* Module index should not be unconditionally overwritten
Uncovered after we switched our CI to generate modules for packages
one-by-one rather than in bulk. This overwrote a complete module index
with an index with a single entry, and broke our downstream Spack
instances that needed the upstream module index.
* Changed the 'include' config section to use 'substitute_path_variables' to allow for Spack config variables to be used (e.g. $spack).
* Fixed a bug with 'include' section path expansion and added a test case for 'include' paths with embedded config variables.
* Cray: fix Blue Waters support
* pkg-config env vars needed on Blue Waters
* cray platform: fix support for user-build MPI on cray machines
* reintroduce cray environment cleaning behind cnl version guard
* cray platform: fix support for user-build MPI on cray machines
Co-authored-by: Gregory <becker33@llnl.gov>
Builds can be stopped before the final install phase due to user requests. Those builds
should not be registered as installed in the database.
We had code intended to handle this but:
1. It caught the wrong type of exception
2. We were catching these exceptions to suppress them at a lower level in the stack
This PR allows the StopIteration to propagate through a ChildError, and catches it
properly. Also added to an existing test to prevent regression.
This fixes a fork bomb in `spack versions`. Recursive generation of pools
to scrape URLs in `_spider` was creating large numbers of processes.
Instead of recursively creating process pools, we now use a single
`ThreadPool` with a concurrency limit.
More on the issue: having ~10 users running at the same time spack
versions on front-end nodes caused kernel lockup due to the high number
of sockets opened (sys-admin reports ~210k distributed over 3 nodes).
Users were internal, so they had ulimit -n set to ~70k.
The forking behavior could be observed by just running:
$ spack versions boost
and checking the number of processes spawned. Number of processes
per se was not the issue, but each one of them opens a socket
which can stress `iptables`.
In the original issue the kernel watchdog was reporting:
Message from syslogd@login03 at May 19 12:01:30 ...
kernel:Watchdog CPU:110 Hard LOCKUP
Message from syslogd@login03 at May 19 12:01:31 ...
kernel:watchdog: BUG: soft lockup - CPU#110 stuck for 23s! [python3:2756]
Message from syslogd@login03 at May 19 12:01:31 ...
kernel:watchdog: BUG: soft lockup - CPU#94 stuck for 22s! [iptables:5603]
* add an --exclude-file option to 'spack mirror create' which allows a user to specify a file of specs to exclude when creating a mirror. this is anticipated to be useful especially when using the '--all' option
* allow specifying number of versions when mirroring all packages
* when mirroring all specs within an environment, include dependencies of root specs
* add '--exclude-specs' option to allow user to specify that specs should be excluded on the command line
* add test for excluding specs
fixes#12527
Mention that specs can be uninstalled by hash also in
the help message. Reference `spack gc` in case people
are looking for ways to clean the store from build time
dependencies.
Use "spec" instead of "package" to avoid ambiguity in
the error message.
* Unify tests for compiler command in the same file
Tests for the "spack compiler" command were previously
scattered among different files.
* Tests should use mutable_config, since they modify the compiler list
Because of the way abstract variants are implemented, the following
spec matrix does not work as intended:
```
matrix:
- [foo]
- [bar=a, bar=b]
exclude:
- bar=a
```
because abstract variants always satisfy any variant of the same
name, regardless of values.
This PR converts abstract variants to whatever their appropriate
type is before running satisfaction checks for the excludes clause
in a matrix.
fixes#16841
Now that the version number of GCC reached double digits, an update
to the regex is needed to recognize gcc-10 as an executable to be
inspected when searching for compilers.
* make_link_relative: added docstring
* make_elf_binaries_relative: added docstring, unit tests
* raise_if_not_relocatable: added docstring, added unit test for exceptional case
* relocate_links: removed unused arguments, added docstring and comments
Also fixed a possible bug that was issuing spurious
warning when a file was relocated successfully
* relocate_text: added docstring and comments, renamed arguments
* relocate_text_bin: added docstring and comments, renamed arguments, unit tests
Problem: when calling `static_to_shared_library` on the `cray` arch, it
produces a non-sensical compiler command with no input files. For
example, when installing lua@5.2.4, it produced:
'gcc -lm -ldl -o /big-long-spack-path/liblua.so.5.2.4'
Solution: do the same thing on `cray` that is done for `linux`
* account for schema validation errors where the associated instance doesn't have a line number
* fix unrelated flake error (but it must be fixed because this PR touches this file and the flake rules have been updated since the last edit to this file)
Allows `all` to be configured non-buildable in packages.yaml.
The following config would only allow zlib to be built by Spack, all other packages would have to be found as externals.
```
packages:
all:
buildable: False
zlib:
buildable: True
```
This change also adds a code path through the spack ci pipelines
infrastructure which supports PR testing on the Spack repository.
Gitlab pipelines run as a result of a PR (either creation or pushing
to a PR branch) will only verify that the packages in the environment
build without error. When the PR branch is merged to develop,
another pipeline will run which results in the generated binaries
getting pushed to the binary mirror.
Providing only $padding or ${padding} results in an attempt to
substitute a padding of maximum system path length, while leaving
room for the parts of the install path spack generates. Providing
$padding-<len> or ${padding-<len>} simply substitutes padding of
the specified length.
Packages built with lmod core_compiler are placed in `Core`.
Other packages may belong in `Core`. For example, python may be built with a proprietary compiler for performance, but belong on the `Core` directory.
With this PR, lmod config can include a `core_specs` list. Any package that satisfies a spec in that list is placed in `Core`, regardless of its compiler or dependencies.
This improves the documentation for `spack external find` in several ways:
* Provide a code example of implementing `determine_spec_details` for a package
* Explain how to define executables to look for (and also e.g. that they are treated as regular expressions and so can pull in unexpected files).
* Add the "why" for a couple of constraints (i.e. explain that this logic only works for build/run deps because it examines `PATH` for executables)
* Spread the docs between build customization and packaging sections
* Add cross-references
* Add a label so that `spack external find` is linked from the command reference.
* Add pmi support (required by ucx, ofi, and gni backends)
* Add support for ucx backend
* Add dependency on MPI for pmi=simplepmi, slurmpmi, or slurmpmi2
* Remove charmpp as an MPI provider since the changes in this PR can
add MPI as a dependency (mentioned previously)
* Install into transport_protocol-OS-arch subdirectory to match
default charmpp installation behavior (which helps dependents find it)
- add docstrings and make parameter names consistent in `relocate.py`
- Make `replace_prefix_*` and other functions private (as they are implementation details)
- remove unused function _replace_prefix_nullterm()
- Add unit tests for `relocate.py` functions
- add patchelf to Travis and use it during tests
- add hello_world fixture with a compiled binary, so we can test relocation
After migrating to `travis-ci.com`, we saw I/O issues in our tests --
tests that relied on `capfd` and `capsys` were failing. We've also seen
this in GitHub actions, and it's kept us from switching to them so far.
Turns out that the issue is that using streams like `sys.stdout` as
default arguments doesn't play well with `pytest` and output redirection,
as `pytest` changes the values of `sys.stdout` and `sys.stderr`. if these
values are evaluated before output redirection (as they are when used as
default arg values), output won't be captured properly later.
- [x] replace all stream default arg values with `None`, and only assign stream
values inside functions.
- [x] fix tests we didn't notice were relying on this erroneous behavior
This adds the `url` alternative `urls` to `package.all_urls`. With
this addition, one can find again new versions with
`spack versions <package>` for packages that are populated with
from mixin mirror `urls`.
Example: `util-macros` from x.org mixin.
* Non-interactive mode for spack checksum; allow passing 'package@version' to spack checksum
* Flake8 fixes
* Update checksum.py
Fix typo
* Update spack-completion script
* Automatically set non-interactive mode if more than one version passed
* Update lib/spack/spack/cmd/checksum.py
Co-Authored-By: Adam J. Stewart <ajstewart426@gmail.com>
* Add documentation and update spack-completion
* Flake8
* Rename option
* Update spack-completion
* Update lib/spack/spack/cmd/checksum.py
Co-Authored-By: Adam J. Stewart <ajstewart426@gmail.com>
* Update checksum.py
* Update stage.py
* Update create.py
Use batch mode when adding a new package
Co-authored-by: Ivan Razumov <ivan.razumov@cern.ch>
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
This fixes some errors with setting up test configuration. These
errors do not cause current Spack tests to fail but do create
red herring issues elsewhere (see #15666). Fixing these errors
leads to more errors in tests that depended on the original
misconfigured state, so those are also addressed here.
This is an update to #16003 which accounts for some unit tests with
conflicting config/mutable_config fixtures. These conflicts were
not exposed until the mutable_config fixture was fixed. Details are
included below. The change which builds on #16003 is prefixed with
"(new)".
* For tests that use the real Spack package repository, the config
needs to avoid using MPI providers that are not intended to be
installed by Spack. Without this, it is possible that Spack tests
which concretize the MPI virtual will end up trying to use an
implementation that it shouldn't (e.g. one that is always
provided externally). See #15666 for an example.
* The mutable_config test fixture was not initializing the scope
roots to the right directories (so the resulting config was empty).
* The current_host fixture in the concretize.py tests was using the
config fixture rather than mutable_config, and was polluting the
config cache for other tests.
* One test in concretize.py was clearing a nonexistent cache
(PackagePrefs._packages_config_cache). This reference has been
removed.
* The test 'test_preferred_compilers' was was depending on cross
test config pollution to succeed. The initial spec before
concretization has been updated to updated to be explicit about
the desired result.
* (new) For tests that use install_mockery and mutable_config,
replace install_mockery with a separate install_mockery_mutable_config
fixture that is exactly the same as install_mockery but uses the
mutable_config fixture to avoid conflicts.
Fixed#15884.
Spack asks every package linked into an environment to tell us how
environment variables should be modified when a spack environment is
activated. As part of this, specs in an environment are symlinked into
the environment's view (see #13249), and the package calculates
environment modifications with *the default view as the prefix*.
All of this works nicely for pointing the user's environment at the view
*if* every package is successfully linked. Unfortunately, right now we
only track what specs "should" be in a view, not which specs actually
are. So we end up calculating environment modifications on things that
aren't linked into thee view, and the exception isn't caught, so lots of
spack commands end up failing.
This fixes the issue by ignoring and warning about specs where
calculating environment modifications fails. So we can still keep using
Spack even if the current environment is incomplete.
We should probably also just avoid computing env modifications *entirely*
for unlinked packages, but right now that is a slow operation (requires a
lot of YAML parsing). We should revisit that when we have some better
state management for views, but the fix adopted here will still be
necessary, as we want spack commands to be resilient to other types of
bugs in `setup_run_environment()` and friends. That code is in packages
and we have to assume it could be buggy when we call it outside of builds
(as it might fail more than just the build).
