call _test_step instead of test_step in Bundle easyblock - #4112
Conversation
|
For testing, I've added this diff to X11: diff --git a/easybuild/easyconfigs/x/X11/X11-20250608-GCCcore-14.3.0.eb b/easybuild/easyconfigs/x/X11/X11-20250608-GCCcore-14.3.0.eb
index 997dc16ba5..b45fc0f546 100644
--- a/easybuild/easyconfigs/x/X11/X11-20250608-GCCcore-14.3.0.eb
+++ b/easybuild/easyconfigs/x/X11/X11-20250608-GCCcore-14.3.0.eb
@@ -51,6 +51,7 @@ default_component_specs = {
components = [
('libpthread-stubs', '0.5', { # 2023-07-18
'checksums': ['593196cc746173d1e25cb54a93a87fd749952df68699aab7e02c085530e87747'],
+ 'runtest': 'Nonsense',
}),
('xorgproto', '2024.1', { # 2024-03-26
'checksums': ['4f6b9b4faf91e5df8265b71843a91fc73dc895be6210c84117a996545df296ce'],without this EasyBlock: https://gist.github.com/Thyre/0432c5e1dc3d89d2541a048394dcc9ee It's worth noting that this naming change shows up in the logs: But I think this fix is worth having this slight change in the logs. We could still improve this by doing some string manipulation to replace a leading comp.run_step(step_name, [lambda x: getattr(x, '%s_step' % step_name)]) |
|
Test report by @Thyre Overview of tested easyconfigs (in order)
Build succeeded (with --ignore-test-failure) for 1 out of 1 (total: 8 mins 48 secs) (1 easyconfigs in total) |
|
Test report by @Thyre Overview of tested easyconfigs (in order)
Build succeeded (with --ignore-test-failure) for 1 out of 1 (total: 8 mins 49 secs) (1 easyconfigs in total) |
(created using
eb --new-pr)Partially addresses #4044.
The
Bundleeasyblock can now correctly handle--ignore-test-failure, but renaming the function on the framework level is still the better option in the long run...