Repository navigation
docs(design): research note for bare-IDF/CMake component builds (#1761) - #1969
Merged
Merged
Conversation
Measured idf.py spike + option costs for the "honor" half of #1664.
|
Important Review skippedAuto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Research deliverable for #1761 (the honor half of #1664, whose reject half landed in #1730). Docs only — no implementation and no behaviour change.
What this adds
docs/design/bare-idf-component-builds.md: option comparison, measured costs and a recommendation for genuine bare-IDF / CMake-component projects, plus a contents entry indocs/design/README.mdand a row indocs/INDEX.md.Acceptance criteria from #1761
fcae3288), gccxtensa-esp-elf 14.2.0, a project with amaincomponent (idf_component_register(... REQUIRES mycomp)) and a second componentcomponents/mycomp,app_main()only,sdkconfig.defaults, targetesp32s3.set-target esp32s315.9 s → coldidf.py build50.3 s (1071 ninja actions / 1064 objects / 108 component archives) on a host at loadavg ~60.sdkconfig73,024 B,compile_commands.json7.39 MB / 959 entries,build/144 MB total.ci/idf_appbare-IDF app is 240,368 B on the same IDF 5.5 line, so the spike is representative.integration=idf-cmakemay only be emitted alongside artifact proof (elf/bin bytes, compile-db entry count, sdkconfig digest).Demand evidence (not speculation)
FastLED itself is an IDF component (root
CMakeLists.txt:79callsidf_component_register) and already buildsci/idf_appin.github/workflows/build_bare_idf.ymlacross IDF 5.5 (esp32/esp32s3/c3/c6/p4) plus 5.1 and 4.4. Meanwhile FastLED'sESP32DEV_IDF5_COMPONENTboard (ci/boards.py:900) declaresframework = arduino, espidf, the exact shape fbuild builds through the Arduino core — so fbuild's current claim to exercise component registration is unproven.Cost highlights (measured)
xtensa-esp-elftoolchain 1.1 GB (the default install is 3.8 GB because it also pulls the 2.1 GB riscv32 toolchain, gdb, ulp and dist tarballs); IDF source 667 MB per IDF version; python env 63 MB.install.shbootstrap does not work unmodified on NixOS/uv-CPython here (brokenensurepip;openocdneedslibusb-1.0.so.0), so a delegating path must provision cmake/ninja/python/constraints itself instead of assuming a working host interpreter.Verification
Docs-only change. No Rust/Python changed, so the local unit/lint gates have nothing to exercise; CI is the check. Numbers above are real output from the spike run, reproduced by the recipe in the note. The spike workspace is scratch (not committed):
/home/niteris/dev/hermes/cache/scratch/idf-spike.Deliberately not done: no code, no new dependency, no change to the
espidfrejection, and #1761 stays open for the implementation follow-ups.