
-
-    make clean --no-print-directory -C ${appdir} BOARD=${board}
-    mmp=$(make --no-print-directory info-modules info-packages -C ${appdir} BOARD=${board})
-
-    make clean --no-print-directory -C ${appdir} BOARD=${board}
-    kmp=$(TEST_KCONFIG=1 make --no-print-directory info-modules info-packages -C ${appdir} BOARD=${board})
-
-
-    bash -c "diff <(echo \"${mmp}\") <(echo \"${kmp}\")"
-}
-
 # if RUN_TESTS is unset (e.g., not passed from the outside),
 # set to 1 if NIGHTLY=1 or if the label "CI: run tests" is set,
 # otherwise set 0.
@@ -531,14 +364,12 @@ compile() {
     # reduces disk IO.
     export CCACHE_TEMPDIR="$(pwd)/.ccache/tmp"
 
-    # compile without Kconfig
     BOARD=${board} make -C${appdir} clean
     CCACHE_BASEDIR="$(pwd)" BOARD=$board TOOLCHAIN=$toolchain RIOT_CI_BUILD=1 \
         make -C${appdir} all test-input-hash -j${JOBS:-4}
     RES=$?
 
-    # test hash is used to cache test results, not for comparing binaries
-    # generated with and without KConfig
+    # test hash is used to cache test results
     test_hash=$(test_hash_calc "$BINDIR")
 
     # run tests

[MrKevinWeiss]

Looks good, lets see what happens! ACK!

[benpicco]

Thank you for helping to bring the Kconfig migration to a close!