Skip to content

Conversation

@jsquyres
Copy link
Member

One reason that we make compile-time sym links is because waaaaay back in olden days, SVN couldn't track sym links. So we made sym links at compile time and avoided the problem.

Per @siegmargross' usage, however, perhaps we should not make compile-time sym links any more. Instead, just:

  • git commit sym links to one directory back
  • remove the compile-time logic to make sym links
  • update .gitignore

The "make dist" tarballs will actually contain copies of files (not sym links), but that's ok -- the sym link is actually useful to developers, not end users.

jsquyres added 3 commits June 10, 2015 14:39
Instead of creating sym links at run time, just commit sym links into
the git repo.  During "make dist", the sym link will get copied,
meaning that an actual file will end up in the tarball (not a sym
link) -- but that's ok.  The sym link is mainly for developer
convenience -- so that we can edit a single source file and have it
compiled as both MPI_<foo> and PMPI_<foo>.
Similar to the previous commit, commit sym links directly in git for
all the profiled versions of the mpif.h Fortran interfaces.
Similar to the previous commit, commit sym links directly in git for
all the profiled versions of the MPI_T interfaces.
@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_distcheck/38/
Test PASSed.

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_cle5.2up02/130/
Test PASSed.

@mellanox-github
Copy link

@mellanox-github
Copy link

Refer to this link for build results (access rights to CI server needed):
http://bgate.mellanox.com/job/gh-ompi-master-pr/620/

jsquyres added 4 commits June 10, 2015 18:41
Similae to the previous commit, commit sym link directly in git
for mpi_portable_platform.h.
Similar to the previous commit, commit sym links directly in git for
all the profiled versions of the OSHMEM C interfaces.
Similar to the previous commit, commit sym links directly in git for
all the profiled versions of the OSHMEM Fortran interfaces.
@jsquyres jsquyres force-pushed the pr/remove-build-time-sym-links branch from 8825711 to c772bb2 Compare June 11, 2015 01:41
@mellanox-github
Copy link

@mellanox-github
Copy link

Refer to this link for build results (access rights to CI server needed):
http://bgate.mellanox.com/job/gh-ompi-master-pr/623/

@hppritcha
Copy link
Member

bot:retest

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_distcheck/40/
Test PASSed.

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_cle5.2up02/133/
Test PASSed.

@jsquyres
Copy link
Member Author

bot:retest

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_distcheck/43/
Test PASSed.

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_cle5.2up02/136/
Test PASSed.

@mellanox-github
Copy link

@mellanox-github
Copy link

Refer to this link for build results (access rights to CI server needed):
http://bgate.mellanox.com/job/gh-ompi-master-pr/625/

@hppritcha
Copy link
Member

not sure why the LANL pull request builder's status didn't show up. Trying again.
bot:retest

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_cle5.2up02/137/
Test PASSed.

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_distcheck/44/
Test PASSed.

@hppritcha
Copy link
Member

bot:retest

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_cle5.2up02/138/
Test PASSed.

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_distcheck/45/
Test PASSed.

@mellanox-github
Copy link

@mellanox-github
Copy link

Refer to this link for build results (access rights to CI server needed):
http://bgate.mellanox.com/job/gh-ompi-master-pr/626/

@jsquyres
Copy link
Member Author

I talked this over with @goodell -- it doesn't seem like this PR has any real positive value.

Unless someone has a strong reason, I'm likely to throw this PR / branch away.

@hppritcha
Copy link
Member

has small positive value right now because it seems to bring out some issue with jenkins and/or github and or something. One more time with this.

bot:retest

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_distcheck/46/
Test PASSed.

@lanl-ompi
Copy link
Contributor

Refer to this link for build results (access rights to CI server needed):
http://jenkins.open-mpi.org/job/ompi_master_pr_cle5.2up02/139/
Test PASSed.

@jsquyres jsquyres closed this Jun 15, 2015
@jsquyres jsquyres deleted the pr/remove-build-time-sym-links branch June 15, 2015 21:10
jsquyres pushed a commit to jsquyres/ompi that referenced this pull request Nov 10, 2015
…ze_flag_and_fca_use_it

v2.x - Add in-finalize indicator, fca fall back to prev barrier if in-finalize
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants