| Summary: | "libcomps: update to 0.1.12" doesn't work due to patch not applying | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] OE-Core | Reporter: | Scott Branden <scott.branden> |
| Component: | devtools / tool chain | Assignee: | Unassigned <unassigned> |
| Status: | RESOLVED WORKSFORME | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium | CC: | jon.richardson, meta.mr.watcher, meta.watcher, richard.purdie |
| Version: | unspecified | ||
| Target Milestone: | 3.1 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | Don't know | |
|
Description
Scott Branden
2019-12-19 21:54:11 UTC
It seems to be an issue with quilt. I had the same problem with pseudo and had to apply this patch locally: commit 0e3323b123ab91c2ab76c71c955eae758a084bdc Author: Jonathan Richardson <jonathan.richardson@broadcom.com> Date: Thu Nov 21 11:49:43 2019 -0800 SOC-8238: Use git as patchtool for pseudo recipe The default patchtool quilt sometimes fails to apply the patch 0001-configure-Prune-PIE-flags.patch even though there is nothing wrong with it. This causes the build to fail. Since all the patches are git patches, the patchtool is being changed to git to work around the issue. Change-Id: I9f48d627ae04bb0524368804597548249e3fcd7d Reviewed-on: http://gerrit-ccxsw.broadcom.net/107661 Reviewed-by: Jonathan Richardson <jonathan.richardson@broadcom.com> Tested-by: Jonathan Richardson <jonathan.richardson@broadcom.com> diff --git a/meta/recipes-devtools/pseudo/pseudo.inc b/meta/recipes-devtools/pseudo/pseudo.inc index 8b34909726..8ad98b386e 100644 --- a/meta/recipes-devtools/pseudo/pseudo.inc +++ b/meta/recipes-devtools/pseudo/pseudo.inc @@ -153,3 +153,5 @@ do_install_append_class-nativesdk () { } BBCLASSEXTEND = "native nativesdk" + +PATCHTOOL = "git" Which host OS is this one? We test these patches on the autobuilder and haven't see this issue... This is on a Ubuntu 18.04 x86-64 host I just tried with the default build/conf again and problem did not happen. Will see if I can reproduce again. It doesn't happen all the time. It's sporadic. But if it happens once it will again eventually. *** Bug 13719 has been marked as a duplicate of this bug. *** We're going to struggle to help with this unless we can reproduce the problem. Is there some way we could reproduce the issue? I agree it's difficult to reproduce random yocto build issues. Especially since once it builds successfully then the output is in the yocto cache and we don't hit the issue any more. To test this and other problems you probably need some sort of automated build system that builds with and without yocto-cache populated. As well as some level of randomness with build thread scheduling by loading the build machine? I haven't seen the issue since because I now use the yocto-cache. I deleted it and tried again and wasn't able to reproduce in 3 attempts. unable to reproduce at this time. can probably close if you like as it is/was random. Not able to reproduce. |