Skip to content

Remove the in-repo benchmark: the family's benchmark is schema's data-driven bench - #2

Merged
gafferongames merged 1 commit into
mainfrom
remove-in-repo-benchmarks
Aug 31, 2026
Merged

Remove the in-repo benchmark: the family's benchmark is schema's data-driven bench#2
gafferongames merged 1 commit into
mainfrom
remove-in-repo-benchmarks

Conversation

@gafferongames

Copy link
Copy Markdown
Contributor

Per mas-bandwidth/schema#196 (owner ruling, 2026-08-31), every serialize-family repo's in-repo benchmark is deleted.

The family's benchmark is now schema's data-driven bench: one variant corpus driven through the real generated codecs in all nine languages. A per-repo bench that nobody runs is a re-activation-and-drift surface — it will come back to life some future day carrying stale methodology and stale numbers. Deletion, not archival: git history is the archive. If the generated-vs-hand-written comparison is ever wanted again, it is remeasured deliberately from scratch, never revived from dormant code.

Pure deletions. Correctness tests, conformance and golden vectors, fuzz corpora and example programs are untouched; the full suite is green.

🤖 Generated with Claude Code

…-driven bench

Per mas-bandwidth/schema#196. Deletes bench/serialize/bench/Bench.java,
the Makefile bench target and its variables, and the README benchmark
section with its pasted numbers. Deletion not archival: a dormant bench
is a re-activation and drift surface, and git history is the archive.

🤖 Generated with [Claude Code](https://claude.com/claude-code)
@gafferongames
gafferongames merged commit afbad19 into main Aug 31, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant