Bug 15254 - Building webkitgtk fails on beaglebone-yocto machine
Summary: Building webkitgtk fails on beaglebone-yocto machine
Status: RESOLVED FIXED
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: oe-core other (show other bugs)
Version: 4.3
Hardware: x86 Multiple
: Medium normal
Target Milestone: 5.1
Assignee: Ross Burton
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2023-10-26 16:29 UTC by Michael Opdenacker
Modified: 2024-10-03 15:01 UTC (History)
3 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: No (bug/feature does not impact docs)


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Michael Opdenacker 2023-10-26 16:29:42 UTC
Configuring webkitgtk on nanbield and kirstone fails for the "beaglebone-yocto" machine (ARM Cortex A8):

| CMake Error at Source/cmake/WebKitFeatures.cmake:321 (message):
|   ENABLE_JIT conflicts with ENABLE_C_LOOP.  You must disable one or the
|   other.

That's because of this line in the recipe:
-DENABLE_JIT=${@bb.utils.contains('TUNE_FEATURES', 'callconvention-hard', 'ON', 'OFF', d)}"

and the fact that the image is built with "callconvention-hard", which enables ENABLE_JIT, which is incompatible with ENABLE_C_LOOP which seems to be set by default.

I tried to modify the recipe (meta/recipes-sato/webkit/webkitgtk_2.40.5.bb) as follows, replacing:
EXTRA_OECMAKE:append:armv7a = " -DENABLE_JIT=${@bb.utils.contains('TUNE_FEATURES', 'callconvention-hard', 'ON', 'OFF', d)}"
by
EXTRA_OECMAKE:append:armv7a = " -DENABLE_JIT=OFF "

Configure works then, but then compiling fails after a long time, from an error that is not detailed by "ninja", just mentioning that one of the compile commands returned with a failed status (which cause doesn't seem to appear in the compiler output).

The error doesn't happen for the qemu86-64 and qemuarm64 machines.

However, doesn't the autobuilder try to build "world" (corresponding to OE-core recipes) for all supported machines?
Comment 1 Michael Opdenacker 2023-10-26 18:52:41 UTC
Also tried on Cortex A7 (stm32mp1 actually). This machine doesn't have this issue.
Comment 2 Ross Burton 2024-04-11 14:52:53 UTC
Bumping, and copying Jon.
Comment 3 Ross Burton 2024-10-03 15:01:24 UTC
Fixed in 26a34fd71659e32e56ccc4f23e79aa62f4bc062d.