The matching engine (pkg/lx, 45.5K LOC incl. tests) and the deterministic
core layered on it (pkg/dex) are one matcher spread across two packages
(dex imported lx in 12 files; lx imported neither). Fold lx into dex and
drop the "lx" name entirely — pkg/dex is now the single matcher package.
Pure rename/fold, ZERO logic change. Behavior proven identical: the merged
pkg/dex runs 457 top-level tests (431 former-lx + 26 former-dex) and 489
test funcs (468 + 21), all green — matching, batch (BatchMatchEqualsSequential,
run-to-run determinism, tile boundary), GPU/CPU parity, truncation, and CL
swap math all pass unchanged.
What moved:
- git mv all 101 pkg/lx/*.go + 1 .c into pkg/dex/ (package lx -> package dex).
- lx's swap_test.go -> swap_math_test.go (only filename collision with dex's
existing swap_test.go).
- Rewrote imports repo-wide: 54 pkg/lx import lines removed; 42 external
importers (api, client, consensus, dchain, dpdk, lxgpu, marketdata, test/,
tests/) repointed to pkg/dex with lx.X -> dex.X; the 12 in-package dex files
had lx.X -> bare X. 717 selector lines rewritten. Stale pkg/lx comment refs
updated (pkg/lxgpu left intact).
Collisions resolved (2, both package-level vars):
- ErrNoLiquidity: lx's pool-level error (dexError, used only in swap.go, never
referenced via qualifier) renamed ErrPoolNoLiquidity to match the existing
ErrPool* convention; dex's router-level ErrNoLiquidity (public, has errors.Is
tests) kept.
- priceMultiplierBig: identical after fold (both big.NewInt(PriceMultiplier));
deduped by deleting dex's copy.
Not folded — pkg/orderbook: it is NOT the GPU seam (that is orderbook_gpu.go,
which lives in the engine and folded in here). pkg/orderbook is a zero-importer,
placeholder-cgo island whose 13 core domain identifiers (Order, OrderBook, Trade,
OrderType, OrderStatus, PriceLevel, Buy, Sell, Market, Limit, Filled,
PartiallyFilled, NewOrderBook) collide incompatibly with the engine's. Folding
it verbatim breaks the build; folding via rename reintroduces the exact parallel
duplication this fold removes. Left for owner decision (recommend deletion as
dead code) — see refactor notes.
Follow-on (separate repos, not touched): 19 files under luxfi/dex import
pkg/lx externally — lux/precompile/dex, lux/precompile-dex/dex,
lux/consensus/examples/01-simple-bridge, lux/_consensus-v12537/examples,
lux/cex/pkg/engine — repoint to pkg/dex after this lands.
Verified: CGO_ENABLED=0 GOWORK=off go build ./... clean; go vet ./... clean;
go test ./... -count=1 all green (22 ok / 0 fail).
Co-authored-by: Hanzo Dev <[email protected]>
dexd is now the ONLY command in the repo. One binary, two run modes:
dexd run as a luxd rpcchainvm PLUGIN. luxd execs the plugin from
dexd plugin its PluginDir with NO positional args (engine address arrives
via the runtime.EngineAddressKey env var) and probes the
version with `dexd version`. No-subcommand == plugin is
REQUIRED to preserve that contract — byte-for-byte the old
cmd/dchain behavior (rpc.Serve over the plugin transport).
dexd run run the STANDALONE D-Chain venue: open the on-disk zapdb,
dexd standalone serve the frozen dex_* surface over ZAP, and drive the
consensus sealer loop (WaitForEvent->BuildBlock->Verify->
Accept). -addr/-db flags preserved from the old cmd/dvenue.
Relocated logic (engine code NOT duplicated — pkg/dchain + pkg/lx untouched):
cmd/dchain/{main,runner} -> internal/plugin (plugin.Run)
cmd/dvenue/{main,engine_*} -> internal/standalone (standalone.Run + engineName)
cmd/dexd/main.go = thin run-mode dispatcher
Removed cmd dirs (12): dchain, dvenue, demo, dex-api-server, dex-server,
dexfillproof, dexseed, fix-gateway, gateway, monitor, persistent-server,
price-feed. Removed committed build artifacts: gateway, dex-server-linux.
Build/ship ONLY dexd:
- Dockerfile: canonical CPU image (CGO=0, ghcr.io/luxfi/dex) running
`dexd run`; no GPU/MLX required.
- Dockerfile.dvenue: GPU operator bundle repointed cmd/dvenue -> cmd/dexd.
- Deleted Dockerfile.{gateway,devnet,minimal,dvenue-cpu} and
docker/{Dockerfile,Dockerfile.cuda,backend/Dockerfile,compose-simple.yml}.
- Makefile rewritten to build/run/ship dexd only.
- docker.yml: dropped the dex-gateway image; dex-venue-image.yml: dvenue->dexd.
- Deleted legacy scripts that built removed/long-gone cmds; fixed deploy.sh +
dev-setup.sh; .gitignore ignores root build binaries.
Also included (coupled, environment-applied, settled): the in-progress
clob -> dex/orderbook terminology refactor across pkg/* + sdk/* + tests/*
(identifier + file renames, e.g. RegisterCLOB->RegisterDEX,
source_clob.go->source_orderbook.go). The relocated standalone venue calls the
renamed API, so it lands in the same consistent snapshot.
Verified (GOWORK=off CGO_ENABLED=0): `go build ./cmd/dexd`, `go build ./...`,
`go vet ./...` all exit 0; `go test ./pkg/dchain` (renamed RegisterDEX socket
path) green. GPU/cgo path not compiled locally (no CUDA); its rename mirrors
the verified nocgo twins and preserves the C ABI symbol names.
Chaos tests (Jepsen-style):
- OrderMatchingAtomicity: verify buy+sell match is atomic
- CLOBPriceTimePriority: 100 orders verify FIFO at same price
- PartitionedMatchingEngine: no phantom fills after partition
- ZMQMessageDelivery: no lost/duplicate orders under transport drops
- NATSFailover: order flow survives disconnect/reconnect
- ConcurrentCancelFill: race cancel vs fill, exactly one wins
- MarketOrderSlippage: VWAP within 10% of best ask
- GatewayLoadShedding: 10K orders, no panics, consistent book
Benchmarks:
- OrderInsert: ~773K ops/sec, 1294ns/order
- OrderMatch: ~15K matches/sec, 66us/match
- CancelOrder: ~608K ops/sec, 0 allocs
- BatchOrders: ~442K orders/sec (1000-order batches)
- ConcurrentMatching: ~16K markets/sec across 100 parallel books
Known: engine has data races in AddOrder.LastOrderID (non-atomic
write after atomic.Add) exposed by concurrent ZMQ/LoadShedding tests.
zeekayandHanzo Dev