summaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)AuthorFilesLines
2016-01-12Cleanup and proper flag satisfaction for partial specsGregory Becker4-13/+17
2016-01-12Implemented flags as lists for subsettingGregory Becker4-44/+72
2016-01-11Fixed concretizationGregory Becker2-1/+3
2015-11-11merged in work from other computer and starting to track down bugsGregory Becker1-3/+5
2015-11-10Eliminated two bugs found by spack test and rewrote the parser to be more saneTodd Gamblin2-82/+23
2015-11-10Fixed bug that spack.db.exists() returned True for anonymous specsGregory Becker3-5/+10
2015-11-10Removed "any-pkg-name" and replaced it with empty string. Also changed cflag ↵Gregory Becker4-28/+35
concretizer to concretize each flag individually, allowing us to have unconcretized FlagMap objects for find and uninstall. Now empty flags in find match any, whereas specifying +cflags=\'\' matches only those with empty strings for flags
2015-11-10Removed cflags from default format string and made them an option within the ↵Gregory Becker2-15/+39
compiler string. Added -f option to find command; with -f, find prints flags
2015-11-10added default compilers into spec and fixed constrain/concretize bugGregory Becker2-25/+20
2015-11-10Changed flag default to "". Updated printing and other logic to match. Seems ↵Gregory Becker3-40/+16
to have solved error in normalize.
2015-11-10partial commit of cflags for debuggingGregory Becker4-12/+43
2015-11-10Commit of compiler flags addition:Gregory Becker9-71/+208
Flags are passed from the command line all the way through build environments to environment variables. Flags are specified using +name=value and values are quoted using escaped quotes when necessary. Future work includes using the flags in the compiler wrapper script and hopefully updating the parser for a gentler user experience of the spec language.
2015-11-10reclaimed the = sign. Architectures now specified by +arch= instead. Decided ↵Gregory Becker7-56/+100
to prepend flag names with + for clarity in spec names and ease of parsing. Also generalized variants, although there is not yet a way to specify a generalized (name=value) variant.
2015-11-10Headless specs and /hash spec specificationGregory Becker4-11/+70
2015-11-10partial commit to merge databaseGregory Becker3-9/+42
2015-11-10partial commit of cflags workGregory Becker6-11/+147
2015-11-04Fix SPACK-93, SPACK-94, GitHub #150Todd Gamblin4-21/+18
- `remove_prefix` was modified to remove from the DB, but the package may not have been added to the DB yet when `remove_prefix` is called from `cleanup`. - Made `remove_prefix` a pure utility function (it just removes the prefix) - Added `installed_db.remove()` call only after the `remove_prefix` in `uninstall`.
2015-11-04Make architecture reflect OS *and* machine. Use Python's platform module.Todd Gamblin1-11/+8
2015-10-30Fixed typo : write_lock() -> write_transaction()Massimiliano Culpo3-3/+3
2015-10-29Merge Greg Becker's build database support into develop.Todd Gamblin22-139/+1641
2015-10-29Merge pull request #124 from scheibelp/features/testinstall-cmdTodd Gamblin6-3/+350
Features/testinstall cmd
2015-10-29Merge pull request #145 from markcmiller86/v-char-bugfixTodd Gamblin1-2/+2
generalize one of the cases to handle many more kinds of formatting o…
2015-10-29Merge pull request #139 from mathstuf/use-uname-for-archTodd Gamblin1-10/+10
architecture: use uname if available
2015-10-27generalize one of the cases to handle many more kinds of formatting optionsmiller861-2/+2
2015-10-27Build database working with simple transaction support; all tests passing.Todd Gamblin12-230/+645
2015-10-27Fix stupid lock bug.Todd Gamblin3-41/+53
- Code simplification ignored case where exception was raised. - If LockError was raised, read and write counts were incremented erroneously. - updated lock test.
2015-10-27Not all package stages have an archive file (e.g. source code repos) but all ofPeter Scheibel1-1/+1
them do have a source_path: use this instead to check whether the package resources were successfully retrieved.
