Sort JSON Keys for Cleaner Configuration Diffs

JSON object order is not usually meaningful to a parser, but it is highly meaningful to a reviewer. When identical keys appear in a stable order, a pull request shows the real change instead of a wall of movement. Key sorting is especially useful for generated configuration, API fixtures, localization files, and policy documents that are edited by multiple people. ## Decide where ordering belongs Sort object keys when the file is treated as a map. Do not sort arrays simply because they contain objects. Array order often represents execution order, preference order, or priority. Changing it can alter behavior. A trustworthy sorter walks through objects recursively while preserving arrays and the order of their elements. The CalcSnippets JSON Key Sorter applies this rule locally and renders indented output. Use it before a review to see the structure clearly. Keep the original file available, and make the formatting-only commit separate from a functional change when possible. Reviewers should not have to infer whether a moved key also changed value. ## Make generated output deterministic If an application writes JSON for a repository, report, or cache key, define its serialization rule in code rather than relying on insertion order from unrelated operations. Deterministic output improves cache behavior, makes snapshots stable, and keeps diffs small. It also makes a checksum more meaningful because equivalent input is less likely to produce a different representation. Sorting does not make JSON valid, secure, or schema-compliant. Parse it first, preserve numeric and string types, and validate against the consuming service's contract. Be careful with duplicate keys: JSON parsers often retain only the final value, which can hide a source mistake. A key sorter cannot recover a duplicate that was already discarded by parsing. ## Review changes at the right level For a configuration change, inspect the value, scope, default, and rollback plan. Sorting is a tool for visibility, not a substitute for that reasoning. Avoid applying automated formatting to a large unrelated directory just before a release. A focused diff, a deterministic serializer, and a test that loads the file turn JSON ordering into a practical reliability improvement.