Add a `spack external find` command that tries to populate
`packages.yaml` with external packages from the user's `$PATH`. This
focuses on finding build dependencies. Currently, support has only been
added for `cmake`.
For a package to be discoverable with `spack external find`, it must define:
* an `executables` class attribute containing a list of
regular expressions that match executable names.
* a `determine_spec_details(prefix, specs_in_prefix)` method
Spack will call `determine_spec_details()` once for each prefix where
executables are found, passing in the path to the prefix and the path to
all found executables. The package is responsible for invoking the
executables and figuring out what type of installation(s) are in the
prefix, and returning one or more specs (each with version, variants or
whatever else the user decides to include in the spec).
The found specs and prefixes will be added to the user's `packages.yaml`
file. Providing the `--not-buildable` option will mark all generated
entries in `packages.yaml` as `buildable: False`
Cray has two machine types. "XC" machines are the larger
machines more common in HPC, but "Cluster" machines are
also cropping up at some HPC sites. Cluster machines run
a slightly different form of the CrayPE programming environment,
and often come without default modules loaded. Cluster
machines also run different versions of some software, and run
a linux distro on the backend nodes instead of running Compute
Node Linux (CNL).
Below are the changes made to support "Cluster" machines in
Spack. Some of these changes are semi-related general upkeep
of the cray platform.
* cray platform: detect properly after module purge
* cray platform: support machines running OSs other than CNL
Make Cray backend OS delegate to LinuxDistro when no cle_release file
favor backend over frontend OS when name clashes
* cray platform: target detection uses multiple strategies
This commit improves the robustness of target
detection on Cray by trying multiple strategies.
The first one that produces results wins. If
nothing is found only the generic family of the
frontend host is used as a target.
* cray-libsci: add package from NERSC
* build_env: unload cray-libsci module when not explicitly needed
cray-libsci is a package in Spack. The cray PrgEnv
modules load it implicitly when we set up the compiler.
We now unload it after setting up the compiler and
only reload it when requested via external package.
* util/module_cmd: more robust module parsing
Cray modules have documentation inside the module
that is visible to the `module show` command.
Spack module parsing is now robust to documentation
inside modules.
* cce compiler: uses clang flags for versions >= 9.0
* build_env: push CRAY_LD_LIBRARY_PATH into everything
Some Cray modules add paths to CRAY_LD_LIBRARY_PATH
instead of LD_LIBRARY_PATH. This has performance benefits
at load time, but leads to Spack builds not finding their
dependencies from external modules.
Spack now prepends CRAY_LD_LIBRARY_PATH to
LD_LIBRARY_PATH before beginning the build.
* mvapich2: setup cray compilers when on cray
previously, mpich was the only mpi implementation to support
cray systems (because it is the MPI on Cray XC systems).
Cray cluster systems use mvapich2, which now supports cray
compiler wrappers.
* build_env: clean pkgconf from environment
Cray modules silently add pkgconf to the user environment
This can break builds that do not user pkgconf.
Now we remove it frmo the environment and add it again if it
is in the spec.
* cray platform: cheat modules for rome/zen2 module on naples/zen node
Cray modules for naples/zen architecture currently specify
rome/zen2. For now, we detect this and return zen for modules
named `craype-x86-rome`.
* compiler: compiler default versions
When detecting compiler default versions for target/compiler
compatibility checks, Spack previously ran the compiler without
setting up its environment. Now we setup a temporary environment
to run the compiler with its modules to detect its version.
* compilers/cce: improve logic to determine C/C++ std flags
* tests: fix existing tests to play nicely with new cray support
* tests: test new functionality
Some new functionality can only be tested on a cray system.
Add tests for what can be tested on a linux system.
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
Since #9481 Python's None is not permitted as a value for
MV variants. The string 'none' is used instead.
Add the same fix for the amgx and lammps packages
If spack is checked out in a git worktree (see [1]), all git-related
commands fail because the `spack_is_git_repo()`-check is not thorough
enough.
When developing in a feature-branch in a seperate worktree, this is
annoying as all unittests regarding git-related spack commands fail,
cluttering the test results with false-positives.
[1]: https://git-scm.com/docs/git-worktree
Change-Id: I94b573a2c0e058e9ccc169e7ee6561626fbb06fd
* For tests that use the real Spack package repository, the config
needs to avoid using MPI providers that are not intended to be
installed by Spack. Without this, it is possible that Spack tests
which concretize the MPI virtual will end up trying to use an
implementation that it shouldn't (e.g. one that is always
provided externally). See #15666 for an example.
* The mutable_config test fixture was not initializing the scope
roots to the right directories (so the resulting config was empty).
* The current_host fixture in the concretize.py tests was using the
config fixture rather than mutable_config, and was polluting the
config cache for other tests.
* One test in concretize.py was clearing a nonexistent cache
(PackagePrefs._packages_config_cache). This reference has been
removed.
* The test 'test_preferred_compilers' was was depending on cross
test config pollution to succeed. The initial spec before
concretization has been updated to updated to be explicit about
the desired result.
* dev-build: --drop-in <shell>
Add a `--drop-in <shell>` option to `spack dev-build`.
This option will automatically run a
`spack build-env <spec> -- <shell>` at the end of a `dev-build`, e.g.
to quickly drop-and-devel into a build phase of a package.
Example usage:
```
spack dev-build --before cmake --drop-in bash openpmd-api@develop
```
* build_env: drop in unit test
Co-authored-by: Greg Becker <becker33@llnl.gov>
Generally speaking, errors that are encountered when attempting to load
command extensions now terminate the running Spack instance.
* Added new exceptions `spack.cmd.PythonNameError` and
`spack.cmd.CommandNameError`.
* New functions `spack.cmd.require_python_name(pname)` and
`spack.cmd.require_cmd_name(cname)` check that `pname` and `cname`
respectively meet requirements, throwing the appropriate error if not.
* `spack.cmd.get_module()` uses `require_cmd_name()` and passes through
exceptions from module load attempts.
* `spack.cmd.get_command()` uses `require_cmd_name()` and invokes
`get_module()` with the correct command-name form rather than the
previous (incorrect) Python name.
* Added New exceptions `spack.extensions.CommandNotFoundError` and
`spack.extensions.ExtensionNamingError`.
* `_extension_regexp` has a new leading underscore to indicate expected
privacy.
* `spack.extensions.extension_name()` raises an `ExtensionNamingError`
rather than using `tty.warn()`.
* `spack.extensions.load_command_extension()` checks command source
existence early and bails out if missing. Also, exceptions raised by
`load_module_from_file()` are passed through.
* `spack.extensions.get_module()` raises `CommandNotFoundError` as
appropriate.
* Spack `main()` allows `parser.add_command()` exceptions to cause
program end.
Tests:
* More common boilerplate has been pulled out into fixtures including
`sys.modules` dictionary cleanup and resource-managed creation of a
simple command extension with specified contents in the source file
for a single named command.
* "Hello, World!" test now uses a command named `hello-world` instead of
`hello` in order to verify correct handling of commands with hyphens.
* New tests for:
* Missing (or misnamed) command.
* Badly-named extension.
* Verification that errors encountered during import of a command are
propagated upward.
Co-authored-by: Massimiliano Culpo <massimiliano.culpo@gmail.com>
This PR introduces trivial refactoring in:
- `get_existing_elf_rpaths`
- `get_relative_elf_rpaths`
- `get_normalized_elf_rpaths`
- `set_placeholder`
mainly to be more consistent with practices used in other
parts of the code and to simplify functions locally. It also
adds or reworks unit tests for these functions and extends
their docstrings.
Co-authored-by: Patrick Gartung <gartung@fnal.gov>
Co-authored-by: Peter J. Scheibel <scheibel1@llnl.gov>
Packages in Spack are classes, and we need to be able to execute class
methods on mock packages. The previous design used instances of a single
MockPackage class; this version gives each package its own class that can
spider depenencies. This allows us to implement class methods like
`possible_dependencies()` on mock packages.
This design change moves mock package creation into the
`MockPackageMultiRepo`, and mock packages now *must* be created from a
repo. This is required for us to mock `possible_dependencies()`, which
needs to be able to get dependency packages from the package repo.
Changes include:
* `MockPackage` is now `MockPackageBase`
* `MockPackageBase` instances must now be created with
`MockPackageMultiRepo.add_package()`
* add `possible_dependencies()` method to `MockPackageBase`
* refactor tests to use new code structure
* move package mocking infrastructure into `spack.util.mock_package`,
as it's becoming a more sophisticated class and it gets lots in `conftest.py`
The variants table in `spack info` is cramped, as the *widest* it can be
is 80 columns. And that's actually only sort of true -- the padding
calculation is off, so it still wraps on terminals of size 80 because it
comes out *slightly* wider.
This change looks at the terminal size and calculates the width of the
description column based on it. On larger terminals, the output looks
much nicer, and on small terminals, the output no longer wraps.
Here's an example for `spack info qmcpack` with 110 columns.
Before:
Name [Default] Allowed values Description
==================== ==================== ==============================
afqmc [off] on, off Install with AFQMC support.
NOTE that if used in
combination with CUDA, only
AFQMC will have CUDA.
build_type [Release] Debug, Release, The build type to build
RelWithDebInfo
complex [off] on, off Build the complex (general
twist/k-point) version
cuda [off] on, off Build with CUDA
After:
Name [Default] Allowed values Description
==================== ==================== ========================================================
afqmc [off] on, off Install with AFQMC support. NOTE that if used in
combination with CUDA, only AFQMC will have CUDA.
build_type [Release] Debug, Release, The build type to build
RelWithDebInfo
complex [off] on, off Build the complex (general twist/k-point) version
cuda [off] on, off Build with CUDA
Update compiler config with bootstrapped compiler when it was already installed and added config defaults to code so mutable_config test fixture works.
To specify an environment for a comment, the user can specify
"spack -e <env>". The documentation incorrectly specified "-E" (which
is actually used to ignore any implicit use of environments).
If the Spack compiler wrapper encounters any "-isystem" option, then
when adding include directories for Spack dependencies, Spack will
use "-isystem" instead of "-I". This prevents Spack-generated "-I"
options from overriding the "-isystem" options generated by the build
system. To ensure that build-system "-isystem" directories are
searched first, Spack places all of its inserted "-isystem"
directories after.
The new ordering of -isystem includes is:
* -isystem from build system (not system directories)
* -isystem from Spack
* -isystem from build system (for directories like /usr/include)
The prior order of "-I" arguments is preserved (although as of this
commit Spack no longer generates -I if -isystem is detected):
* -I from build system (not system directories)
* -I from Spack (only if there are no "-isystem" options)
* -I from build system (for directories like /usr/include)
Since #16132, we've consolidated the setting of FORCE_UNSAFE_CONFIGURE to
`autotools.py`, so we don't need to use it in packages like `coreutils`,
in our commands, or in our container recipes.
- [x] Remove FORCE_UNSAFE_CONFIGURE from packages
- [x] Remove FORCE_UNSAFE_CONFIGURE from container recipes
- [x] Remove FORCE_UNSAFE_CONFIGURE from `spack ci` command
This commit sets the `FORCE_UNSAFE_CONFIGURE` environment variable to 1 in autotools builds.
We see a lot of builds popping up and complaining about `FORCE_UNSAFE_CONFIGURE`. This behavior is not actually part of `autoconf` per se. It comes from this patch to `mknod.m4`, which is used by a lot of autoconf builds:
* https://lists.gnu.org/archive/html/bug-gnulib/2010-07/msg00282.html
Which originated from this problem that someone had on AIX:
* https://lists.gnu.org/archive/html/bug-gnulib/2010-07/msg00279.html
The gist of the problem seems to be that they want to check whether `mknod` can do something as root, but instead of checking whether they're running as root and using `su` or something to test this, they just made it harder to run `configure` as root.
This seems very ad hoc and this is one of many checks that are run as root in `configure`. Many of them run before this check, so it's not clear that the `FORCE_UNSAFE_CONFIGURE` thing is even preventing bad things from happening.
So:
1. This only happens in `autotools` builds, so we should go ahead and put it into `autotools.py` instead of in the global build environment, and
2. The variable does too little and provides a false sense of security in the first place, so we'll just disable it and avoid the nuisance. If we really feel strongly about this we can put some warnings in Spack about running as root, but at the top level, not in the middle of an already running script like `configure`.
* SourceForge: Mirror Mixin
Add a mixing class for direct `CNAME`s to sourceforge mirrors.
Since the main gateway servers are often down, this could reduce
timeouts and fetch errors for sourceforge.net hosted software.
* SourceForge: unspectacular mirror replacement
add mirrors to all sourceforge packages with trivial
download logic.
tested fetch of latest version of each of these packages
with various mirrors before committing.
* SourceForge: xz
the author homepage is chronocially overrun and this is the offical
upload with many mirrors.
`DYLD_LIBRARY_PATH` can frequently break builtin macOS software when
pointed at Spack libraries. This is because it takes *higher* precedence
than the default library search paths, which are used by system software.
`DYLD_FALLBACK_LIBRARY_PATH`, on the other hand, takes lower precedence.
At first glance, this might seem bad, because the software installed by
Spack in an environment needs to find *its* libraries, and it should not
use the defaults. However, Spack's isntallations are always `RPATH`'d,
so they do not have this problem.
`DYLD_FALLBACK_LIBRARY_PATH` is thus useful for things built in an
environment that need to use Spack's libraries, that don't set *their*
RPATHs correctly for whatever reason. We now prefer it to
`DYLD_LIBRARY_PATH` in modules and in environments because it helps a
little bit, and it is much less intrusive.
provided (#15662).
Prior to this fix, the checked Spec object would not be populated, and
concretization would fail.
Co-authored-by: Marc Allen <mrcall@amazon.com>
`spack test` has a spurious '[+] ' in the output:
```
lib/spack/spack/test/install.py .........[+] ......
```
Output is properly suppressed:
```
lib/spack/spack/test/install.py ...............
```
Makes the following changes:
* (Fixes#15620) tty configuration was failing when stdout was
redirected. The implementation now creates a pseudo terminal for
stdin and checks stdout properly, so redirections of stdin/out/err
should be handled now.
* Handles terminal configuration when the Spack process moves between
the foreground and background (possibly multiple times) during a
build.
* Spack adjusts terminal settings to allow users to to enable/disable
build process output to the terminal using a "v" toggle, abnormal
exit cases (like CTRL-C) could leave the terminal in an unusable
state. This is addressed here with a special-case handler which
restores terminal settings.
Significantly extend testing of process output logger:
* New PseudoShell object for setting up a master and child process
and configuring file descriptor inheritance between the two
* Tests for "v" verbosity toggle making use of the added PseudoShell
object
* Added `uniq` function which takes a list of elements and replaces
any consecutive sequence of duplicate elements with a single
instance (e.g. "112211" -> "121")
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
The performance improvements done in #14693 where leaving the DB in an inconsistent state when specs were removed from it. This PR updates the DB internal state whenever the DB is written to a file.
Note that we still cannot properly enumerate installed dependents, so there is a TODO in this code. Fixing that will require the dependents dictionaries in specs to be re-keyed (either by hash, or not keyed at all -- a list would do). See #11983 for details.
Reading the database repeatedly can be quite slow. We need a way to speed
up Spack when it reads the DB multiple times, but the DB has not been
modified between reads (which is nearly all the time).
- [x] Add a file containing a unique uuid that is regenerated at database
write time. Use this uuid to suppress re-parsing the database
contents if we know a previous uuid and the uuid has not changed.
- [x] Fix mutable_database fixture so that it resets the last seen
verifier when it resets.
- [x] Enable not rereading the database immediately after a write. Make
the tests reset the last seen verifier in between tests that use the
database fixture.
- [x] make presence of uuid module optional
Removed the code that was converting the old index.yaml format into
index.json. Since the change happened in #2189 it should be
considered safe to drop this (untested) code.
* only override spec prefix for non-external packages
* add test that environment shell modifications respect explicitly-specified prefixes for external packages
* add clarifying comment
spack.util.environment_after_sourcing_files compares the local
environment against a shell environment after having sourced a
file; but this ends up including the default shell profile and
rc, which might differ from the local environment.
To change this, compare against the default shell environment,
expressed here as 'source /dev/null'.
According to my nightly CI/CD tests, x.org is another large provider
of software in common build chains that is often down.
Added a hand-selected amount of mirrors that is well up-to-sync.
Tested with `util-macros` that has a quite "recent" patch release.
Other packages to follow in an individual PR.
Makes the following changes:
* (Fixes#15620) tty configuration was failing when stdout was
redirected. The implementation now creates a pseudo terminal for
stdin and checks stdout properly, so redirections of stdin/out/err
should be handled now.
* Handles terminal configuration when the Spack process moves between
the foreground and background (possibly multiple times) during a
build.
* Spack adjusts terminal settings to allow users to to enable/disable
build process output to the terminal using a "v" toggle, abnormal
exit cases (like CTRL-C) could leave the terminal in an unusable
state. This is addressed here with a special-case handler which
restores terminal settings.
Significantly extend testing of process output logger:
* New PseudoShell object for setting up a master and child process
and configuring file descriptor inheritance between the two
* Tests for "v" verbosity toggle making use of the added PseudoShell
object
* Added `uniq` function which takes a list of elements and replaces
any consecutive sequence of duplicate elements with a single
instance (e.g. "112211" -> "121")
Co-authored-by: Adam J. Stewart <ajstewart426@gmail.com>
* Moved link to the right place in the docs
* Fixed a few minor issues in extensions docs
Fixed a typo, added a subsubsection for better
navigation, reworded "modules in Python" as
"Python packages"
sourceware.org is often quite overrun and times out or results in
certificate errors.
Since libffi, bzip2, elfutils, etc. are quite fundamental in
build chains, lets add some official mirrors.
libffi, bzip2, elfutils, lvm2, valgrind: add mirrors
The performance improvements done in #14693 where leaving the DB in an inconsistent state when specs were removed from it. This PR updates the DB internal state whenever the DB is written to a file.
Note that we still cannot properly enumerate installed dependents, so there is a TODO in this code. Fixing that will require the dependents dictionaries in specs to be re-keyed (either by hash, or not keyed at all -- a list would do). See #11983 for details.
* Skip collection of compiler link paths if compiler does not define a verbose flag
* modules config bug: allow user to configure a compiler without an explicit entry for loaded modules
* Add capability for detecting build number for Arm compilers
* Fixing fleck8 errors and updating test_arm_version_detection function for more detailed Arm compielr version detection
* Ran flake8 locally and corrected errors
* Altering Arm compielr version check to remove else clause and be more consistent with other compielr version checks. Added test case so both the 'if' and 'else' conditionals of the Arm compiler version check have a test case
Co-authored-by: EC2 Default User <ec2-user@ip-172-31-7-135.us-east-2.compute.internal>
spack.util.environment_after_sourcing_files compares the local
environment against a shell environment after having sourced a
file; but this ends up including the default shell profile and
rc, which might differ from the local environment.
To change this, compare against the default shell environment,
expressed here as 'source /dev/null'.
* only override spec prefix for non-external packages
* add test that environment shell modifications respect explicitly-specified prefixes for external packages
* add clarifying comment
Currently, to force Spack to use an external MPI, you have to specify `buildable: False`
for every MPI provider in Spack in your packages.yaml file. This is both tedious and
fragile, as new MPI providers can be added and break your workflow when you do a
git pull.
This PR allows you to specify an entire virtual dependency as non-buildable, and
specify particular implementations to be built:
```
packages:
all:
providers:
mpi: [mpich]
mpi:
buildable: false
paths:
mpich@3.2 %gcc@7.3.0: /usr/packages/mpich-3.2-gcc-7.3.0
```
will force all Spack builds to use the specified `mpich` install.
Removed provider_index use of 'import from' and refactored a few routines to a further subclassing of _IndexBase for implementing user defined bindings of provider specs.
* relocate: removed import from statements
* relocate: renamed *Exception to *Error
This aims at consistency in naming with both
the standard library (ValueError, AttributeError,
etc.) and other errors in 'spack.error'.
Improved existing docstrings
* relocate: simplified search function by un-nesting conditionals
The search function that searches for patchelf has been
refactored to remove deeply nested conditionals.
Extended docstring.
* relocate: removed a condition specific to unit tests
* relocate: added test for _patchelf
Our unit tests run many times. Any unit test which actually installs
a package (which involves fetching code on the internet) is a severe
bug because it runs an installation many times (i.e. re-downloading
the same package for each version of Python that we run unit tests
for).
This reverts commit 25893f1, which added tests that install real
packages.
If the Python used by Spack does not include Setuptools, then
'spack test' will fail because Spack's vendored pytest dependency
imports and uses Setuptools in some of its functions. It turns out
that Spack doesn't use the functionality those methods enable, so
this PR removes those functions and thereby allows 'spack test' to
run without Setuptools.
For any Spack test using Spack's YAML configuration, avoid using real
Spack configuration that has been cached by other tests and Spack
startup logic. Previously this was only done for tests using
'mutable_config' (i.e. those which expected to *change* the
configuration of Spack), but in fact all tests that read Spack config
should use it.
This was an issue when running tests within an environment, because
compiler configuration ends up being queried earlier, and the user's
real config "leaks" into the cache. Outside an environment, the cache
is never set until tests touch it, so we weren't seeing this issue.
