adopt module_load_environment in Bundle easyblock

[lexming]

Update of Bundle easyblock for #3527

Also fixes issue #3528

[boegel]

@lexming I suspect that #3509 will get merged soon, which will cause trouble/conflicts here.

Maybe we can revert the changes to the Bundle easyblock made here for now, and postpone that to another PR to avoid a complex conflict resolution situation?

[Thyre]

@lexming I suspect that #3509 will get merged soon, which will cause trouble/conflicts here.

Maybe we can revert the changes to the Bundle easyblock made here for now, and postpone that to another PR to avoid a complex conflict resolution situation?

I think that this should not cause a merge conflict. #3509 doesn't touch the lines changed by this PR and comp, used by this PR, is still defined.
However, having both module_load_environment and make_module_req_guess in the same EasyBlock is probably not a good idea.

I think there are two options:

• We can move the changes in Bundle to a separate PR
• Update this PR if enhance generic Bundle easyblock to transfer module requirements of components, but do not create logfile in components #3509 is merged before this one.

[lexming]

Having both module_load_environment and make_module_req_guess is indeed undesirable. I propose that #3509 gets merged first and then I'll update this PR to adapt those changes to the new approach.

I changed the scope of this PR to only concern Bundle.

[boegel]

@lexming I think this PR should only get merged if all other easyblocks have been updated to use module_load_environment rather than customizing make_module_req_guess, since with the changes made here customized make_module_req_guess methods for components will be ignored?

[boegel]

@lexming Taking this one step further: shouldn't we be taking into account here that one of bundle components is being installed with an easyblock that has not migrated to module_load_environment yet?

We'll do so for all easyblocks under our control, but we also need to take into account that people may have custom easyblocks that are still using make_module_req_guess, so we should at least consider it too still (along with a deprecation warning, perhaps)?

[boegel]

@lexming You should sync up your branch with current 5.0.x, since the Bundle easyblock now uses make_module_req_guess in two places (after the merge of #3509, which was synced into 5.0.x in #3542)

[lexming]

@boegel LGTM, will send some tests...

[lexming]

Test report by @lexming

Overview of tested easyconfigs (in order)

• SUCCESS buildenv-default-foss-2023a.eb
• SUCCESS OpenSSL-3.eb
• SUCCESS Perl-bundle-CPAN-5.36.1-GCCcore-12.3.0.eb
• SUCCESS gfbf-2023a.eb
• SUCCESS jiter-0.4.1-GCCcore-12.3.0.eb
• SUCCESS X11-20230603-GCCcore-12.3.0.eb

Build succeeded for 6 out of 6 (6 easyconfigs in total)
obelia - Linux Fedora Linux 41 (KDE Plasma), x86_64, 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz, Python 3.9.21
See https://gist.github.com/lexming/974c98d6234323edaa157681d887c484 for a full test report.

[lexming]

Checked the resulting module files and they have the same variables as before this PR.

[Thyre]

Might be also worth checking FlexiBLAS, where we ran into #2733 before #3509

[Thyre]

Testing one of the EasyConfigs from easybuilders/easybuild-easyconfigs#21582

Test report by @Thyre

Overview of tested easyconfigs (in order)

• SUCCESS intel-compilers-2025.0.0-CUDA-12.8.0.eb

Build succeeded for 1 out of 1 (1 easyconfigs in total)
Linux - Linux Arch Linux UNKNOWN, x86_64, AMD Ryzen 7 7800X3D 8-Core Processor, 1 x NVIDIA NVIDIA GeForce RTX 3070, 565.77, Python 3.13.1
See https://gist.github.com/Thyre/ea3d5e4bf30df680feabf4444567aae4 for a full test report.

[Thyre]

Test report by @Thyre

Overview of tested easyconfigs (in order)

• SUCCESS FlexiBLAS-3.4.4-GCC-13.3.0.eb

Build succeeded for 1 out of 1 (1 easyconfigs in total)
Linux - Linux Arch Linux UNKNOWN, x86_64, AMD Ryzen 7 7800X3D 8-Core Processor, 1 x NVIDIA NVIDIA GeForce RTX 3070, 565.77, Python 3.13.1
See https://gist.github.com/Thyre/66a519ac64dc1e4dfcffef30fccfebb0 for a full test report.

[Thyre]

Test report by @Thyre

Overview of tested easyconfigs (in order)

* **SUCCESS** _FlexiBLAS-3.4.4-GCC-13.3.0.eb_

Build succeeded for 1 out of 1 (1 easyconfigs in total) Linux - Linux Arch Linux UNKNOWN, x86_64, AMD Ryzen 7 7800X3D 8-Core Processor, 1 x NVIDIA NVIDIA GeForce RTX 3070, 565.77, Python 3.13.1 See https://gist.github.com/Thyre/66a519ac64dc1e4dfcffef30fccfebb0 for a full test report.

Module file paths looks good:

[...]
prepend_path("CMAKE_PREFIX_PATH", root)
prepend_path("CPATH", pathJoin(root, "include"))
prepend_path("CPATH", pathJoin(root, "include/flexiblas"))
prepend_path("LD_LIBRARY_PATH", pathJoin(root, "lib"))
prepend_path("LIBRARY_PATH", pathJoin(root, "lib"))
prepend_path("MANPATH", pathJoin(root, "share/man"))
prepend_path("PATH", pathJoin(root, "bin"))
prepend_path("PKG_CONFIG_PATH", pathJoin(root, "lib/pkgconfig"))
prepend_path("XDG_DATA_DIRS", pathJoin(root, "share"))
setenv("EBROOTFLEXIBLAS", root)
setenv("EBVERSIONFLEXIBLAS", "3.4.4")
setenv("EBDEVELFLEXIBLAS", pathJoin(root, "easybuild/FlexiBLAS-.3.4.4-GCC-13.3.0-easybuild-devel"))

-- Built with EasyBuild version 5.0.0beta1

[boegel]

lgtm