On 05-05-2026 15:01, Richard Purdie wrote:
CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know
the content is safe.
On Tue, 2026-05-05 at 01:16 -0700, Varatharajan, Deepesh via
lists.openembedded.org wrote:
From: Deepesh Varatharajan<[email protected]>
While running testsuite set CLANG_ENABLE_TESTSUITE = "1"
Signed-off-by: Deepesh Varatharajan<[email protected]>
---
meta-selftest/recipes-devtools/clang/clang_%.bbappend | 2 ++
meta-selftest/recipes-devtools/clang/lld_%.bbappend | 2 ++
meta-selftest/recipes-devtools/clang/llvm_%.bbappend | 2 ++
3 files changed, 6 insertions(+)
create mode 100644 meta-selftest/recipes-devtools/clang/clang_%.bbappend
create mode 100644 meta-selftest/recipes-devtools/clang/lld_%.bbappend
create mode 100644 meta-selftest/recipes-devtools/clang/llvm_%.bbappend
diff --git a/meta-selftest/recipes-devtools/clang/clang_%.bbappend
b/meta-selftest/recipes-devtools/clang/clang_%.bbappend
new file mode 100644
index 0000000000..32c236b972
--- /dev/null
+++ b/meta-selftest/recipes-devtools/clang/clang_%.bbappend
@@ -0,0 +1,2 @@
+# Enable CLANG testsuite
+CLANG_ENABLE_TESTSUITE = "1"
diff --git a/meta-selftest/recipes-devtools/clang/lld_%.bbappend
b/meta-selftest/recipes-devtools/clang/lld_%.bbappend
new file mode 100644
index 0000000000..32c236b972
--- /dev/null
+++ b/meta-selftest/recipes-devtools/clang/lld_%.bbappend
@@ -0,0 +1,2 @@
+# Enable CLANG testsuite
+CLANG_ENABLE_TESTSUITE = "1"
diff --git a/meta-selftest/recipes-devtools/clang/llvm_%.bbappend
b/meta-selftest/recipes-devtools/clang/llvm_%.bbappend
new file mode 100644
index 0000000000..32c236b972
--- /dev/null
+++ b/meta-selftest/recipes-devtools/clang/llvm_%.bbappend
@@ -0,0 +1,2 @@
+# Enable CLANG testsuite
+CLANG_ENABLE_TESTSUITE = "1"
Having bbappends inside meta-selftest isn't something we've done before
and feels a bit odd. I think it will cause yocto-check=layer to fail
too.
Why wouldn't we just set this variable inside the tests that need it?
As you mentioned, yocto-check-layer is failing due to a signature
difference with this change. The reason for placing this in
meta-selftest is that we’ve observed intermittent issues when reusing
sstate, where the required flags are not always enabled, leading to
testsuite failures. To make the behavior more consistent and avoid these
sstate-related inconsistencies, we opted to set the variable in
meta-selftest, since it is included by default when running the
testsuite. That said, I understand the concern with this approach given
the layer check failure. Is there a better way you’d recommend handling
this kind of sstate-related issue? Regards, Deepesh
Cheers,
Richard
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#236480):
https://lists.openembedded.org/g/openembedded-core/message/236480
Mute This Topic: https://lists.openembedded.org/mt/119157532/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