`spack test` has a spurious '[+] ' in the output:
```
lib/spack/spack/test/install.py .........[+] ......
```
Output is properly suppressed:
```
lib/spack/spack/test/install.py ...............
```
Reading the database repeatedly can be quite slow. We need a way to speed
up Spack when it reads the DB multiple times, but the DB has not been
modified between reads (which is nearly all the time).
- [x] Add a file containing a unique uuid that is regenerated at database
write time. Use this uuid to suppress re-parsing the database
contents if we know a previous uuid and the uuid has not changed.
- [x] Fix mutable_database fixture so that it resets the last seen
verifier when it resets.
- [x] Enable not rereading the database immediately after a write. Make
the tests reset the last seen verifier in between tests that use the
database fixture.
- [x] make presence of uuid module optional
Spack currently cannot run as a background process uninterrupted because some of the logging functions used in the install method (especially to create the dynamic verbosity toggle with the v key) cause the OS to issue a SIGTTOU to Spack when it's backgrounded.
This PR puts the necessary gatekeeping in place so that Spack doesn't do anything that will cause a signal to stop the process when operating as a background process.
Spack currently cannot run as a background process uninterrupted because some of the logging functions used in the install method (especially to create the dynamic verbosity toggle with the v key) cause the OS to issue a SIGTTOU to Spack when it's backgrounded.
This PR puts the necessary gatekeeping in place so that Spack doesn't do anything that will cause a signal to stop the process when operating as a background process.
This makes sure that a package's fetch_options are used when fetching
new versions to checksum. This allows working around problems with
slow servers or those requiring a cookie to be set.
Bug: Spack hangs on some Cray machines
Reason: The TERM environment variable is necessary to run bash -lc "echo $CRAY_CPU_TARGET", but we run that command within env -i, which wipes the environment.
Fix: Manually forward the TERM environment variable to env -i /bin/bash -lc "echo $CRAY_CPU_TARGET"
When trying to use an upstream Spack repository, as of f2aca86 Spack
was attempting to write to the upstream DB based on a new metadata
directory added in that commit. Upstream DBs are read-only, so this
should not occur.
This adds a check to prevent Spack from writing to the upstream DB
fixes#15449
Before this PR a call to pkg.url_for_version was modifying
class attributes determining different results for subsequents
calls and an error when the urls was empty.
This recovers the old behavior of replace_prefix_bin that was
modified to work with elf binaries by prefixing os.sep to new prefix
until length is the same as old prefix.
Testing the install StopIteration exception resulted in an attribute error:
AttributeError: 'StopIteration' object has no attribute 'message'
This PR adds a unit test and resolves that error.
The new build process, introduced in #13100 , relies on a spec's dependents in addition to their dependencies. Loading a spec from a yaml file was not initializing the dependents.
- [x] populate dependents when loading from yaml
The distributed build PR (#13100) -- did not check the install status of dependencies when using the `--only package` option so would refuse to install a package with the claim that it had uninstalled dependencies whether that was the case or not.
- [x] add install status checks for the `--only package` case.
- [x] add initial set of tests
This change stores packages' configure arguments during build and makes
use of them while refreshing module files. This fixes problems such as in
#10716.
Bug: Spack hangs on some Cray machines
Reason: The TERM environment variable is necessary to run bash -lc "echo $CRAY_CPU_TARGET", but we run that command within env -i, which wipes the environment.
Fix: Manually forward the TERM environment variable to env -i /bin/bash -lc "echo $CRAY_CPU_TARGET"
- [x] move some logic for handling virtual packages from the `spack
dependencies` command into `spack.package.possible_dependencies()`
- [x] rework possible dependencies tests so that expected and actual
output are on the left/right respectively
When trying to use an upstream Spack repository, as of f2aca86 Spack
was attempting to write to the upstream DB based on a new metadata
directory added in that commit. Upstream DBs are read-only, so this
should not occur.
This adds a check to prevent Spack from writing to the upstream DB
* try extend path to solve PyQt5.sip not found issue
* disable private sip installation in sippackage class
* undo manual PyQt5 dir creation in py-sip site-packages dir
* fix typo
* fix typo
* also apply fix to PyQt4
* tidy up
* flake8 and tidy up
* tidy and undo hardcoding of python_include_dir
* replace hardcoded python inc dir
* fix minor issues
* rethink include dir variable name
* improve style
* add new versions
* implement new sip setup to qsci installation
* set sip-incdir correctly for the new setup
* setup extend_path thing before qsci python bindings
* take care of conflict
* flake8
* also extend for PyQt4
* improve style
* improve style
* SipPackage build sys should depend on py-sip
* consolidate extend_path fixes into SipPackage
* fix typo
* fix bugs
* flake8
* revert sip doc to pre-resource setup
* import os module
* flake8
Co-authored-by: Sinan81 <sbulut@3vgeomatics.com>
Add a 'define_from_variant` helper function to CMake-based Spack
packages to convert package variants into CMake arguments. For
example:
args.append('-DFOO=%s' % ('ON' if '+foo' in self.spec else 'OFF'))
can be replaced with:
args.append(self.define_from_variant('foo'))
The following conversions are handled automatically:
* Flag variants will be converted to CMake booleans
* Multivalued variants will be converted to semicolon-separated strings
* Other variant values are converted to CMake string arguments
This also adds a 'define' helper method to convert any variable to
a CMake argument. It has the same conversion rules as
'define_from_variant' (but operates directly on values rather than
requiring the user to supply the name of a package variant).
* Buildcache: Install into non-default directory layouts
Store a dictionary mapping of original dependency prefixes to dependency hashes
Use the loaded spec to grab the new dependency prefixes in the new directory layout.
Map the original dependency prefixes to the new dependency prefixes using the dependency hashes.
Use the dependency prefixes map to replace original rpaths with new rpaths preserving the order.
For mach-o binaries, use the dependency prefixes map to replace the dependency library entires for libraries and executables and the replace the library id for libraries.
On Linux, patchelf is used to replace the rpaths of elf binaries.
On macOS, install_name_tool is used to replace the rpaths and dependency libraries of mach-o binaries and the id of mach-o libraries.
On Linux, macholib is used to replace the dependency libraries of mach-o binaries and the id of mach-o libraries.
Binary text with padding replacement is attempted for all binaries for the following paths:
spack layout root
spack prefix
sbang script location
dependency prefixes
package prefix
Text replacement is attempted for all text files using the paths above.
Symbolic links to the absolute path of the package install prefix are replaced, all others produce warnings.
PR #15212 added a new connect_timeout option that can be overridden
using fetch_options but had to specified per-version. This adds a new
per-package variable that can be used to override fetch_options for
all versions in the package. This includes connect_timeout as well
as 'cookie' (e.g. for the jdk package).
Packages can combine package-level fetch_options with per-version
fetch_options, in which case the version fetch_options completely
override the package-level fetch_options.
This commit includes tests for the added behavior.
fixes#15449
Before this PR a call to pkg.url_for_version was modifying
class attributes determining different results for subsequents
calls and an error when the urls was empty.
* add --skip-unstable-versions option to 'spack mirror create' which skips sources/resource for packages if their version is not stable (i.e. if they are the head of a git branch rather than a fixed commit)
* '--skip-unstable-versions' should skip all VCS sources/resources, not just those which are not cachable
Allows spack.config InternalConfigScope and Configuration.set() to
handle keys with trailing ':' to indicate replacement vs merge
behavior with respect to lower priority scopes.
Lists may now be replaced rather than merged (this behavior was
previously only available for dictionaries).
This commit adds tests for the new behavior.
Testing the install StopIteration exception resulted in an attribute error:
AttributeError: 'StopIteration' object has no attribute 'message'
This PR adds a unit test and resolves that error.
This recovers the old behavior of replace_prefix_bin that was
modified to work with elf binaries by prefixing os.sep to new prefix
until length is the same as old prefix.
Removed the code that was converting the old index.yaml format into
index.json. Since the change happened in #2189 it should be
considered safe to drop this (untested) code.
Spack's fflags are meant for both f77 and fc. Therefore, they must
be passed as FFLAGS and FCFLAGS to the configure scripts of
Autotools-based packages.
The distributed build PR (#13100) -- did not check the install status of dependencies when using the `--only package` option so would refuse to install a package with the claim that it had uninstalled dependencies whether that was the case or not.
- [x] add install status checks for the `--only package` case.
- [x] add initial set of tests
connect_timeout can be used to increase the time Spack waits for the
server to answer. This can be used to work around slow connections or
servers.
Fixes#14700
* CudaPackage: add support for Tesla K80 and older CUDA
* Flake8 fixes
* Fix cuda_arch when no arch is set
* Fine-tune cuda_arch=37,50 supported CUDA versions
* CUDA 6.5+ supports SM_37
* Add @svenevs as a maintainer
The new build process, introduced in #13100 , relies on a spec's dependents in addition to their dependencies. Loading a spec from a yaml file was not initializing the dependents.
- [x] populate dependents when loading from yaml
* Buildcache command: add install option -o/--otherarch
This will allow matching specs from other archs, for example
installing macOS buildcaches on linux hosts.
* spack commands --update-completion
args.specs is a list, which results in output like this:
```
eval `spack load --sh ['libxml2', 'xz']`
```
We want this instead:
```
eval `spack load --sh libxml2 xz`
```
This change stores packages' configure arguments during build and makes
use of them while refreshing module files. This fixes problems such as in
#10716.
* Emit a sensible error message if compiler's target is overly specific
fixes#14798fixes#13733
Compiler specifications require a generic architecture family as
their target. This commit improves the error message that is
displayed to users if they edit compilers.yaml and use an overly
specific name.
The hashing logic looks for function calls that are Spack directives.
It expects that when a Spack directive is used that it is referenced
directly by name, and that the directive function is not itself
retrieved by calling another function. When the hashing logic
encountered a function call where the function was determined
dynamically, it would fail (attempting to access a name attribute
that does not happen to exist in this case).
This updates the hashing logic to filter out function calls where the
function is determined dynamically when looking for uses of Spack
directives.
Spack now requires an exact match of the compiler version
requested by the user. A loose constraint can be given to
Spack by using a version range instead of a concrete version
(e.g. 4.5: instead of 4.5).
Sometimes one needs to preserve the (relative order) of
mtimes on installed files. So it's better to just copy
over all the metadata from the source tree to the install
tree. If permissions need fixing, that will be done anyway
afterwards.
One major use case are resource()s:
They're unpacked in one place and then copied to their
final place using install_tree(). If the resource is a
source tree using autoconf/automake, resetting mtimes
uncorrectly might force unwanted autoconf/etc calls.
