The read spine demands unconditionally: the tournament is over and the demand won - #76
Merged
Merged
Conversation
…e demand won SERIALIZE_READ_SPINE_DEMAND landed default-off in #74 on remark evidence alone; the tournament measured it. Armed alongside the schema emitter's matching read demand (Apple clang 21, -O3, schema bench, tournament-air era) the read rows swept with zero regressions — batch read +68%, rigidbody_at_rest read +244%, rigidbody_moving read +232%, inputpacket read +187%, testdata read +109% — and the confirmation-air pass (r2, window 0.0%, controls clean) reconfirmed the sweep. Per the feature lifecycle ruling, the fastest features win and get turned on: the read spine — BitReader::ReadBits/ReadAlign and ReadStream's six per-field methods, the exact mirror of the demanded write set — is now spelled SERIALIZE_ALWAYS_INLINE unconditionally, the switch macro and its #if plumbing are deleted, and the rationale comment states the measured verdict plainly. The compressed-float and string bodies stay undemanded, deliberately: their cost is the work, not the call. Tests pass Release, Debug, and both CI C++03 legs; with no switch left there is only one mode to test. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
SERIALIZE_READ_SPINE_DEMAND(landed default-OFF in #74) ships: the read spine —BitReader::ReadBits/ReadAlignandReadStream's six per-field methods, the exact mirror of the demanded write set — is now spelledSERIALIZE_ALWAYS_INLINEunconditionally. The switch macro, its#ifplumbing, and its docs are deleted; the rationale comment now states the measured verdict plainly. The compressed-float and string bodies stay undemanded, deliberately — their cost is the work, not the call.Why
The lifecycle ruling: #ifdef features are tournaments, not settlements — the fastest features win and get turned on, the losers get removed. This one won.
Measured (Apple clang 21, -O3, schema bench, Air-on-power quiet windows, armed alongside the schema emitter's matching read demand):
The confirmation-air pass (r2, window 0.0%, controls clean) reconfirmed the sweep. Ledgers: schema
bench/results/tournament-air/(on main) and schema branchconfirmation-air.Verification
__SIZEOF_INT128__): ALL TESTS PASS🤖 Generated with Claude Code