Compile error if CONFIG_FREERTOS_HZ != 1000 #6955
Conversation
|
|
|
Ha, looks like the CI build ran afoul of this check. From the conversation when that check was added:
Working as intended, I suppose... this change explicitly makes it so the default config does not build. Tracing through the CI scripts, it looks like it starts with a standard espressif/idf:release-v4.4 container: arduino-esp32/.github/workflows/push.yml Line 82 in c93bf11 Is the right thing to add some steps to that workflow that create an sdkconfig.defaults (or point to an sdkconfig.ci)? Or something else? |
|
(By the way @atanisoft, I attempted to request re-review in github, but it failed, seemingly due to this issue: community/community#11738) |
VojtechBartoska
added
Area: Arduino as ESP-IDF component
|
In order for CI to pass, we either need to ensure that proper option is selected here or have the CI add |
Based on my understanding of the Kconfig language, you can't set reverse dependencies to force an int option (like FREERTOS_HZ) to a specific value based on a different config. So unless you know some trick I don't, I'll look into some sort of |
|
(Let's see if this fixes the CI. Sorry about the reviewer flip-flopping, I was confused by the interface.) |
|
Thanks @egnor ! CI passes now :) |
* Enable configuring target variants (espressif#7019) This change enables picking the right board configuration from the variants/ folder. Previously, we would always pick the default configuration (e.g., "esp32" instead of "heltec_wifi_lora_32_V2"). * Update .gitignore (espressif#7021) Add entries to gitignore * Update esptool to version 4.2.1 (espressif#7127) * Update esptool to version 4.2.1 * Fix esptool for MacOS * Esptool v4.2.1 for CI Platformio (espressif#7147) to complete the test. Co-authored-by: Jason2866 <[email protected]> * Update get.py to support python 3.10+ (espressif#7166) * Update get.py to support python 3.10+ * Use try/except to remove version check * fixed names on the VID and PID for boards (espressif#7144) somehow had duplicated the same name across all my boards. the PID's and VID's were correct, but the name was the same for all of them. that is fixed now. * Compile error if CONFIG_FREERTOS_HZ != 1000 (espressif#6955) * Compile error if CONFIG_FREERTOS_HZ != 1000 * add a check at the CMake level, per feedback * Set CONFIG_FREERTOS_HZ=1000 in CI test of Arduino-as-component * Adding u-blox NORA-W10 series (ESP32-S3) (espressif#7191) * Create pins_arduino.h * Update boards.txt * Update boards.txt * Fixed espressif/esp-rainmaker#152 (espressif#7171) * Update PlatformIO build scripts (espressif#7200) This update includes the following: - Implemented an additional build step that produces an adjusted bootloader image with updated headers according to selected flash mode and size values. This step is only executed for debugging or uploading via debug probes. - Implemented a basic mechanism to dynamically add an extra UF2 bootloader image if corresponding partition is selected (e.g. for Adafruit targets) - Minor code formatting Co-authored-by: Dr. Christian Kohlschütter <[email protected]> Co-authored-by: Me No Dev <[email protected]> Co-authored-by: Paul Price <[email protected]> Co-authored-by: Daniel Egnor <[email protected]> Co-authored-by: Michael Ammann <[email protected]> Co-authored-by: Sanket Wadekar <[email protected]> Co-authored-by: Valerii Koval <[email protected]>
Description of Change
Since
delay()won't work properly withCONFIG_FREERTOS_HZnot set to 1000, this adds a_Static_assertto verify that fact. (The standard Arduino build already sets the value properly; this is for people using the Arduino platform as a component within their own ESP-IDF build.)Tests scenarios
Tested with a platformio build that left
CONFIG_FREERTOS_HZat the default (100) and verified build failure.Then added
CONFIG_FREERTOS_HZ=1000toconfig.defaults, erasedconfig.esp32and built again, verifying success this time.(This particular build happens to be for the OLIMEX ESP32-POE, but none of this has any impact on runtime-- runtime code is unchanged.)
Related links
Addresses #6946