If the mimetype returned from `file -h -b --mime-type` contains slashes
in its subtype, the tuple returned from `spack.relocate.mime_type` will
have a size larger than two, which leads to errors.
Change-Id: I31de477e69f114ffdc9ae122d00c573f5f749dbb
Fixes#9394Closes#13217.
## Background
Spack provides the ability to enable/disable parallel builds through two options: package `parallel` and configuration `build_jobs`. This PR changes the algorithm to allow multiple, simultaneous processes to coordinate the installation of the same spec (and specs with overlapping dependencies.).
The `parallel` (boolean) property sets the default for its package though the value can be overridden in the `install` method.
Spack's current parallel builds are limited to build tools supporting `jobs` arguments (e.g., `Makefiles`). The number of jobs actually used is calculated as`min(config:build_jobs, # cores, 16)`, which can be overridden in the package or on the command line (i.e., `spack install -j <# jobs>`).
This PR adds support for distributed (single- and multi-node) parallel builds. The goals of this work include improving the efficiency of installing packages with many dependencies and reducing the repetition associated with concurrent installations of (dependency) packages.
## Approach
### File System Locks
Coordination between concurrent installs of overlapping packages to a Spack instance is accomplished through bottom-up dependency DAG processing and file system locks. The runs can be a combination of interactive and batch processes affecting the same file system. Exclusive prefix locks are required to install a package while shared prefix locks are required to check if the package is installed.
Failures are communicated through a separate exclusive prefix failure lock, for concurrent processes, combined with a persistent store, for separate, related build processes. The resulting file contains the failing spec to facilitate manual debugging.
### Priority Queue
Management of dependency builds changed from reliance on recursion to use of a priority queue where the priority of a spec is based on the number of its remaining uninstalled dependencies.
Using a queue required a change to dependency build exception handling with the most visible issue being that the `install` method *must* install something in the prefix. Consequently, packages can no longer get away with an install method consisting of `pass`, for example.
## Caveats
- This still only parallelizes a single-rooted build. Multi-rooted installs (e.g., for environments) are TBD in a future PR.
Tasks:
- [x] Adjust package lock timeout to correspond to value used in the demo
- [x] Adjust database lock timeout to reduce contention on startup of concurrent
`spack install <spec>` calls
- [x] Replace (test) package's `install: pass` methods with file creation since post-install
`sanity_check_prefix` will otherwise error out with `Install failed .. Nothing was installed!`
- [x] Resolve remaining existing test failures
- [x] Respond to alalazo's initial feedback
- [x] Remove `bin/demo-locks.py`
- [x] Add new tests to address new coverage issues
- [x] Replace built-in package's `def install(..): pass` to "install" something
(i.e., only `apple-libunwind`)
- [x] Increase code coverage
* Buildcache creation change the way prefix is copied to workdir.
* install_tree copies hardlinked files
* tarfile creates hardlinked files on extraction.
* create a temporary tarfile from prefix and extract it to workdir
* Use temp tarfile to move workdir to prefix to preserve hardlinks instead of copying
It's often useful to run a module with `python -m`, e.g.:
python -m pyinstrument script.py
Running a python script this way was hard, though, as `spack python` did
not have a similar `-m` option. This PR adds a `-m` option to `spack
python` so that we can do things like this:
spack python -m pyinstrument ./test.py
This makes it easy to write a script that uses a small part of Spack and
then profile it. Previously thee easiest way to do this was to write a
custom Spack command, which is often overkill.
Fixes#10019
If multiple instances of a package were installed in a single
instance of Spack, and they differed in terms of dependencies, then
"spack find" would not distinguish specs based on their dependencies.
For example if two instances of X were installed, one with Y and one
with Z, then "spack find X ^Y" would display both instances of X.
Using `sys.executable` to run Python in a sub-shell doesn't always work in a virtual environment as the `sys.executable` Python is not necessarily compatible with any loaded spack/other virtual environment.
- revert use of sys.executable to print out subshell environment (#14496)
- try instead to use an available python, then if there *is not* one, use `sys.executable`
- this addresses RHEL8 (where there is no `python` and `PYTHONHOME` issue in a simpler way
When removing packages from a view, extensions were being deactivated
in an arbitrary order. Extensions must be deactivated in preorder
traversal (dependents before dependencies), so when this order was
violated the view update would fail.
This commit ensures that views deactivate extensions based on a
preorder traversal and adds a test for it.
Despite trying very hard to keep dicts out of our hash algorithm, we seem
to still accidentally add them in ways that the tests can't catch. This
can cause errors when hashes are not computed deterministically.
This fixes an error we saw with Python 3.5, where dictionary iteration
order is random. In this instance, we saw a bug when reading Spack
environment lockfiles -- The load would fail like this:
```
...
File "/sw/spack/lib/spack/spack/environment.py", line 1249, in concretized_specs
yield (s, self.specs_by_hash[h])
KeyError: 'qcttqplkwgxzjlycbs4rfxxladnt423p'
```
This was because the hashes differed depending on whether we wrote `path`
or `module` first when recomputing the build hash as part of reading a
Spack lockfile. We can fix it by ensuring a determistic iteration order.
- [x] Fix two places (one that caused an issue, and one that did
not... yet) where our to_node_dict-like methods were using regular python
dicts.
- [x] Also add a check that statically analyzes our to_node_dict
functions and flags any that use Python dicts.
The test found the two errors fixed here, specifically:
```
E AssertionError: assert [] == ['Use syaml_dict instead of ...pack/spack/spec.py:1495:28']
E Right contains more items, first extra item: 'Use syaml_dict instead of dict at /Users/gamblin2/src/spack/lib/spack/spack/spec.py:1495:28'
E Full diff:
E - []
E + ['Use syaml_dict instead of dict at '
E + '/Users/gamblin2/src/spack/lib/spack/spack/spec.py:1495:28']
```
and
```
E AssertionError: assert [] == ['Use syaml_dict instead of ...ack/architecture.py:359:15']
E Right contains more items, first extra item: 'Use syaml_dict instead of dict at /Users/gamblin2/src/spack/lib/spack/spack/architecture.py:359:15'
E Full diff:
E - []
E + ['Use syaml_dict instead of dict at '
E + '/Users/gamblin2/src/spack/lib/spack/spack/architecture.py:359:15']
```
This commit introduces a `--no-check-signature` option for
`spack install` so that unsigned packages can be installed. It is
off by default (signatures required).
VSX alitvec extensions are supported by PowerISA from v2.06 (Power7+), but might
not be listed in features.
FMA has been supported by PowerISA since Power1, but might not be listed in
features.
This commit adds these features to all the power ISA family sets.
Add an optional 'submodules_delete' field to Git versions in Spack
packages that allows them to remove specific submodules.
For example: the nervanagpu submodule has become unavailable for the
PyTorch project (see issue 19457 at
https://github.com/pytorch/pytorch/issues/). Removing this submodule
allows 0.4.1 to build.
* Initialize _cached_specs at the file level and check for spec in it before searching mirrors in try_download_spec.
* Make _cached_specs a set to avoid duplicates
* Fix packaging test
* Ignore build_cache in stage when spec.yaml files are downloaded.
`spack -V` previously always returned the version of spack from
`spack.spack_version`. This gives us a general idea of what version
users are on, but if they're on `develop` or on some branch, we have to
ask more questions.
This PR makes `spack -V` check whether this instance of Spack is a git
repository, and if it is, it appends useful information from `git
describe --tags` to the version. Specifically, it adds:
- number of commits since the last release tag
- abbreviated (but unique) commit hash
So, if you're on `develop` you might get something like this:
$ spack -V
0.13.3-912-3519a1762
This means you're on commit 3519a1762, which is 912 commits ahead of
the 0.13.3 release.
If you are on a release branch, or if you are using a tarball of Spack,
you'll get the usual `spack.spack_version`:
$ spack -V
0.13.3
This should help when asking users what version they are on, since a lot
of people use the `develop` branch.
This PR adds a new command to Spack:
```console
$ spack containerize -h
usage: spack containerize [-h] [--config CONFIG]
creates recipes to build images for different container runtimes
optional arguments:
-h, --help show this help message and exit
--config CONFIG configuration for the container recipe that will be generated
```
which takes an environment with an additional `container` section:
```yaml
spack:
specs:
- gromacs build_type=Release
- mpich
- fftw precision=float
packages:
all:
target: [broadwell]
container:
# Select the format of the recipe e.g. docker,
# singularity or anything else that is currently supported
format: docker
# Select from a valid list of images
base:
image: "ubuntu:18.04"
spack: prerelease
# Additional system packages that are needed at runtime
os_packages:
- libgomp1
```
and turns it into a `Dockerfile` or a Singularity definition file, for instance:
```Dockerfile
# Build stage with Spack pre-installed and ready to be used
FROM spack/ubuntu-bionic:prerelease as builder
# What we want to install and how we want to install it
# is specified in a manifest file (spack.yaml)
RUN mkdir /opt/spack-environment \
&& (echo "spack:" \
&& echo " specs:" \
&& echo " - gromacs build_type=Release" \
&& echo " - mpich" \
&& echo " - fftw precision=float" \
&& echo " packages:" \
&& echo " all:" \
&& echo " target:" \
&& echo " - broadwell" \
&& echo " config:" \
&& echo " install_tree: /opt/software" \
&& echo " concretization: together" \
&& echo " view: /opt/view") > /opt/spack-environment/spack.yaml
# Install the software, remove unecessary deps and strip executables
RUN cd /opt/spack-environment && spack install && spack autoremove -y
RUN find -L /opt/view/* -type f -exec readlink -f '{}' \; | \
xargs file -i | \
grep 'charset=binary' | \
grep 'x-executable\|x-archive\|x-sharedlib' | \
awk -F: '{print $1}' | xargs strip -s
# Modifications to the environment that are necessary to run
RUN cd /opt/spack-environment && \
spack env activate --sh -d . >> /etc/profile.d/z10_spack_environment.sh
# Bare OS image to run the installed executables
FROM ubuntu:18.04
COPY --from=builder /opt/spack-environment /opt/spack-environment
COPY --from=builder /opt/software /opt/software
COPY --from=builder /opt/view /opt/view
COPY --from=builder /etc/profile.d/z10_spack_environment.sh /etc/profile.d/z10_spack_environment.sh
RUN apt-get -yqq update && apt-get -yqq upgrade \
&& apt-get -yqq install libgomp1 \
&& rm -rf /var/lib/apt/lists/*
ENTRYPOINT ["/bin/bash", "--rcfile", "/etc/profile", "-l"]
```
* Add binary_distribution::get_spec which takes concretized spec
Add binary_distribution::try_download_specs for downloading of spec.yaml files to cache
get_spec is used by package::try_install_from_binary_cache to download only the spec.yaml
for the concretized spec if it exists.
