- configuration fails on newer versions when enabling sse/avx/avx2
- prefer an older version that will build with these features

* cpu: differentiate flags used for pristine LLVM vs. Apple's version

* cuda: fix conflict statements for x86-64 targets

fixes #13462

This build system mixin was not updated after the support for specific
targets has been merged.

* Updated the version range of cuda that conflicts with gcc@8:

* Updated the version range of cuda that conflicts with gcc@8: for ppc64le

* Relaxed conflicts for version > 10.1

* Updated versions in conflicts

Co-Authored-By: Axel Huebl <[email protected]>

Currently, query arguments in the Spack core are documented on the
Database._query method, where the functionality is defined.

For users of the spack python command, this makes the python builtin
method help less than ideally useful, as help(spack.store.db.query)
and help(spack.store.db.query_local) do not show relevant information.

This PR updates the doc attributes for the Database.query and
Database.query_local arguments to mirror everything after the first
line of the Database._query docstring.

* Travis CI: Test Python 3.8

* Fix use of deprecated cgi.escape method

* Fix version comparison

* Fix flake8 F811 change in Python 3.8

* Make flake8 happy

* Use Python 3.8 for all test categories

* Make package preferences a soft failure for targets, instead of a hard failure.
* Added unit tests for preferences expressed via packages.yaml

This fixes a UnicodeDecodeError in the sbang patching function.

…13488)

* fix issue where spack.util.url.join() failed to correctly handle absolute path components
* add url util tests

This changes Spack environments so that the YAML file associated with the environment is *only* written when necessary (i.e., if it is changed *by spack*).  The lockfile is still written out as before.

There is a larger question here of which part of Spack should be responsible for setting defaults in config files, and how we can get rid of empty lists and data structures currently cluttering files like `compilers.yaml`.  But that probably requires a rework of the default-setting validator in `spack.config`, as well as the code that uses `spack.config`.  This will at least help for `spack.yaml`.

…13545)

- [x] insert at beginning of list so fetch grabs local mirrors before remote resources
- [x] update the S3FetchStrategy so that it throws a SpackError if the fetch fails.  
      Before, it was throwing URLError, which was not being caught in stage.py.
- [x] move error handling out of S3FetchStrategy and into web_util.read_from_url()
- [x] pass string instead of URLError to SpackWebError

#13578)

* Make relative binaries relocate text files properly
* rb strings aren't valid in python 2
* move perl to new interface for setup_environment family methods

* environments: make shell modifications partially unconditional

* flake

* missing module name

* add regression test

* flake

This fixes a regression introduced in #10792.  `spack uninstall` in an
environment would not match concrete query specs properly after the index
hash of enviroments changed.

- [x] Search by DAG hash for specs to remove instead of by build hash

If you do this in a spack environment:

    spack add hdf5+hl

hdf5+hl will be the root added to the `spack.yaml` file, and you should
really expect `hdf5+hl` to display as a root in the environment.

- [x] Add decoration to roots so that you can see the details about what
      is required to build.
- [x] Add a test.

Binaries with relative RPATHS currently do not relocate strings
hard-coded in binaries

This PR extends the best-effort relocation of strings hard-coded
in binaries to those whose RPATHs have been relativized.

…link path (which includes the file name) (#13727)

Add a configuration option to suppress gpg warnings during binary
package verification. This only suppresses warnings: a gpg failure
will still fail the install. This allows users who have already
explicitly trusted the gpg key they are using to avoid seeing
repeated warnings that it is self-signed.

…ch (#13734)

* remove reference to `spack.store` in method definition

Referencing `spack.store` in method definition will cache the `spack.config.config` singleton variable too early, before we have a chance to add command line and environment scopes.

* votca: help concretizer

* flake8

* Add version check when building with clang

* Update for flake8 formatting

`@run_after` and `@when` are incompatible, issue #12736.

* Add elfutils-0.178

* Disable debuginfod server

* Fix flake8

* Fixed x86-64 optimization flags for clang
* Fixed expected results in unit tests

Before the flags used where the one for llc, the underlying compiler from LLVM IR to machine assembly. It turns out that the semantic of `-march`, `-mtune` and `-mcpu` changes from clang front-end to llc.

I found no definitive reference for the flags submitted in this PR, but I checked the assembly on a vectorizable function using Godbolt's web-site.

`spack module loads` and `spack module find` previously failed if any upstream modules were missing.  This prevented it from being used with upstreams (or, really, any spack instance) that blacklisted modules.

This PR makes module finding is now more lenient (especially for blacklisted modules).

- `spack module find` now does not report an error if the spec is blacklisted
  - instead, it prints a single warning if any modules will be omitted from the loads file
  - It comments the missing modules out of the loads file so the user can see what's missing
  - Debug messages are also printed so users can check this with `spack -d...`

- also added tests for new functionality

Keep CUDA 10.1.243 as the preferred version until the issue
of including implementation details of libcu++ is addressed.

`spack module loads` and `spack module find` previously failed if any upstream modules were missing.  This prevented it from being used with upstreams (or, really, any spack instance) that blacklisted modules.

This PR makes module finding is now more lenient (especially for blacklisted modules).

- `spack module find` now does not report an error if the spec is blacklisted
  - instead, it prints a single warning if any modules will be omitted from the loads file
  - It comments the missing modules out of the loads file so the user can see what's missing
  - Debug messages are also printed so users can check this with `spack -d...`

- also added tests for new functionality

v0.13.2

This release contains major performance improvements for Spack environments, as well as some bugfixes and minor changes.

