hash

README

This utility produces a consistent filename from three pieces of information. This program is intended to produce filenames for files that represent information about a group of packages. Its intended first use-case is to provide a consistent filename for the file containing a JSON list of archive.Spec structs. The current setup is to group the building of packages by:

1. Project name
2. Tag (release version)
3. Revision

For all builds containing the same project name AND the same tag AND the same revision number, the input will be an array of archive.Specs. All specs in that array MUST have the same project name, tag, and revision. Because these are all the same, and they must be unique, those three pieces of information can uniquely represent the build group. As such, they are the inputs to this program and determine its output.

Consistent filenames are important because they avoid relying too heavily on ADO templating which is difficult to read and reason about. Because the specs remain immutable throughout the build process, we can calculate values deterministically throughout the build to consistently reference files named by this program.

Usage:
  -project string
    	name of the project
  -tag string
    	tag of artifact
  -revision string
    	revision