| Summary: | recipetool has poor python recipe failure mode | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] OE-Core | Reporter: | Richard Purdie <richard.purdie> |
| Component: | Scripts and Tools | Assignee: | Ross Burton <ross.burton> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium+ | CC: | adrian.freihofer, randy.macleod, ross.burton, tgamblin, tim.orling |
| Version: | 0.0.0 | ||
| Target Milestone: | 5.0 M3 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |
|
Description
Richard Purdie
2024-02-09 09:36:34 UTC
The related oe-selftest failure came in with 5206056589ca6b0671b0009393f526828e2c665a. I can confirm that this commit was tested on a host where import tomllib in Python works. Improving recipetool so that it exits with a correct error code should be done, I agree. But that will not fix the error of the oe-selftest on the AB. The selftest checks if the bb file has been created. But this is not the case if recipetool has no tomllib or tomli available. I see different possible solutions: - Make tomllib a host dependency and fix the AB by installing this Python package - Keep tomllib optional and let recipe tool ask the user for installing it when it is required. If this is the preferred solution, I need to improve the two test cases. Just saw the patches from RP on master-next. Looks like this problem is already solved. Thank you. I've sent a patch to clean up the horrible exception dumping. Tested master-next with Python 3.12 on Fedora 39 again. Looks good! Thank you. Fixed in 1a3ba4c312844d80ae382912b319e60ad8b30737. |