update scipy easyblock for scipy >= 1.9.0 to use meson/ninja #2862
Conversation
|
figuring out the best way to ensure all the tests are run in a way they don't fail silently |
|
Test report by @jfgrimm |
This comment was marked as outdated.
This comment was marked as outdated.
|
Test report by @jfgrimm Overview of tested easyconfigs (in order)
Build succeeded for 0 out of 2 (2 easyconfigs in total) |
|
test failures actually are present in old builds, they were just ignored (see #2237) |
|
Test report by @jfgrimm Overview of tested easyconfigs (in order)
Build succeeded for 2 out of 2 (2 easyconfigs in total) |
easybuild/easyblocks/s/scipy.py
Outdated
| extra_vars = ({ | ||
| 'enable_slow_tests': [False, "Run scipy test suite, including tests marked as slow", CUSTOM], | ||
| # ignore tests by default to maintain behaviour in older easyconfigs | ||
| 'ignore_test_result': [True, "Run scipy test suite, but ignore result (only log)", CUSTOM], |
There was a problem hiding this comment.
We should set this to None, and auto-enable for scipy versions >= 1.9.0 unless it's explicitly set to False or True in easyconfig?
easybuild/easyblocks/s/scipy.py
Outdated
|
|
||
|
|
||
| class EB_scipy(FortranPythonPackage): | ||
| class EB_scipy(FortranPythonPackage, MesonNinja): |
There was a problem hiding this comment.
I'm a bit surprised that using MesonNinja works to install scipy as an extension, since that generic easyblock doesn't derive from ExtensionEasyBlock (which is sort of required for stuff you want to install both as extension and stand-alone).
That's probably because MesonNinja is only used as "secondary" easyblock, next to FortranPythonPackage, but to make this future-proof (there will come a time where deriving from FortranPythonPackage is useless since the scipy versions we still support are all >= 1.9.0) we may need to be a bit more careful here (which could be as simple as changing MesonNinja to derive from ExtensionEasyBlock, rather than from EasyBlock like it does now).
There was a problem hiding this comment.
so I had a bit of a closer look, and the key thing that makes this work as-is seems to be calling an __init__ (e.g. PythonPackages.__init__ ) which derives from ExtensionEasyBlock
With the new changes, the meson build still makes use of PythonPackage anyway (to get the python command, and for the sanity check), so now removing FortranPythonPackage would not break anything.
|
Test report by @boegel Overview of tested easyconfigs (in order)
Build succeeded for 8 out of 8 (8 easyconfigs in total) |
easybuild/easyblocks/s/scipy.py
Outdated
| if self.cfg['ignore_test_result'] is None: | ||
| self.cfg['ignore_test_result'] = True | ||
| # ignore installopts inherited from FortranPythonPackage | ||
| self.cfg['installopts'] = "" |
There was a problem hiding this comment.
We should at least also log the original installopts value if we're going to do a hard reset here...
There was a problem hiding this comment.
updated to strip the ones inherited from PythonPackage, perhaps that's better?
From a build of easybuilders/easybuild-easyconfigs#16912 started at 11:19:59 - so with 016118a |
Co-authored-by: Simon Branford <[email protected]>
simplify logic to set default value for ignore_test_result in scipy easyblock
|
Test report by @jfgrimm Overview of tested easyconfigs (in order)
Build succeeded for 14 out of 14 (11 easyconfigs in total) |
|
Test report by @boegel Overview of tested easyconfigs (in order)
Build succeeded for 11 out of 11 (11 easyconfigs in total) |
(created using
eb --new-pr)