| Summary: | Package EFL - build error | ||
|---|---|---|---|
| Product: | buildroot | Reporter: | jhiljanen |
| Component: | Other | Assignee: | unassigned |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | buildroot |
| Priority: | P5 | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Host: | Target: | ||
| Build: | |||
|
Description
jhiljanen
2022-01-20 12:55:10 UTC
Hi jhiljanen, I've tried with Buildroot master branch and I've found 2 missing dependencies: - libinput - libxkbcommon and this deserves a patch But once enabled them efl builds correctly. Which Buildroot version are you using? (In reply to Giulio Benetti from comment #1) I pulled the master from git. I don't have BR2_PACKAGE_EFL_EEZE enabled, because I'm using BusyBox init system. Ah ok, now I can reproduce with BR2_PACKAGE_EFL_EEZE not enabled. This must be fixed. Thanks for reporting. I'm going to fix it. Best regards Hi jhiljanen, I've finally managed the problem. When you state: "I don't have BR2_PACKAGE_EFL_EEZE enabled, because I'm using BusyBox init system." means you're not enabling a dynamic /dev management for BusyBox. To do this you have to enable something that is not: BR2_ROOTFS_DEVICE_CREATION_STATIC so for example: BR2_ROOTFS_DEVICE_CREATION_DYNAMIC_EUDEV This way BR2_PACKAGE_EFL_EEZE will be selectable and efl will build successfully. Here the problem is that efl doesn't really support at the moment the capability to disable eeze, because lot of their modules that depend on eeze don't handle the case of #ifndef HAVE_EEZE/#endif. So the real bug is in efl itself. Here we have 2 ways to go: 1) make efl to depend on BR2_PACKAGE_HAS_UDEV 2) fix efl code and send patches upstream I'm still not sure what to do, I'm going to ask to their Mailing list if they will accept patches through Mailing List by sending with git send-mail(option 2), otherwise I will send a patch for option 1) to Buildroot. Hope this helps Best regards Giulio Benetti Hi jhiljanen, I've found an easier fix with the help of Carsten Haitzler in Enlightment ML. Here is the pending patch: https://patchwork.ozlabs.org/project/buildroot/patch/20220124100920.3005564-1-giulio.benetti@benettiengineering.com/ Best regards (In reply to Giulio Benetti from comment #5) Thank you for your work! (In reply to Giulio Benetti from comment #5) Hi Giulio, I already sent a patch in December: https://patchwork.ozlabs.org/project/buildroot/patch/20211230221904.3064022-1-fontaine.fabrice@gmail.com/ I didn't take time to take into account the remarks from Romain Naour however I think that this approach is better. Best Regards, Fabrice |