The Spec parser currently calls `spec.traverse()` after every parse, in
order to set the platform if it's not set. We don't need to do a full
traverse -- we can just check the platforrm as new specs are parsed.
This takes about a second off the time required to import all packages in
Spack (from 8s to 7s).
- [x] simplify platform-setting logic in `SpecParser`.
`filename_for_package_name()` and `dirname_for_package_name()`
automatically construct a Spec from their arguments, which adds a fair
amount of overhead to importing lots of packages. Removing this removes
about 11% of the runtime of importing all packages in Spack (9s -> 8s).
- [x] `filename_for_package_name()` and `dirname_for_package_name()` now
take a string `pkg_name` arguments instead of specs.
* `Environment.__init__` is now synchronized with all writing operations
* `spack uninstall` now synchronizes its updates to any associated environment
* A side effect of this is that the environment is no longer updated piecemeal as specs are uninstalled - all specs are removed from the environment before they are uninstalled
This commit makes two fundamental corrections to tests:
1) Changes 'matches' to the correct 'match' argument for 'pytest.raises' (for all affected tests except those checking for 'SystemExit');
2) Replaces the 'match' argument for tests expecting 'SystemExit' (since the exit code is retained instead) with 'capsys' error message capture.
Both changes are needed to ensure the associated exception message is actually checked.
Updates to environments were not multi-process safe, which prevented them from taking advantage of parallel builds as implemented in #13100. This is a minimal set of changes to enable `spack install` in an environment to be parallelized:
- [x] add an internal lock, stored in the `.spack-env` directory,
to synchronize updates to `spack.yaml` and `spack.lock`
- [x] add `Environment.write_transaction` interface for this lock
- [x] makes use of `Environment.write_transaction` in `install`,
`add`, and `remove` commands
- `uninstall` is not synchronized yet; that is left for a future PR.
Spack commands referring to upstream-installed specs by hash have
been broken since 6b619da (merged September 2019), which added a new
Database function specifically for parsing hashes from command-line
specs; this function was inappropriately attempting to acquire locks
on upstream databases.
This PR updates the offending function to avoid locking upstream
databases and also updates associated tests to catch regression
errors: the upstream database created for these tests was not
explicitly set as an upstream (i.e. initialized with upstream=True)
so it was not guarding against inappropriate accesses.
* Unified environment modifications in config files
fixes#13357
This commit factors all the code that is involved in
the validation (schema) and parsing of environment modifications
from configuration files in a single place. The factored out
code is then used for module files and compiler configuration.
Attributes were separated by dashes in `compilers.yaml` files and
by underscores in `modules.yaml` files. This PR unifies the syntax
on attributes separated by underscores.
Unit testing of environment modifications in compilers
has been refactored and simplified.
Using `sys.executable` to run Python in a sub-shell doesn't always work in a virtual environment as the `sys.executable` Python is not necessarily compatible with any loaded spack/other virtual environment.
- revert use of sys.executable to print out subshell environment (#14496)
- try instead to use an available python, then if there *is not* one, use `sys.executable`
- this addresses RHEL8 (where there is no `python` and `PYTHONHOME` issue in a simpler way
Openblas target is now determined automatically upon inspection of
`TargetList.txt`. If the spack target is a generic architecture family
(like x86_64 or aarch64) the DYNAMIC_ARCH setting is used
instead of targeting a specific microarchitecture.
Instead of another script, this adds a simple argument to `spack
commands` that updates the completion script. Developers can now just
run:
spack commands --update-completion
This should make it simpler for developers to remember to run this
*before* the tests fail. Also, this version tab-completes.
Previously the `spack load` command was a wrapper around `module load`. This required some bootstrapping of modules to make `spack load` work properly.
With this PR, the `spack` shell function handles the environment modifications necessary to add packages to your user environment. This removes the dependence on environment modules or lmod and removes the requirement to bootstrap spack (beyond using the setup-env scripts).
Included in this PR is support for MacOS when using Apple's System Integrity Protection (SIP), which is enabled by default in modern MacOS versions. SIP clears the `LD_LIBRARY_PATH` and `DYLD_LIBRARY_PATH` variables on process startup for executables that live in `/usr` (but not '/usr/local', `/System`, `/bin`, and `/sbin` among other system locations. Spack cannot know the `LD_LIBRARY_PATH` of the calling process when executed using `/bin/sh` and `/usr/bin/python`. The `spack` shell function now manually forwards these two variables, if they are present, as `SPACK_<VAR>` and recovers those values on startup.
- [x] spack load/unload no longer delegate to modules
- [x] refactor user_environment modification calculations
- [x] update documentation for spack load/unload
Co-authored-by: Todd Gamblin <tgamblin@llnl.gov>
This PR adds a `--format=bash` option to `spack commands` to
auto-generate the Bash programmable tab completion script. It can be
extended to work for other shells.
Progress:
- [x] Fix bug in superclass initialization in `ArgparseWriter`
- [x] Refactor `ArgparseWriter` (see below)
- [x] Ensure that output of old `--format` options remains the same
- [x] Add `ArgparseCompletionWriter` and `BashCompletionWriter`
- [x] Add `--aliases` option to add command aliases
- [x] Standardize positional argument names
- [x] Tests for `spack commands --format=bash` coverage
- [x] Tests to make sure `spack-completion.bash` stays up-to-date
- [x] Tests for `spack-completion.bash` coverage
- [x] Speed up `spack-completion.bash` by caching subroutine calls
This PR also necessitates a significant refactoring of
`ArgparseWriter`. Previously, `ArgparseWriter` was mostly a single
`_write` method which handled everything from extracting the information
we care about from the parser to formatting the output. Now, `_write`
only handles recursion, while the information extraction is split into a
separate `parse` method, and the formatting is handled by `format`. This
allows subclasses to completely redefine how the format will appear
without overriding all of `_write`.
Co-Authored-by: Todd Gamblin <tgamblin@llnl.gov>
The gpg2 command isn't always around; it's sometimes called gpg. This is
the case with the brew-installed version, and it's breaking our tests.
- [x] Look for both 'gpg2' and 'gpg' when finding the command
- [x] If we find 'gpg', ensure the version is 2 or higher
- [x] Add tests for version detection.
- [x] Factored to a common place the fixture `testing_gpg_directory`, renamed it as
`mock_gnupghome`
- [x] Removed altogether the function `has_gnupg2`
For `has_gnupg2`, since we were not trying to parse the version from the output of:
```console
$ gpg2 --version
```
this is effectively equivalent to check if `spack.util.gpg.GPG.gpg()` was found. If we need to ensure version is `^2.X` it's probably better to do it in `spack.util.gpg.GPG.gpg()` than in a separate function.
Despite trying very hard to keep dicts out of our hash algorithm, we seem
to still accidentally add them in ways that the tests can't catch. This
can cause errors when hashes are not computed deterministically.
This fixes an error we saw with Python 3.5, where dictionary iteration
order is random. In this instance, we saw a bug when reading Spack
environment lockfiles -- The load would fail like this:
```
...
File "/sw/spack/lib/spack/spack/environment.py", line 1249, in concretized_specs
yield (s, self.specs_by_hash[h])
KeyError: 'qcttqplkwgxzjlycbs4rfxxladnt423p'
```
This was because the hashes differed depending on whether we wrote `path`
or `module` first when recomputing the build hash as part of reading a
Spack lockfile. We can fix it by ensuring a determistic iteration order.
- [x] Fix two places (one that caused an issue, and one that did
not... yet) where our to_node_dict-like methods were using regular python
dicts.
- [x] Also add a check that statically analyzes our to_node_dict
functions and flags any that use Python dicts.
The test found the two errors fixed here, specifically:
```
E AssertionError: assert [] == ['Use syaml_dict instead of ...pack/spack/spec.py:1495:28']
E Right contains more items, first extra item: 'Use syaml_dict instead of dict at /Users/gamblin2/src/spack/lib/spack/spack/spec.py:1495:28'
E Full diff:
E - []
E + ['Use syaml_dict instead of dict at '
E + '/Users/gamblin2/src/spack/lib/spack/spack/spec.py:1495:28']
```
and
```
E AssertionError: assert [] == ['Use syaml_dict instead of ...ack/architecture.py:359:15']
E Right contains more items, first extra item: 'Use syaml_dict instead of dict at /Users/gamblin2/src/spack/lib/spack/spack/architecture.py:359:15'
E Full diff:
E - []
E + ['Use syaml_dict instead of dict at '
E + '/Users/gamblin2/src/spack/lib/spack/spack/architecture.py:359:15']
```
Rework Spack's continuous integration workflow to be environment-based.
- Add the `spack ci` command, which replaces the many scripts in `bin/`
- `spack ci` decouples the CI workflow from the spack instance:
- CI is defined in a spack environment
- environment is in its own (single) git repository, separate from Spack
- spack instance used to run the pipeline is up to the user
- A new `gitlab-ci` section in environments allows users to configure how
specs in the environment should be mapped to runners
- Compilers can be bootstrapped in the new pipeline workflow
- Add extensive documentation on pipelines (see `pipelines.rst` for further details)
- Add extensive tests for pipeline code
* Reorder GNU mirrors (#14395)
As @adamjstewart commented in #14395, GNU suggests to use
their mirror. So reorder the mirror to the top.
GNU Doc: https://www.gnu.org/prep/ftp.en.html
* Use spack.util.url.join for URLs in GNU mirrors (#14395)
One should not use os.path.join for URLs. This does only
work on POSIX systems.
Instead use spack.util.url.join.
So every part in spack uses the same url joining method.
When removing packages from a view, extensions were being deactivated
in an arbitrary order. Extensions must be deactivated in preorder
traversal (dependents before dependencies), so when this order was
violated the view update would fail.
This commit ensures that views deactivate extensions based on a
preorder traversal and adds a test for it.
* Spack can uninstall unused specs
fixes#4382
Added an option to spack uninstall that removes all unused specs i.e.
build dependencies or transitive dependencies that are left
in the store after the specs that pulled them in have been removed.
* Moved the functionality to its own command
The command has been named 'spack autoremove' to follow the naming used
for the same functionality by other widely known package managers i.e.
yum and apt.
* Speed-up autoremoving specs by not locking and re-reading the scratch DB
* Make autoremove work directly on Spack's store
* Added unit tests for the new command
* Display a terser output to the user
* Renamed the "autoremove" command "gc"
Following discussion there's more consensus around
the latter name.
* Preserve root specs in env contexts
* Instead of preserving specs, restrict gc to the active environment
* Added docs
* Added a unit test for gc within an environment
* Updated copyright to 2020
* Updated documentation according to review
Rephrased a couple of sentences, added references to
`spack find` and dependency types.
* Updated function naming and docstrings
* Simplified computation of unused specs
Since the new approach uses private attributes of the DB
it has been coded as a method of that class rather than a
freestanding function.
