summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2015-12-19Merge pull request #226 from epfl-scitas/packages/mpichesTodd Gamblin2-5/+38
mpich, mvapich2: changes the filter for the wrappers mpi{cc, cxx, f77, fc}
2015-12-19Merge pull request #227 from trws/fetch-depsTodd Gamblin4-4/+25
fetch: add options to fetch missing or all deps
2015-12-19Merge pull request #230 from nolta/fix-env-mod-sys-typeTodd Gamblin2-6/+12
fix environment module path when $SYS_TYPE isn't defined
2015-12-19Merge branch 'pull/256' into developTodd Gamblin5-11/+25
2015-12-19Fix empty format strings in some packages.Todd Gamblin3-8/+6
2015-12-19Fix stale code in python version test so that it actually works.Todd Gamblin2-5/+2
2015-12-19new package: cfitsioMike Nolta1-0/+19
2015-12-19fix env module sys_type on non-LLNL systemsMike Nolta1-4/+4
2015-12-19spack python: add -c optionMike Nolta1-2/+8
Allows passing program in as a string. Example: $ spack python -c 'print 2+3' 5 Also imports spack module by default into the environment.
2015-12-19Merge pull request #253 from LLNL/bugfix/244-uninstall-errorsGregory Lee5-13/+37
Bugfix/244 uninstall errors
2015-12-19Merge branch 'develop' of https://github.com/llnl/spack into developGregory L. Lee1-3/+3
2015-12-19added mrnet 5.0.1Gregory L. Lee1-0/+1
2015-12-18Merge pull request #252 from JohnWestlund/developTodd Gamblin1-3/+3
remove possible trailing slash in _sp_prefix or _sp_share_dir
2015-12-18Fix #244: errors on uninstallTodd Gamblin2-9/+23
- Extension logic didn't take conditional deps into account. - Extension methods now check for whether the extnesion is in the extendee map AND whether the dependency is actually present in the spec yet.
2015-12-18minor formatting.Todd Gamblin1-0/+8
2015-12-18Fix import error.Todd Gamblin1-0/+1
2015-12-18remove possible trailing slash in _sp_prefix or _sp_share_dirJohn Westlund1-3/+3
2015-12-18Format to 80 char width.Todd Gamblin1-4/+5
2015-12-17Merge pull request #236 from nolta/cmake-enable-sslTodd Gamblin1-1/+2
cmake: enable ssl, for https downloads
2015-12-17Merge pull request #205 from trws/shell-arg-handling-fixTodd Gamblin1-1/+7
fix for the shell arg handling issues
2015-12-17Merge pull request #250 from davidbeckingsale/features/lmodTodd Gamblin1-0/+26
Added Lmod package
2015-12-17Added Lmod packageDavid Beckingsale1-0/+26
2015-12-17Merge pull request #86 from jgalarowicz/developTodd Gamblin17-11/+1240
Add support for Krell Institute related products.
2015-12-17Change back the previous boost change and change the mrnet variant name to ↵Jim Galarowicz7-17/+16
lwthreads.
2015-12-16Merge pull request #246 from LLNL/features/245-pdb-debugTodd Gamblin1-0/+5
added option to enable pdb debug
2015-12-16added option to enable pdb debugGregory L. Lee1-0/+5
2015-12-16Merge pull request #167 from mathstuf/hdf5-fix-mpi-dependencyTodd Gamblin5-7/+17
Hdf5 fix mpi dependency
2015-12-16Merge pull request #239 from LLNL/bugfix/238-db-specs-not-concreteGregory Lee3-4/+17
Fix #238: Database specs need to be concrete when read in as well.
2015-12-16Merge pull request #241 from epfl-scitas/packages/gnuplotTodd Gamblin7-1/+174
gnuplot : added package
2015-12-16Merge pull request #242 from epfl-scitas/packages/gslTodd Gamblin1-0/+46
gsl : added package
2015-12-16Merge pull request #243 from nolta/patch-2Todd Gamblin1-2/+2
minor comment tweak in modules.py
2015-12-16minor comment tweak in modules.pyMike Nolta1-2/+2
2015-12-16gsl : added packagealalazo1-0/+46
2015-12-16gnuplot : modifications to build on RHLE 6Massimiliano Culpo3-2/+7
2015-12-15Fix #238: Database specs need to be concrete when read in as well.Todd Gamblin3-4/+17
2015-12-15cmake: enable ssl, for https downloadsMike Nolta1-1/+2
By default cmake builds its own curl, without SSL support. This patch enables SSL when building cmake, fixing the following error: error: downloading 'https://...' failed status_code: 1 status_string: "Unsupported protocol" log: Protocol "https" not supported or disabled in libcurl
2015-12-15gnuplot : added packagealalazo6-0/+168
2015-12-14Merge pull request #233 from LLNL/bugfix/206-value-errorTodd Gamblin1-2/+4
Fix #206: need to make deps AND root concrete when read in.
2015-12-14Fix #206: need to make deps AND root concrete when read in.Todd Gamblin1-2/+4
2015-12-14Merge pull request #232 from LLNL/bugfix/package-cache-217Todd Gamblin3-16/+26
Bugfix/package cache 217
2015-12-14Merge pull request #229 from epfl-scitas/packages/fftwTodd Gamblin1-0/+96
fftw : added package
2015-12-14fftw : added packagealalazo1-0/+96
2015-12-13Fix build of krell tools to use spack with minimal impact on rest of spack ↵Jim Galarowicz7-66/+52
packages.
2015-12-11Fix merge conflict with papi package.Jim Galarowicz240-423/+1169
2015-12-11Make internal hash dep sort order match external one.Todd Gamblin1-1/+2
2015-12-11Fix #217: update spec_dag test for new `_cmp_key`.Todd Gamblin1-3/+5
2015-12-11Fix #217: Use MUCH faster hashing, reduce number of DAG copies.Todd Gamblin2-12/+18
This changes the hash algorithm so that it does much less object allocation and copying, and so that it is correct. The old version of `_cmp_key()` would call `sorted_deps`, which would call `flat_dependencies` to get a list of dependencies so that it could sort them in alphabetical order. This isn't necessary in the `_cmp_key()`, and in fact we want more DAG structure than that to be included in the `_cmp_key()`. The new version constructs a tuple without copying the Spec DAG, and the tuple contains hashes of sub-DAGs that are computed recursively in-place. This is way faster than the previous algorithm and reduces the numebr of copies significantly. It is also a correct DAG hash. Example timing and copy counts for the different hashing algorithms we've tried: Original (wrong) Spec hash: ``` 106,170 copies real 0m5.024s user 0m4.949s sys 0m0.104s ``` Spec hash using YAML `dag_hash()`: ``` 3,794 copies real 0m5.024s user 0m4.949s sys 0m0.104s New no-copy, no-YAML hash: ``` 3,594 copies real 0m2.543s user 0m2.435s sys 0m0.104s ``` So now we have a hash that is correct AND faster. The remaining ~3k copies happen mostly during concretization, and as all packages are initially loaded. I believe this is because Spack currently has to load all packages to figure out virtual dependency information; it could also be becasue there ar a lot of lookups of partial specs in concretize. I can investigate this further.
2015-12-11Fix #217: Make package cache use DAG hash instead of sorted deps.Todd Gamblin1-7/+8
- Gets rid of last vestige of old-style specs. - Uses new hashing for lookup
2015-12-10fix path resolution for mirror packages, especially with dependency fetchingTom Scogland2-1/+2
2015-12-10hdf5: update dependents to declare MPI supportBen Boeckel4-3/+4
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2024-09-19 09:49:49 +0000