| Summary: | eSDK: create test case for populate_sdk_ext using shared states | ||
|---|---|---|---|
| Product: | [Yocto Project Subprojects] eSDK | Reporter: | Jose Perez C <jose.perez.carranza> |
| Component: | eSDK | Assignee: | Unassigned <unassigned> |
| Status: | NEW --- | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium | CC: | anibal.limon, brian.avery, randy.macleod, richard.purdie, ross.burton |
| Version: | 2.4 | ||
| Target Milestone: | Future | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |
| Bug Depends on: | 11623, 11489 | ||
| Bug Blocks: | |||
|
Description
Jose Perez C
2017-05-10 17:34:29 UTC
i would broaden this to have a build of <something large like sato-dev>,with sdk and esdk all being built from sstate. furthermore, it's crucial that the build path for the sstate used and the build path for this test be different. Its also crucial the paths are different *and* the original build path is no longer present and has been deleted. Working on it. Assigned to Joshua due need info for the best way to avoid hard-coding when using sstates. As this is very similar to bug #11500, for which I've already implemented a new buildstep to remove files, I've gone ahead and begun implementation. Essentially we need to add extra buildsteps on end of an existing buildset (i.e. nightly-qa-extras) which deletes tmp, then builds an eSDK in a different TMPDIR. Some buildsteps to run this test have now been enabled on the Yocto Autobuilder: http://git.yoctoproject.org/clean/cgit.cgi/yocto-autobuilder/commit/?id=5d80b30d55c134aeca63643b4ef3611a1845c6d1 However this isn't complete because the way to ensure that shared state objects are used it to generate a locked signature file and use that when running the build. However, as reported in bug #11623, when auto.conf includes a relative path to the generated locked-sigs.inc this value is propagated to the simulated eSDK environment where bitbake is run during eSDK generation, but the file isn't and its absence causes a build error like: ERROR: core-image-sato-1.0-r0 do_populate_sdk_ext: Failed to generate filtered task list for extensible SDK: ### Shell environment set up for builds. ### You can now run 'bitbake <target>' Common targets are: core-image-minimal core-image-sato meta-toolchain meta-ide-support You can also run generated qemu images with a command like 'runqemu qemux86' ERROR: bitbake failed: ERROR: ParseError at /srv/autobuilder/yocto-autobuilder/yocto-worker/nightly-qa-extras/build/build/newtmp/work/qemux86_64-poky-linux/core-image-sato/1.0-r0/sdk-ext/image/tmp-renamed-sdk/conf/local.conf:23: Could not include required file ../locked-sigs.inc ERROR: core-image-sato-1.0-r0 do_populate_sdk_ext: Function failed: copy_buildsystem ERROR: Logfile of failure stored in: /srv/autobuilder/yocto-autobuilder/yocto-worker/nightly-qa-extras/build/build/newtmp/work/qemux86_64-poky-linux/core-image-sato/1.0-r0/temp/log.do_populate_sdk_ext.49644 NOTE: recipe core-image-sato-1.0-r0: task do_populate_sdk_ext: Failed ERROR: Task (/srv/autobuilder/yocto-autobuilder/yocto-worker/nightly-qa-extras/build/meta/recipes-sato/images/core-image-sato.bb:do_populate_sdk_ext) failed with exit code '1' NOTE: Tasks Summary: Attempted 7344 tasks of which 6862 didn't need to be rerun and 1 failed. Moving milestone as dependent bug #11623 isn't going to make M2 This is blocked on the locked-sigs bug, which has moved target milestone again. I'm assigning this issue to Paul so that he can track the dependency and disposition this issue to an appropriate implementer once the dependency bug is resolved. Still worth doing when we have the people available. -- YP bug review. Still worth doing when we have the people available. -- YP bug review. |