2015-10-27Lock acquires return True/False depending on whether they got POSIX lock.Todd Gamblin1-4/+15
2015-10-26architecture: remove custom mac_type methodBen Boeckel1-13/+0
2015-10-26architecture: use uname if availableBen Boeckel1-1/+14
2015-10-26Fetch errors were also terminating runs of test-install with system exit, soPeter Scheibel3-1/+10
stage.fetch() was updated to raise a FetchError instead of calling tty.die(). Output is the same for spack install in case of a fetch error.
2015-10-26python: use the setdefault method on dictBen Boeckel3-16/+7
It allows more concise code and skips some key lookups.
2015-10-26Making SpackError reference consistent.Peter Scheibel1-2/+1
2015-10-26Automatically create a 'test-output' directory in the current directory if noPeter Scheibel1-1/+4
output path is specified. Test output files are placed in this directory. Furthermore the filenames now have the prefix "test" (but otherwise are the string representation of the spec ID as before).
2015-10-24Working Lock class, now uses POSIX fcntl locks, extensive unit test.Todd Gamblin3-84/+350
- llnl.util.lock now uses fcntl.lockf instead of flock - purported to have more NFS compatibility. - Added an extensive test case for locks. - tests acquiring, releasing, upgrading, timeouts, shared, & exclusive cases.
2015-10-24Add a multiprocess Barrier class to use for testing parallel code.Todd Gamblin1-1/+49
2015-10-23The HTML number conversion regex operating against a byte string will onlyPeter Scheibel1-1/+2
convert individual bytes, so therefore incorrectly converts utf-8 encoded characters. Decoding byte strings to unicode objects results in correct HTML number encodings.
2015-10-231. Rename CommandError -> InstallErrorPeter Scheibel2-5/+7
2. InstallError now subclasses SpackError vs. StandardError (so it is now handled by the spack shell script)
2015-10-221. Added CommandError exception to build_environmentPeter Scheibel2-15/+28
2. The parent of a failed child process in build_environment.fork no longer calls sys.exit - instead it raises a CommandError (from [1]) 3. test-install command now attempts to install all packages even if one fails
2015-10-22Generate test results (designated as skipped) for parents of failed dependenciesPeter Scheibel2-24/+45
2015-10-21Merge pull request #129 from mathstuf/separate-prefixTodd Gamblin1-6/+8
spack: split spack_root from prefix
2015-10-18Add Spack logo.Todd Gamblin1-3/+3
2015-10-17Better info variant output.Todd Gamblin1-3/+13
2015-10-15Created unit test for core logic in test-install command.Peter Scheibel3-12/+145
2015-10-15Better description for test-install commandPeter Scheibel1-1/+1
2015-10-15Minor edit for clarity (generate output for single top level spec vs. iteratingPeter Scheibel1-2/+1
through collection of size 1)
2015-10-15Originally I enforced specifying 1 top-level package with the test-installPeter Scheibel1-5/+4
command by having it consume exactly 1 positional argument (i.e. by removing "nargs=argparse.REMAINDER") but this does not work when configuring dependencies of a top-level package (which show up as additional positional args). Instead now there is an explicit check to ensure there is only 1 top-level package.
2015-10-151. Specifying the output file path for test-install is now an option (vs. anPeter Scheibel2-12/+22
argument). The default path is [package id].xml in the CWD where test-install is called from. 2. Fixed a bug with package.build_log_path (which was added in this branch). 3. keep_stage for package.do_install is now set. This allows uninstalling and reinstalling packages without (re) downloading them.
2015-10-15Move logic for tracking the build log into package.py (since that is what isPeter Scheibel2-9/+10
managing the build log) and expose as package.build_log_path.
2015-10-15Update test failure output: don't include the entire build log, just lines whichPeter Scheibel1-5/+10
mention errors (or if no such lines can be found, output the last 10 lines from the log).