nghttp2 depends on python, and requires the install directory to exist,
and contain the python module directory which it will eventually
install the python modules into.
* trilinos: update hdf5 dependency to use 'hdf5+hl'
Without this change - I get the following error [triggered by 161dca6d5c9810f808 changes]
$ ./bin/spack install trilinos+hdf5
==> Error: An unsatisfiable variant constraint has been detected for spec:
hdf5@1.10.1%gcc@7~cxx~debug~fortran~hl+mpi+pic+shared~szip~threadsafe arch=linux-fedora27-x86_64
while trying to concretize the partial spec:
netcdf+mpi
^m4
^mpich@3.3a2%gcc@7 device=ch3 +hydra netmod=tcp +pmi+romio~verbs arch=linux-fedora27-x86_64
^zlib@1.2.11%gcc@7+optimize+pic+shared arch=linux-fedora27-x86_64
netcdf requires hdf5 variant +hl, but spec asked for ~hl
* fix flake8 warning
* added relion v 2.1
* relion 2.1 - misssed comma
* relion now pulling from master branch - removed 2.03 - added cuda 8 as dependency given 9 does not yet work with relion per https://github.com/3dem/relion/issues/296
* r-igraph: find old packages in the archive
Fixes#6104.
* r-irlba: find old packages in the archive
Fixes#6101.
* r-lazyeval: find old packages in the archive
Fixes#6102.
* packages: set list_url for all packages hosted on CRAN
* packages: consistently use the CRAN archive URLs
Aspell compiles in the paths it looks for rather than looking relative
to its binary location. There are also no environment variables to
redirect it other than changing the path to the configuration file it
looks at.
- hdf5's +hl variant is required because the build uses libhdf5_hl.
- sz is only compatible up to version 1.4.10 because the library has
been renamed to libSZ in 1.4.11.
* Fixed the datatype flag in the LBANN package.
Fixed where the datatype flag should be set.
* Added variant to OpenCV to disable building DNN support
* Fixed a bug in how the elemental package tries to find its Fortran
library. The current method does not work with Intel compilers.
* Added necessary paths for cuDNN and CUB.
* Addressed flake8 issues in Elemental package.
* Fixed several flake8 errors. Added variant to build with or without
support for arbitrary-precision floating point arithmetic.
* Added the proper CMake flag when MPFR support is disabled.
* Fixed a bug in how the cudnn path was being set when there was no
cudnn requested. Also requested that OpenCV is built without eigen
support.
* Fixed another bug where a dependency may not be there and thus should
not be always checked.
* Added a package for the MDAnalysis toolkit.
* Changed dependency for the eigen library to not include support for
multiple-precision (arbitrary-precision) floating point arithmetic,
since OpenCV does not support these data types.
* Fixed incorrect implementation
* dockbook: remove update to dependent environment
Remove override of setup_dependent_environment in both docbook-xml
and docbook-xsl, which seems to generate a warning from spack about
"Suspicious requests to set or unset 'XML_CATLOG_FILES' found."
* flux: update version/dependencies to 0.8.0
Update flux-core to v0.8.0. Clean up dependencies and remove unneeded
extra invocation of ./autogen.sh when building from @master.
* flux: modify environment for Lua tests
Modify Spack environment to add `./?.lua` to LUA_PATH so that
fluxometer.lua can be found by tests during `make check` run
from `spack install --tests all flux`
* flux: fix broken flux version when on master
When building @master with spack, the flux version test from
`make check` fails because spack is using a shallow clone. Add
a setup section for @master that calls git pull --depth=50 --tags
to pull down tags and (hopefully) enough information for git-describe
to work properly.
A recent update (#5907) to htslib added a different URL for an old
version of htslib.
Now the package is using that URL as the pattern for the newer
versions too.
I have a vague memory of running into this before, that it's a known
issue.
This fixes it by adding an explicit `url_for_version` routine.
