| Summary: | misconfigured "diff" utility | ||
|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] OE-Core | Reporter: | Juro Bystricky <juro.bystricky> |
| Component: | core | Assignee: | Jussi Kukkonen <jku> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Medium | CC: | jku, meta.mr.watcher, meta.watcher, randy.macleod |
| Version: | 2.3 | ||
| Target Milestone: | 2.4 M2 | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | Don't know | |
|
Description
Juro Bystricky
2017-05-03 21:55:13 UTC
Is this known to break something (other than manually running "diff -l")? Fixing the path is easy (set ac_cv_path_PR_PROGRAM) but then we'd have to RDEPEND on coreutils to actually have /usr/bin/pr ... Assuming that its only diff -l that is broken then hardcoding the path and not adding a dependency seems right. Juro, did you discover this by using diff -l, or some other way? I found out by using diffoscope, binary comparison of two different builds, showing the build directory path. Another possible way (I think) to fix this would be to remove "pr" from HOSTTOOLS in the recipe. Removing pr from HOSTTOOLS causes libav's configure to moan. Lets tell diff what pr to use. Jussi, mind taking this? commit f1a5c7c53be36bbee3b7476b336bab5a04c2d46d Author: Jussi Kukkonen <jussi.kukkonen@intel.com> Date: Fri Jun 9 10:36:18 2017 +0300 diffutils: prevent host path leaks, set pr path Note that this just removes a host path leakage: diffutils still doesn't depend on coreutils so 'pr' might not actually be there on image (this only breaks "diff -l" so no big deal). (From OE-Core rev: 3f7fd738981752a8c34c6cf4a1f2d4c035b081ca) |