* allow missing modules if they are blacklisted (#13540)
* speed up environment activation (#13557)
* mirror path works for unknown versions (#13626)
* environments: don't try to modify run-env if a spec is not installed (#13589)
* use semicolons instead of newlines in module/python command (#13904)
* verify.py: os.path.exists exception handling (#13656)
* Document use of the maintainers field (#13479)
* bugfix with config caching (#13755)
* hwloc: added 'master' version pointing at the HEAD of the master branch (#13734)
* config option to allow gpg warning suppression (#13744)
* fix for relative symlinks when relocating binary packages (#13727)
* allow binary relocation of strings in relative binaries (#13724)

* apply strict constraint checks for patches, otherwise Spack may incorrectly treat a version range constraint as satisfied when mixing x.y and x.y.z versions
* add mixed version checks to version comparison tests

* New package: swipl

* Removed boilerplate, added maintainer

* Fixed python 2 incompatibility

* Update var/spack/repos/builtin/packages/swipl/package.py

Co-Authored-By: Adam J. Stewart <[email protected]>

* swipl: Added conflict with intel compiler

py-matplotlib: No constraint necessary for py3 but py2 has an upper bound
py-numpy: No constraint necessary for py3 but py2 has an upper bound

* add tensorflow

Change-Id: Id778c68d148cc42f0b478a9d10a8f937cb54cdc6

* make bazel and tensorflow build

Change-Id: Iae9005e8f4dcc8f1ed36ea9337d2430aeebb291f

* fix flake8

Change-Id: Ib05529dd796eab4a8855a5d7775cc4efea8e479d

* 2nd flake8 attempt

Change-Id: I46224be3a374b2a65793048b0c5178ea64adbd78

* replace md5 sums with sha256

* add version 1.13.2

* bazel() -> bazel('build',...

* specify versions of bazel dependency

* build with CUDA

* add TODOs

* add more todo"s

* improve enum34 dependency

* py-future is a dependency as of v1.14

* Update var/spack/repos/builtin/packages/tensorflow/package.py

Co-Authored-By: Adam J. Stewart <[email protected]>

* Update var/spack/repos/builtin/packages/tensorflow/package.py

Co-Authored-By: Adam J. Stewart <[email protected]>

* Update var/spack/repos/builtin/packages/tensorflow/package.py

Co-Authored-By: Adam J. Stewart <[email protected]>

* Update var/spack/repos/builtin/packages/tensorflow/package.py

Co-Authored-By: Adam J. Stewart <[email protected]>

* enable nccl, cuda by default

* explain patches

* add todo

* remove unnecessary copt_flag

* use join

* join argument must be an iterable

* split long line; use same opts for non-cuda build

* without opt flags, configure hangs

* introduce build phases; re-arrange

* undo mistake

* restore unset tmp_path

* as of v1.14, nccl_install_path is parsed correctly, hence change ...prefix.lib to ...prefix

* now, version 1.14 compiles successfully with cuda

* add version 2.1.0

* specify bazel dependency for version 2.1.0-rc0

* account for deprecated bazel opts for v2.1.0-rc0

* disable mkldnn contraction kernel

* Flake8 fixes

* md5 -> sha256

* Fix TF and TF-estimator version deps

* Don't just comment out patch

* Add myself as a maintainer

* Patch py-astor to support newer py-setuptools

* Add more versions and bazel version constraints

* Add a build phase

* Add note about configure interactivity

* dev-build -> build-env

* Disable iOS build

* Use correct optimization flags

* Add variants for all possible features

* nccl isn't always a dependency

* Specify correct dependency versions for each release

* Libs may not be in lib or lib64

* Add py-opt-einsum package

* Add newer version of py-protobuf

* Add newer version of py-wrapt

* Fix Python 2.6 syntax error

* Code review

* Set more env vars for older versions

* Add more env vars, fix bazel versions, add conflicts

* Fix config options

* Specify version that support --config args

* Add py-future dependency for Python 2

* Fix cuda config flag and compute capabilities

* Fix installation on macOS, add unit tests

* Override cuda variant default to True on non-macOS

* Rename tensorflow to py-tensorflow

* Has to extend something

* Fix os.symlink call

* convert cuda_arc values to capabilities

* restore nccl prefix path for v1.13.1

* Revert to v2

* Remove extraneous period

* Add new version of jdk/openjdk

* More stable cuda_arch formatting

* Fix bazel unit tests

* Fix symlinking

* Fix unit tests

* +gcp by default until build error figured out

Add version 2019.09.  Adjust the iteration through versions to
eliminate the need for version_list.

* Minimal BundlePackage build system doc

* Add link to new bundlepackage file

* Fixed link bug and added create command example

* Add h5sh package

* Fix requirements and update version

* [WIP] Convert llvm spackage to use the monorepo

* Correcting python dependencies

* Adding develop version for llvm monorepo

* Python 2.6 Fix

* Build Flang fork in a different root_cmakelists_dir

* Formatting Fix

* Including flang package changes

* Removing explicit llvm dependencies variants

* flake8 fix

* Updating patches and llvm+flang logic

* Updating env setup API

* Add top level git

* Conflicting flang variant with other variants