* Add bowtie2@2.3.0, fix dependencies and sbangs
Add support for bowtie2@2.3.0
- digest
- a patch for 2.3.0 that parallels the existing package. Truth be
told it builds (for me) without this, but I'm assuming that they're
there for a reason...).
- tune up dependencies
- need tbb
- don't need readline or zlib
Several things were installed with sbang's that use `/usr/bin/env` to
fine perl or python. Fix the dependency and clean up the sbang lines.
* Fix python exe name, avoid path banging
I'd cut and pasted the python bit from the perl bit and missed one
reference to perl.
While I'm there, use the cleaner `spec['perl'].command` instead of
banging together the path from its bits.
* Fix up the "when" constraints on the dependencies
Get the edge cases right.
- 2.2.5 doesn't need tbb, 2.3.[01] do.
- 2.3.1 needs readline and zlib.
The go team released 1.9.2 which includes fixes for some things
that 1.9.1 broke:
> ... include fixes to the compiler, linker, runtime, documentation, go command, and the crypto/x509, database/sql, log, and net/smtp packages. They include a fix to a bug introduced in Go 1.9.1 and Go 1.8.4 that broke "go get" of non-Git repositories under certain conditions.
* Exodus: skip the -G "Unix Makefiles" part
The problem is that spack passes -G "Unix Makefiles" into cmake, which normally
works. But in the Exodus package, it is being passed into a bash wrapper
script. In there, the $@ then loses the information about "Unix Makefiles"
being just one argument, and in effect passes -G Unix Makefiles into the cmake
(without quotes), and so cmake only sees -G Unix, and then fails. This is a
known problem with bash with no simple solutions. As a workaround, this patch
skips the first two arguments, i.e., -G and "Unix Makefiles". This makes it
work.
Fixes#5895.
* Port exodusii to cmake
The cmake options were taken from the cmake-exodus bash script and ported to
spack directly.
* Use variant forwarding to forward the 'mpi'
Now instead of
spack install exodusii~mpi^netcdf~mpi^hdf5~mpi
one can just use
spack install exodusii~mpi
* sw4lite: fix build errors and add variants
* sw4lite: change linking against blas and lapack
* change order of blas and lapack
* satisfy flake8 requirements
* Update package.py
* Add the custom paraview lib directory structure to the library paths in the paraview module file.
* Fixing flake8 issues.
* Checking if lib64 exists for paraview module file generation, else use lib.
* Fixing more flake8 problems I introduced.
Since LLVM 3.9 Clang can use the libc++ library by default using the
CLANG_DEFAULT_CXX_STDLIB cmake configuration variable, without having to
specify the -stdlib=libc++ option on the clang++ command line.
This commit makes clang++ use libc++ by default for LLVM 3.9 and later if the
libcxx variant is on.
Fixes#5942.
Chasing a performance regression has lead me to this change, going from default optimization gives a significant performance win. The sweet spot for zlib is apparently `-O2`, both `-Ofast` and `-O3` are slightly worse (regression is about 3% compared with `-O2) in my testing.
Happy to share my methodology with people so we can benchmark on a wider variety of systems.
* Add package for scalpel@0.5.3
Scalpel's a bit of a mess, it expects it's users to just unpack the
tarball, build it in the resulting directory and install that
directory onto their PATH. My install step recapitulates this into
prefix.bin. The alternative was rewiring their scripts (perl), which
use `FindBin` and expect things to be located in the same dir that the
script itself is.
Sigh.
Lightly tested on CentOS 7.
* Flake8 cleanup
* Additional flake8 cleanup
* Added procedure to edit sbangs of the parallel perl scripts.
* Specify the types of perl dependency
Adding ", type=('build', 'run')" to the dependency declaration to clarify when and how perl is required
* flake8 cleanup
The problem was that the configure script was not using spack's compiler
wrappers. We now pass the proper compiler wrapper using the CC argument
explicitly.
Fixes#5892.
