Skip to content

Tags

Tags give the ability to mark specific points in history as being important
  • 1.19.0
    8b58d19d · Bump version to 1.19.0 ·
    Release 1.19.0
    
    - Change man page extension from .roff to its chapter.
    - Inform user that jobs must be added in order for Litani to run a build.
  • 1.18.0
    e72241cb · Bump version to 1.18.0 ·
    Release 1.18.0
    
    - Add new transform-jobs command
    - Add --fast option to Litani test suite
    - Add --output-dir flag to Litani test suite
    - Litani will no longer print errors when not connected to a tty device
    - Add documentation for new flags
  • 1.17.0
    97897e71 · Bump version to 1.17.0 ·
    Release 1.17.0
    
    - Litani will no longer crash when running in a directory where one of the
    subdirectories contains a cache pointer from an earlier run.
    - Litani will distinguish on the pipeline HTML page the jobs that are currently
    running from ones whose execution has not started.
  • 1.16.0
    8b4a4ffb · Bump version to 1.16.0 ·
    Release 1.16.0
    
    This is a bugfix release.
    
    - Litani will no longer crash when rendering dependency graphs of command lines
      containing HTML entities
    - Litani will no longer crash when the graph rendering output fails in general.
  • 1.15.0
    8579a51f · Bump version to 1.15.0 ·
    Release 1.15.0
    
    This is a bugfix release. It changes how Litani spawns commands so that it can
    correctly terminate the command and all descendant processes if the job has an
    associated timeout.
  • 1.14.0
    f7e1701e · Bump version to 1.14.0 ·
    Release 1.14.0
    
    - Fix bugs in Graphviz output for pipeline graphs
    - Add a flag to make graph output optional
    - Add a pipeline summary at the top of the dashboard
    - Fix spelling
    - Only render pipeline graph when dot is installed
    - Fix --phony-outputs documentation
    - Fix line length
    - Fix crash in graph when job has no description
    - Improve usability of pipeline graph
    - Add documentation for outcome table
    - Add more useful synopses to each man page
    - Add support for custom stage names
  • 1.13.0
    53e35c3c · Bump version to 1.13.0 ·
    Release 1.13.0
  • 1.12.0
    d42428f5 · Bump version to 1.12.0 ·
    Release 1.12.0
    
    - The --phony-outputs flag now has a slightly changed semantic
        Outputs passed to --phony-outputs no longer need to be also passed
        to --outputs to be considered an output. That is, the set of outputs
        of a job is the union of the outputs and the phony outputs. This is
        a backward-compatible change that should make configure scripts
        shorter.
    
    - The `dump-run` command is more reliable when being run in parallel
  • 1.11.1
    ea0a5ce8 · Bump version to 1.11.1 ·
    Release 1.11.1
    
    - Make dependency on graphviz dot optional
  • 1.11.0
    ec080957 · Bump version to 1.11.0 ·
    Release 1.11.0
    
    - Added --phony-outputs flag to litani add-job
    - Added litani dump-run command
    - Added auto-generated documentation for run.json schema
    - Added more documentation
  • 1.10.0
    8b23cb74 · Bump version to 1.10.0 ·
    Release 1.10.0
    
    - "aux" key added to run API
        This is intended for users to fill out their own custom data when
        processing the run.json file. The run.json file has a tightly
        specified schema that allows for no additions, but the "aux" key
        provides a way to add arbitrary data to a run.
    
    - UI improvements
        The parallelism graph now shows the number of processing cores as a
        horizontal line, and the job memory usage section of the dashboard
        now displays a message if no memory trace was captured.
  • 1.9.0
    bba4c824 · Bump version to 1.9.0 ·
    Release 1.9.0
    
    - Manual pages
        Litani now has work-in-progress documentation, currently including
        man pages for the three most commonly-used subcommands
    
    - Gnuplot dependency is now optional
        This is to support users who run Litani in minimal containers
    
    - Improved usability of output directory symlink
        Litani init now prints the path to the symlink by default, unless
        the new --no-print-out-dir flag was passed
  • 1.8.0
    f7da9cc6 · Bump version to 1.8.0 ·
    Release 1.8.0
    
    - Job pool
        Users can now add jobs to a particular "job pool", which limits the
        parallelism of those jobs. Multiple pools can be created, each with
        their own parallelism limits. This is intended to inhibit the
        parallelism of jobs that use a contended resource, like disk I/O or
        network bandwidth.
    
    - UI improvement on pipeline page
        Each job on the pipeline page now has a hyperlink to that job,
        making it easy to link to particularly interesting jobs hosted on
        the web.
    
    - Memory profiler
        Litani can now profile the memory of particular jobs and display
        graphs of those jobs' memory use on the dashboard. This can be
        helpful when deciding when to place particular jobs into a job pool.
    
    - Parallelism measurement
        Litani now tracks how many jobs are running in parallel at any one
        moment, and displays that as a graph on the dashboard. This can be
        helpful when trying to increase CPU core utilization or trying to
        use cloud resources as efficiently as possible.
  • 1.7.1
    047c9f15 · Bump version to 1.7.1 ·
    Release 1.7.1
    
    This minor release fixes a bug with the graph renderer.
  • 1.7.0
    5ab245fb · Bump version to 1.7.0 ·
    Release 1.7.0
    
    - The --outcome-table flag can be used to specify a JSON file that
      Litani uses to decide whether a job was successful.
    
    - The --output-directory flag can be used to specify where Litani should
      write its output files.
    
    - The --output-symlink flag can be used to create a symbolic link to the
      output directory.
    
    - The --output-prefix flag is an alternative to --output-directory that
      specifies a directory, into which Litani will create a unique
      directory for writing its output files.
    
    - Litani now displays a table of contents on each pipeline page. Each
      entry in the table of contents is a hyperlink that points to a job on
      the page.
    
    - Litani now renders each pipeline as a graph and displays that graph on
      the pipeline page, allowing users to see the dependency layout of all
      the jobs in the pipeline.
  • 1.6.0
    b6522950 · Bump version to 1.6.0 ·
    Release 1.6.0
  • 1.5.0
    a1b040d1 · Bump version to 1.5.0 ·
    Release 1.5.0
    
    - The implementation of directory locking no longer releases the lock
      upon creation. This fixes a problem where external processes could
      release a lock that a different process had acquired.
  • 1.4.0
    fe8d0f91 · Bump version to 1.4.0 ·
    Release 1.4.0
    
    - lib.litani now exposes a LockableDirectory API that can be used to
      lock directories against concurrent access by more than one process.
      External processes can use this API to lock report directories while
      reading or copying them.
    
    - Litani now touches a file called `.litani-expired` in HTML report
      directories as soon as a new report directory has been written, and
      after the `html` symbolic link has been updated to point to the new
      directory. Litani then deletes all report directories that are both
      expired and unlocked.
  • 1.3.0
    Release 1.3.0
    
    Litani now has a top-level command called `print-capabilities` that
    prints out a list of features in either human or machine-readable
    format.  The intention is that any new API-changing features will be
    accompanied by a new entry in this list
  • 1.2.0
    172bbc91 · Bump version to 1.2.0 ·
    Release 1.2.0
    
    - Litani now renders the HTML directory atomically. The `html` directory is now
      a symbolic link that is updated only after the report has been completely
      written.
    - Added dark mode for all report pages