The imports in `spec.py` are getting to be pretty unwieldy.
- [x] Remove all of the `import from` style imports and replace them with
`import` or `import as`
- [x] Remove a number names that were exported by `spack.spec` that
weren't even in `spack.spec`
Previously, `spack test` automatically passed all of its arguments to
`pytest -k` if no options were provided, and to `pytest` if they were.
`spack test -l` also provided a list of test filenames, but they didn't
really let you completely narrow down which tests you wanted to run.
Instead of trying to do our own weird thing, this passes `spack test`
args directly to `pytest`, and omits the implicit `-k`. This means we
can now run, e.g.:
```console
$ spack test spec_syntax.py::TestSpecSyntax::test_ambiguous
```
This wasn't possible before, because we'd pass the fully qualified name
to `pytest -k` and get an error.
Because `pytest` doesn't have the greatest ability to list tests, I've
tweaked the `-l`/`--list`, `-L`/`--list-long`, and `-N`/`--list-names`
options to `spack test` so that they help you understand the names
better. you can combine these options with `-k` or other arguments to do
pretty powerful searches.
This one makes it easy to get a list of names so you can run tests in
different orders (something I find useful for debugging `pytest` issues):
```console
$ spack test --list-names -k "spec and concretize"
cmd/env.py::test_concretize_user_specs_together
concretize.py::TestConcretize::test_conflicts_in_spec
concretize.py::TestConcretize::test_find_spec_children
concretize.py::TestConcretize::test_find_spec_none
concretize.py::TestConcretize::test_find_spec_parents
concretize.py::TestConcretize::test_find_spec_self
concretize.py::TestConcretize::test_find_spec_sibling
concretize.py::TestConcretize::test_no_matching_compiler_specs
concretize.py::TestConcretize::test_simultaneous_concretization_of_specs
spec_dag.py::TestSpecDag::test_concretize_deptypes
spec_dag.py::TestSpecDag::test_copy_concretized
```
You can combine any list option with keywords:
```console
$ spack test --list -k microarchitecture
llnl/util/cpu.py modules/lmod.py
```
```console
$ spack test --list-long -k microarchitecture
llnl/util/cpu.py::
test_generic_microarchitecture
modules/lmod.py::TestLmod::
test_only_generic_microarchitectures_in_root
```
Or just list specific files:
```console
$ spack test --list-long cmd/test.py
cmd/test.py::
test_list test_list_names_with_pytest_arg
test_list_long test_list_with_keywords
test_list_long_with_pytest_arg test_list_with_pytest_arg
test_list_names
```
Hopefully this stuff will help with debugging test issues.
- [x] make `spack test` send args directly to `pytest` instead of trying
to do fancy things.
- [x] rework `--list`, `--list-long`, and add `--list-names` to make
searching for tests easier.
- [x] make it possible to mix Spack's list args with `pytest` args
(they're just fancy parsing around `pytest --collect-only`)
- [x] add docs
- [x] add tests
- [x] update spack completion
Test configuration files (except modules.yaml) were in the root level of
test/data, but should really just be in their own directory. The absence
of modules.yaml was also breaking module tests if we got module
preferences after tests started, as the mock modules.yaml was not in the
test directory.
The module hook would previously fail if there were no enabled module types.
- Instead of looking for a `KeyError`, default to empty list when the
config variable is not present.
- Convert lambdas to real functions for clarity.
- Remove legacy yaml_version_check() hook
- Remove the pre_run hook from `hook/__init__.py` and `main.py`
We want to discourage the use of pre-run hooks because they have to run
at startup. To keep Spack fast, we should do things like this lazily
instead of in hooks that require spidering directories full of modules.
Continuing to shave small bits of time off startup --
`spack.cmd.common.arguments` constructs many `Args` objects at module
scope, which has to be done for all commands that import it. Instead of
doing this at load time, do it lazily.
- [x] construct Args objects lazily
- [x] remove the module-scoped argparse fixture
- [x] make the mock config scope set dirty to False by default (like the
regular scope)
This *seems* to reduce load time slightly
Previously, fixtures like `config`, `database`, and `store` were
module-scoped, but frequently used as test function arguments. These
fixtures swap out global on setup and restore them on teardown. As
function arguments, they would do the right set-up, but they'd leave the
global changes in place for the whole module the function lived in. This
meant that if you use `config` once, other functions in the same module
would inadvertently inherit the mock Spack configuration, as it would
only be torn down once all tests in the module were complete.
In general, we should module- or session-scope the *STATE* required for
these global objects (as it's expensive to create0, but we shouldn't
module-or session scope the activation/use of them, or things can get
really confusing.
- [x] Make generic context managers for global-modifying fixtures.
- [x] Make session- and module-scoped fixtures that ONLY build filesystem
state and create objects, but do not swap out any variables.
- [x] Make seeparate function-scoped fixtures that *use* the session
scoped fixtures and actually swap out (and back in) the global
variables like `config`, `database`, and `store`.
These changes make it so that global changes are *only* ever alive for a
singlee test function, and we don't get weird dependencies because a
global fixture hasn't been destroyed.
`PackagePrefs` has had a class-level cache of data from `packages.yaml` for
a long time, but it complicates testing and leads to subtle errors,
especially now that we frequently manipulate custom config scopes and
environments.
Moving the cache to instance-level doesn't slow down concretization or
the test suite, and it just caches for the life of a `PackagePrefs`
instance (i.e., for a single cocncretization) so we don't need to worry
about global state anymore.
- [x] Remove class-level caches from `PackagePrefs`
- [x] Add a cached _spec_order object on each `PackagePrefs` instance
- [x] Remove all calls to `PackagePrefs.clear_caches()`
Commands like `spack blame` were printig poorly when redirected to files,
as colify reverts to a single column when redirected. This works for
list data but not tables.
- [x] Force a table by always passing `tty=True` from `colify_table()`
In "spack info" the Variants header currently has two blank
lines under it. That's too much. It looks like the actual
content belongs to something else.
Instead underline the headers to make things more obvious.
This commit removes the `python_version.py` unit test module
and the vendored dependencies `pyqver2.py` and `pyqver3.py`.
It substitutes them with an equivalent check done using
`vermin` that is run as a separate workflow via Github Actions.
This allows us to delete 2 vendored dependencies that are unmaintained
and substitutes them with a maintained tool.
Also, updates the list of vendored dependencies.
`ViewDescriptor.regenerate()` calls `get_all_specs()`, which reads
`spec.yaml` files, which is slow. It's fine to do this once, but
`view.remove_specs()` *also* calls it immediately afterwards.
- [x] Pass the result of `get_all_specs()` as an optional parameter to
`view.remove_specs()` to avoid reading `spec.yaml` files twice.
`ViewDescriptor.regenerate()` was copying specs and stripping build
dependencies, which clears `_hash` and other cached fields on concrete
specs, which causes a bunch of YAML hashes to be recomputed.
- [x] Preserve the `_hash` and `_normal` fields on stripped specs, as
these will be unchanged.
`spack install` previously concretized, writes the entire environment
out, regenerated views, then wrote and regenerated views
again. Regenerating views is slow, so ensure that we only do that once.
- [x] add an option to env.write() to skip view regeneration
- [x] add a note on whether regenerate_views() shouldn't just be a
separate operation -- not clear if we want to keep it as part of write
to ensure consistency, or take it out to avoid performance issues.
Environments need to read the DB a lot when installing all specs.
- [x] Put a read transaction around `install_all()` and `install()`
to avoid repeated locking
Our `LockTransaction` class was reading overly aggressively. In cases
like this:
```
1 with spack.store.db.read_transaction():
2 with spack.store.db.write_transaction():
3 ...
```
The `ReadTransaction` on line 1 would read in the DB, but the
WriteTransaction on line 2 would read in the DB *again*, even though we
had a read lock the whole time. `WriteTransaction`s were only
considering nested writes to decide when to read, but they didn't know
when we already had a read lock.
- [x] `Lock.acquire_write()` return `False` in cases where we already had
a read lock.
If a write transaction was nested inside a read transaction, it would not
write properly on release, e.g., in a sequence like this, inside our
`LockTransaction` class:
```
1 with spack.store.db.read_transaction():
2 with spack.store.db.write_transaction():
3 ...
4 with spack.store.db.read_transaction():
...
```
The WriteTransaction on line 2 had no way of knowing that its
`__exit__()` call was the last *write* in the nesting, and it would skip
calling its write function.
The `__exit__()` call of the `ReadTransaction` on line 1 wouldn't know
how to write, and the file would never be written.
The DB would be correct in memory, but the `ReadTransaction` on line 4
would re-read the whole DB assuming that other processes may have
modified it. Since the DB was never written, we got stale data.
- [x] Make `Lock.release_write()` return `True` whenever we release the
*last write* in a nest.
Lock transactions were actually writing *after* the lock was
released. The code was looking at the result of `release_write()` before
writing, then writing based on whether the lock was released. This is
pretty obviously wrong.
- [x] Refactor `Lock` so that a release function can be passed to the
`Lock` and called *only* when a lock is really released.
- [x] Refactor `LockTransaction` classes to use the release function
instead of checking the return value of `release_read()` / `release_write()`
`ViewDescriptor.regenerate()` checks repeatedly whether packages are
installed and also does a lot of DB queries. Put a read transaction
around the whole thing to avoid repeatedly locking and unlocking the DB.
`Environment.added_specs()` has a loop around calls to
`Package.installed()`, which can result in repeated DB queries. Optimize
this with a read transaction in `Environment`.
Checks for deprecated specs were repeatedly taking out read locks on the
database, which can be very slow.
- [x] put a read transaction around the deprecation check
BundlePackages use a noop fetch strategy. The mirror logic was assuming
that the fetcher had a resource to cach after performing a fetch. This adds
a special check to skip caching if the stage is associated with a
BundleFetchStrategy. Note that this should allow caching resources
associated with BundlePackages.
When updating a mirror, Spack was re-retrieving all patches (since the
fetch logic for patches is separate). This updates the patch logic to
allow the mirror logic to avoid this.
Since cache_mirror does the fetch itself, it also needs to do the
checksum itself if it wants to verify that the source stored in the
mirror is valid. Note that this isn't strictly required because fetching
(including from mirrors) always separately verifies the checksum.
The targets for the cosmetic paths in mirrrors were being calculated
incorrectly as of fb3a3ba: the symlinks used relative paths as targets,
and the relative path was computed relative to the wrong directory.
When creating a cosmetic symlink for a resource in a mirror, remove
it if it already exists. The symlink is removed in case the logic to
create the symlink has changed.
* Some packages (e.g. mpfr at the time of this patch) can have patches
with the same name but different contents (which apply to different
versions of the package). This appends part of the patch hash to the
cache file name to avoid conflicts.
