Created attachment 188244 [details] science/ALPSCore-* /!\ CAVEAT EMPTOR /!\ I am not really sure whether I should set CMAKE_*_COMPILER_LAUNCHER, because poudriere mentions symlinking ccache for the the compilers -- so mpi wrappers would already use it. Is that the case with direct build from ports, too? The open source ALPSCore project, which is based on ALPS (Algorithms and Libraries for Physics Simulations) project, provides a collection of physics libraries and applications, with a focus on simulations of lattice models and strongly correlated systems. This project strives to provide a convenient set of maintained, well-documented, and reusable components for developing condensed matter physics simulation code, to help application developers make commonly used and proven computational algorithms available to a non-expert community. WWW: https://github.com/ALPSCore/ALPSCore
Created attachment 188245 [details] science/ALPSCore-serial: portlint -AC
Created attachment 188246 [details] science/ALPSCore-openmpi2: portlint -AC
Created attachment 188247 [details] science/ALPSCore-serial: poudriere testport
Created attachment 188248 [details] science/ALPSCore-openmpi2: poudriere testport
Comment on attachment 188248 [details] science/ALPSCore-openmpi2: poudriere testport Whoops.
Created attachment 188251 [details] science/ALPSCore-openmpi2: poudriere testport
Thanks for your submission! In the future, you don't need to attach portlint/poudriere logs. It's sufficient to just say that they pass.
Hi, Thanks, will stop uploading. I vaguely remember them being required some time ago? Cheers, Mike
(In reply to Michael Danilov from comment #8) Opinions can vary, but I think that people who regularly submit ports are considered trustworthy, therefore there should be enough to just say it. I did this for a long time without problems. Committer has to rerun it anyways. Thanks! Yuri
I think, I will merge ALPSCore-openmpi2 into ALPSCore-serial, and make openmpi2/serial port options, since they can't be installed at the same time.
Buy I heard that with port options one can't have both available in the binary repo?
(In reply to Michael Danilov from comment #11) There is an upcoming 'package flavor' patch that will make possible to have both. I am not sure how exactly, but they are working on this. Yuri
Then go ahead. What about ccache? Should those args be passed to cmake?
(In reply to Michael Danilov from comment #13) ccache was broken when I tried it last time: https://github.com/ccache/ccache/issues/200 They need to fix it before we can use it.
(In reply to Michael Danilov from comment #13) I'll remove the ccache part. It seems this should be set by the framework in Mk/, not by individual ports. It only improves the speed of build, and not the final result.
(In reply to Yuri Victorovich from comment #15) Go ahead. If ccache is not used building without poudriere, the solution to the problem should belong to Mk.
Created attachment 188258 [details] Updated patch My current version.
Created attachment 188259 [details] Updated patch
Created attachment 188263 [details] Updated patch Normally build w/outs tests. Tests are build only for the test target. Disabled the documentation option because it is broken.
Comment on attachment 188263 [details] Updated patch Looks good, thank you!
Oh, and maybe there is a way to disable individual tests in CMake? Should I try to comment them out with a patchfile?
(In reply to Michael Danilov from comment #21) The current patch disables tests during build.
Yes, I'm asking if we could maybe keep the tests running and exclude the failing ones?
(In reply to Michael Danilov from comment #23) Ah, I see. Maybe, but that's the point of tests to show it when some are failing? IMO, all tests should run unless they are legitimately waived, like when they don't apply or something. Currently, one failure is unexpected, the rest are known failures. https://github.com/ALPSCore/ALPSCore/issues/388#issuecomment-346971059
Committed with changes. Thank you for your contribution!
A commit references this bug: Author: yuri Date: Sun Dec 3 19:29:13 UTC 2017 New revision: 455424 URL: https://svnweb.freebsd.org/changeset/ports/455424 Log: New port: science/ALPSCore: Generic algorithms and utilities for condensed matter physics PR: 223850 Submitted by: Michael Danilov <mike.d.ft402@gmail.com> Approved by: tcberner (mentor) Differential Revision: https://reviews.freebsd.org/D13237 Changes: head/science/ALPSCore/ head/science/ALPSCore/Makefile head/science/ALPSCore/distinfo head/science/ALPSCore/pkg-descr head/science/ALPSCore/pkg-plist head/science/Makefile