Cod3xDocs
reference

Historical Evidence Index

A map from Cod3x engineering rules back to the repository history that motivated or tested them.

Cod3x is deliberately willing to cite ugly intermediate history.

This is not a changelog. It is an index of commits that are especially useful when researching why a rule exists, how an optimization evolved, or what assumption failed.

The initial corpus comes from complete Git bundles of St4sh, DreamScripts, Starwind Builder, and Rubic0n. Hashes below refer to those repositories. Every receipt is a direct link so a reader or agent can inspect the change instead of manually hunting for it.

S3ctors S3cret St4sh

CommitWhat it is useful for
4f488e47Cod3x parser fast path; less allocation and up to roughly 30% reported improvement
940fe460Batch-scoped caching of expensive C++/engine-derived condition inputs
bfe28c01Derived resolver state was not invalidated with its source cache
b148ee31Generation/epoch identity for stale deferred work
c100e5ebMaking corrupted presence collection fatal rather than continuing ambiguously
e3e21b64core.quit() does not terminate local control flow immediately
f834f17eReal engine content made a supposedly-required pathgrid optional
3649154cEngine userdata was treated like an ordinary Lua table
fbe9f127“same playlist” fast path skipped dependent silence/playback state
8f3e075dopenmw.types was incorrectly modeled as available in menu scripts
c3e70aa5Hoisting stable gameSelf.id in a combat hot path
02a00845Stabilizing target positions by avoiding repeated bounding-box queries
de85472bLock-on rewrite focused on raycasts/runtime operations rather than Lua cosmetics
ae7561dPhysics and rendering raycasts answered different questions

Design genealogies

CommitWhat it is useful for
f6f73098First internal S3lf wrapper around repeated OpenMW object interpretation
9ffbd3f9Standalone S3lf release
257aabfdMigration of S3lf into H3lp Yours3lf
cd16c7ccNamed nullFunction primitive in H3
d40f379cS3maphore promotion from swappable handlers to H3 StateMachine
3b1f8a75CamHelper migrated from T4rg3t5 into H3
bf230995ImageAtlas added to H3
224d20c5H4ND's first ImageAtlas consumer
51cfc16cDedicated ImageAtlas next-frame operation
8fa3d40aImageAtlas current-tile update fix
8d8f2025CamHelper hot-path rewrite
dcf5bf21S3maphore's move from update polling to event-driven resolution
80b31628Keeping PlaylistRules cache machinery private
07df2d73First caching helper for PlaylistRules
186e469dDeploying the caching helper across the rules
32153de8Event-driven clearing of combat-target caches
a1a431b1Per-target cache invalidation while preserving dynamic-stat caching
73038e4aClearing combat caches when an actor leaves combat
c077f20fCorrecting cache clearing at combat start
e485c601Local track-change handler registration instead of direct internal events
57eb4b69Bounded actor queue for combat scanning
8cad4c1fDynamic frame-time batch sizing experiment
17cfca78Reverting dynamic batch sizing in favor of a fixed bound
c967a3d2Capped actor scheduler with a target revisit interval
7cb5a731SSS utilities promoted into H3 after duplication was proven
a9afe80aisOpenMW promoted into H3
d6a45c62S3maphore deletion after H3 took ownership of isOpenMW
30ca0dbeDedicated clear operation extracted from repeated consumers
bf0b6505H3 clear with a Rubic0n fast path and Lua fallback
7a2580caProtectedTable checks for adding new values
94f60f76ProtectedTable diagnostics for failed inputs
e6b7f1a2ProtectedTable transparently indexes settings
4d78a2d5ProtectedTable metatable protection
5cbc5ed6ProtectedTable promoted as an interface
36ef9e14ProtectedTable state made directly readable and writable
e0122c97Existing storage sections accepted by ProtectedTable
5d2ece29Writable storage groups supported by ProtectedTable
9d5c7c2dMore precise cached-setting invalidation
27c5eb07Explicitly disabling ProtectedTable subscriptions
5f051eeeProtectedTable hot-path allocation removal
2499548eProtectedTable lookup optimization
e6a18a18ProtectedTable optimization and annotation pass
e3501182PlaylistRules constructor removed in favor of direct state ownership
cec95881Removing an unnecessary S3maphore constructor/object boundary
382df1e0Flattening CellPresence into PlaylistState
76358648Removing the redundant staticList representation
b202141cAdding per-cell presence scope
8d574d40Coroutine-driven global presence sweep
d7f38bb5Large StaticCollection optimization pass
342d77ddEvent-driven cell and combat-target tracking
5d608a26StaticCollection switched to cached string primitives
5b1537b5Sorting playlists at load time for deterministic resolution

DreamScripts

CommitWhat it is useful for
219c6e4bLegitimate protected execution at an arbitrary script/host boundary
28062542Cache the loaded module value rather than an intermediate compiled chunk
046694b6Centralize stale event-registration cleanup at the load/reload lifecycle boundary
c8737a3A protected error path that failed to print the traceback
0e7eb8dThe same lesson, learned loudly enough to deserve a second receipt
3523eafHypothesis: numeric GMSTs may not need conversion
39ba4b1Immediate revert when that hypothesis broke behavior
1e3bbc4Actual finding: the relevant GMST values arrived as userdata
1db40dbInitial FFI-string optimization experiment
b575991, 7ef025c, d171ef4Reverting the FFI-string experiment while isolating the failure
9f569d0Finding that the strings were not the actual bug; an argument was
ad910b8Reapplying the mechanism after the real failure was understood

Starwind Builder

CommitWhat it is useful for
a3c598bHistorical runtime context probing helper based on failed require calls
fe57c24Introduction of the protectedTable abstraction from earlier CHIM work
8d83601Adding caching to the magic storage/proxy surface
1d12418Expanding the proxy into a shared manager interface
25c3661Shadow cache invalidation broke when subscription behavior was overridden
2b94a9fInitial lock-on/camera manager architecture
f90baecLock-on texture accidentally recreated continuously

Starwind is particularly useful because many ideas later found cleaner forms in H3 or current St4sh code. Treat it as evolutionary evidence, not current doctrine.

Rubic0n

Rubic0n includes substantial LuaJIT/OpenResty upstream history. The commits below are selected because they illuminate DreamWeave runtime investigation relevant to OpenMW workloads.

CommitWhat it is useful for
7737682eOptional GC telemetry counters
c50281e9GC allocation telemetry expansion
aa5a699fExact-size userdata allocation cache experiment
b57f7756Userdata finalizer lookup cache experiment
edb8fe9cWholesale revert of the userdata-cache experiments
d1e1a8fcDirect C userdata finalizer path
3f9e260cNon-resurrecting C finalizer mode
36a21256GC pacing tuned for OpenMW-style userdata churn
c7f2c204Gated batched direct C finalizers
6797075eFixing GC accounting for batched finalizers
75a9a247Pacing batched finalizers by GC budget
8e6520a7Optional PC position from jit.util.tracesnap() for trace analysis
6c23f555Making the sandbox bypass an explicit opt-in build facility
a04b64a5Experimental paged allocator front end
8acb8986Vector-userdata allocation harnesses

How to use this index

When a Cod3x rule feels oddly specific, do not merely repeat the rule.

Open the relevant commit and ask:

  1. what the old code believed;
  2. what runtime/engine behavior contradicted it;
  3. whether the fix changed correctness, performance, or both;
  4. whether the lesson still applies to current OpenMW;
  5. whether a newer implementation refined the lesson again.

The point of history is not authority.

The point is reproducible context.