* Some exceptions which occur during fetching are not a subclass of
SpackError and therefore do not have a 'message' attribute. This
updates the logic for mirroring a single spec (add_single_spec)
to produce an appropriate error message in that case (where before
it failed with an AttributeError)
* In various circumstances, a mirror can contain the universal storage
path but not a cosmetic symlink; in this case it would not generate
a symlink. Now "spack mirror create" will create a symlink for any
package that doesn't have one.
`ViewDescriptor.regenerate()` calls `get_all_specs()`, which reads
`spec.yaml` files, which is slow. It's fine to do this once, but
`view.remove_specs()` *also* calls it immediately afterwards.
- [x] Pass the result of `get_all_specs()` as an optional parameter to
`view.remove_specs()` to avoid reading `spec.yaml` files twice.
`ViewDescriptor.regenerate()` was copying specs and stripping build
dependencies, which clears `_hash` and other cached fields on concrete
specs, which causes a bunch of YAML hashes to be recomputed.
- [x] Preserve the `_hash` and `_normal` fields on stripped specs, as
these will be unchanged.
`spack install` previously concretized, writes the entire environment
out, regenerated views, then wrote and regenerated views
again. Regenerating views is slow, so ensure that we only do that once.
- [x] add an option to env.write() to skip view regeneration
- [x] add a note on whether regenerate_views() shouldn't just be a
separate operation -- not clear if we want to keep it as part of write
to ensure consistency, or take it out to avoid performance issues.
Environments need to read the DB a lot when installing all specs.
- [x] Put a read transaction around `install_all()` and `install()`
to avoid repeated locking
Our `LockTransaction` class was reading overly aggressively. In cases
like this:
```
1 with spack.store.db.read_transaction():
2 with spack.store.db.write_transaction():
3 ...
```
The `ReadTransaction` on line 1 would read in the DB, but the
WriteTransaction on line 2 would read in the DB *again*, even though we
had a read lock the whole time. `WriteTransaction`s were only
considering nested writes to decide when to read, but they didn't know
when we already had a read lock.
- [x] `Lock.acquire_write()` return `False` in cases where we already had
a read lock.
If a write transaction was nested inside a read transaction, it would not
write properly on release, e.g., in a sequence like this, inside our
`LockTransaction` class:
```
1 with spack.store.db.read_transaction():
2 with spack.store.db.write_transaction():
3 ...
4 with spack.store.db.read_transaction():
...
```
The WriteTransaction on line 2 had no way of knowing that its
`__exit__()` call was the last *write* in the nesting, and it would skip
calling its write function.
The `__exit__()` call of the `ReadTransaction` on line 1 wouldn't know
how to write, and the file would never be written.
The DB would be correct in memory, but the `ReadTransaction` on line 4
would re-read the whole DB assuming that other processes may have
modified it. Since the DB was never written, we got stale data.
- [x] Make `Lock.release_write()` return `True` whenever we release the
*last write* in a nest.
Lock transactions were actually writing *after* the lock was
released. The code was looking at the result of `release_write()` before
writing, then writing based on whether the lock was released. This is
pretty obviously wrong.
- [x] Refactor `Lock` so that a release function can be passed to the
`Lock` and called *only* when a lock is really released.
- [x] Refactor `LockTransaction` classes to use the release function
instead of checking the return value of `release_read()` / `release_write()`
`ViewDescriptor.regenerate()` checks repeatedly whether packages are
installed and also does a lot of DB queries. Put a read transaction
around the whole thing to avoid repeatedly locking and unlocking the DB.
Users can now list mirrors of the main url in packages.
- [x] Instead of just a single `url` attribute, users can provide a list (`urls`) in the package, and these will be tried by in order by the fetch strategy.
- [x] To handle one of the most common mirror cases, define a `GNUMirrorPackage` mixin to handle all the standard GNU mirrors. GNU packages can set `gnu_mirror_path` to define the path within a mirror, and the mixin handles setting up all the requisite GNU mirror URLs.
- [x] update all GNU packages in `builtin` to use the `GNUMirrorPackage` mixin.
- Add an optional argument so that `possible_dependencies()` will report
missing dependencies.
- Add a test to ensure it works.
- Ignore missing dependencies in `possible_dependencies()` by default.
- this version allows getting possible dependencies of multiple packages
or specs at once.
- New method handles calling `PackageBase.possible_dependencies` multiple
times and passing `visited` dict around.
`Environment.added_specs()` has a loop around calls to
`Package.installed()`, which can result in repeated DB queries. Optimize
this with a read transaction in `Environment`.
Checks for deprecated specs were repeatedly taking out read locks on the
database, which can be very slow.
- [x] put a read transaction around the deprecation check
doesn't understand a custom, user-defined compiler version. However, if
the compiler's version check fails, you can't build anything with the
custom compiler.
- [x] Be more lenient: fall back to the custom compiler version and use
it verbatim if the version check fails.
`pgcc -V` was failing on power machines because it returns 2 (despite
correctly printing version information). On x86_64 machines the same
command returns 0 and doesn't cause an error.
- [x] Ignore return value of 2 for pgcc when doign a version check
Vendors for ARM come out of `/proc/cpuinfo` as hex numbers instead of readable strings.
- Add support for associating vendor names with the hex numbers.
- Also move these mappings from Python code to `microarchitectures.json`
- Move darwin feature name mappings to `microarchitectures.json` as well
* when constructing package hash, default to including a method in the content hash if we can't determine whether it would be included by examining the AST
* add a test for updated content-hash calculations
* refactor content hash tests to eliminate repeated lines
BundlePackages use a noop fetch strategy. The mirror logic was assuming
that the fetcher had a resource to cach after performing a fetch. This adds
a special check to skip caching if the stage is associated with a
BundleFetchStrategy. Note that this should allow caching resources
associated with BundlePackages.
When updating a mirror, Spack was re-retrieving all patches (since the
fetch logic for patches is separate). This updates the patch logic to
allow the mirror logic to avoid this.
Since cache_mirror does the fetch itself, it also needs to do the
checksum itself if it wants to verify that the source stored in the
mirror is valid. Note that this isn't strictly required because fetching
(including from mirrors) always separately verifies the checksum.
The targets for the cosmetic paths in mirrrors were being calculated
incorrectly as of fb3a3ba: the symlinks used relative paths as targets,
and the relative path was computed relative to the wrong directory.
When creating a cosmetic symlink for a resource in a mirror, remove
it if it already exists. The symlink is removed in case the logic to
create the symlink has changed.
* pytest: add __init__ files for all test subdirs
* add licenses to empty files
* Fix Sphinx warning message about comment within docstring
* Further fixes to Sphinx docstring
* fix docstring in generate_package_index() refering to "public" keys as "signing" keys
* use explicit kwargs in push_to_url()
* simplify url_util.parse() per tgamblin's suggestion
* replace standardize_header_names() with the much simpler get_header()
* add some basic tests
* update s3_fetch tests
* update S3 list code to strip leading slashes from prefix
* correct minor warning regression introduced in #11117
* add more tests
* flake8 fixes
* add capsys fixture to mirror_crud test
* add get_header() tests
* use get_header() in more places
* incorporate review comments
This PR allows virtual packages to be added to the specs list using
the add command.
Virtual packages are already allowed in named lists in spack
environments/stacks, and they are already allowed in the specs list
when added using the yaml directly.
I have, more than once, tried to install the list of things that need
to build the docs, only to discover that the list doesn't use Spack's
package names. I'm tired of facepalming....
While I was there I touched up the prose about activating the new
Python packages; activating a python package doesn't add anything to
your PYTHONPATH, it links things into a directory that's *already* on
your PYTHONPATH. Note that this all presupposes that you're using
that same python....
* CUDA HeaderList: Unit Test
* Spec Header Dirs: Only first include/
Avoid matching recurringly nested include paths that usually
refer to internally shipped libraries in packages.
Example in CUDA Toolkit, shipping a libc++ fork internally
with libcu++ since 10.2.89:
`<prefix>/include/cuda/some/more/details/include/` or
`<prefix>/include/cuda/std/detail/libcxx/include`
regex: non-greedy first match of include
Co-Authored-By: Massimiliano Culpo <massimiliano.culpo@gmail.com>
* CUDA: Re-Enable 10.2.89 as Default
* 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
`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
`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
* 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.
* Add a transaction around repeated calls to `spec.prefix` in the activation process
* cache the computation of home in the python package to speed up setting deps
* ensure that module-scope variables are only set *once* per module
* Add a transaction around repeated calls to `spec.prefix` in the activation process
* cache the computation of home in the python package to speed up setting deps
* ensure that module-scope variables are only set *once* per module
`mirror_archive_path` was failing to account for the case where the fetched version isn't known to Spack.
- [x] don't require the fetched version to be in `Package.versions`
- [x] add regression test for mirror paths when package does not have a version
Extensions have been available for a while and the overall design
seems solid enough to be feasible for extensions without losing
backward compatibility.
* Some packages (e.g. mpfr at the time of this patch) can have patches
with the same name but different contents (which apply to different
versions of the package). This appends part of the patch hash to the
cache file name to avoid conflicts.
* Some exceptions which occur during fetching are not a subclass of
SpackError and therefore do not have a 'message' attribute. This
updates the logic for mirroring a single spec (add_single_spec)
to produce an appropriate error message in that case (where before
it failed with an AttributeError)
* In various circumstances, a mirror can contain the universal storage
path but not a cosmetic symlink; in this case it would not generate
a symlink. Now "spack mirror create" will create a symlink for any
package that doesn't have one.
* Add process to determine aarch64 microarchitecture
* add microarchitectures for thunderx2 and a64fx
* Add optimize flags for gcc on aarch64 family processors thunderx2 and a64fx.
* Add optimize flags for clang on aarch64 family processors thunderx2 and a64fx
* Add testing for thunderx2 and a64fx microarchitectures
* 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
* 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.
* 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.
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.
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.
when making a package relative, relocate links relative to link directory
rather than the full link path (which includes the file name) because `os.path.relpath` expects a directory.
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.
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.
* Docs update for deprecated `spack sha256`
* Added macOS shasum
* Update lib/spack/docs/packaging_guide.rst
Co-Authored-By: Adam J. Stewart <ajstewart426@gmail.com>
- [x] Use higher contrast terminal output font
- [x] Use higher contrast code block background color than default
- [x] Use a noticeable prompt character
See also https://github.com/spack/spack-tutorial/pull/10.
`mirror_archive_path` was failing to account for the case where the fetched version isn't known to Spack.
- [x] don't require the fetched version to be in `Package.versions`
- [x] add regression test for mirror paths when package does not have a version
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.
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.
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
* 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
- [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
- [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
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`.
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`.
