gitnexus 1.6.10-rc.21 → 1.6.10-rc.211

package/dist/core/ingestion/process-processor.js

@@ -3,7 +3,7 @@
  *
  * Detects execution flows (Processes) in the code graph by:
  * 1. Finding entry points (functions with no internal callers)
- * 2. Tracing forward via CALLS edges (BFS)
+ * 2. Tracing forward via CALLS edges (DFS)
  * 3. Grouping and deduplicating similar paths
  * 4. Labeling with heuristic names
  *
@@ -19,6 +19,16 @@ const DEFAULT_CONFIG = {
     maxProcesses: 75,
     minSteps: 3, // 3+ steps = genuine multi-hop flow (2-step is just "A calls B")
 };
+/** Zeroed counters, mutated in place by the walk. */
+const emptyTruncation = () => ({
+    truncated: false,
+    entryPointCandidatesDropped: 0,
+    entryPointsUnexplored: 0,
+    walksCutByBudget: 0,
+    tracesDepthCapped: 0,
+    calleesDropped: 0,
+    processesDropped: 0,
+});
 // ============================================================================
 // MAIN PROCESSOR
 // ============================================================================
@@ -27,8 +37,16 @@ const DEFAULT_CONFIG = {
  *
  * This runs AFTER community detection, using CALLS edges to trace flows.
  */
-export const processProcesses = async (knowledgeGraph, memberships, onProgress, config = {}) => {
+export const processProcesses = async (knowledgeGraph, memberships, onProgress, config = {}, 
+/**
+ * Places the program reaches outward — fetch calls and ORM queries, each with
+ * a file and a line (R3-6). Attributed to their enclosing function to form the
+ * sink set; omitted, behaviour is exactly as before.
+ */
+outwardActionSites = []) => {
     const cfg = { ...DEFAULT_CONFIG, ...config };
+    const sinkFunctions = buildSinkFunctionSet(knowledgeGraph, outwardActionSites);
+    const isSink = (nodeId) => sinkFunctions.has(nodeId);
     onProgress?.('Finding entry points...', 0);
     // Build lookup maps
     const membershipMap = new Map();
@@ -38,31 +56,153 @@ export const processProcesses = async (knowledgeGraph, memberships, onProgress,
     const nodeMap = new Map();
     for (const n of knowledgeGraph.iterNodes())
         nodeMap.set(n.id, n);
+    // Declared before Step 1 because `findEntryPoints` has a ceiling of its own
+    // (see `ENTRY_POINT_CANDIDATE_LIMIT`) and reports it through the same record.
+    const truncation = emptyTruncation();
     // Step 1: Find entry points (functions that call others but have few callers)
-    const entryPoints = findEntryPoints(knowledgeGraph, reverseCallsEdges, callsEdges);
+    const entryPoints = findEntryPoints(knowledgeGraph, reverseCallsEdges, callsEdges, truncation);
     onProgress?.(`Found ${entryPoints.length} entry points, tracing flows...`, 20);
     onProgress?.(`Found ${entryPoints.length} entry points, tracing flows...`, 20);
     // Step 2: Trace processes from each entry point
     const allTraces = [];
+    let tracedEntryPoints = 0;
     for (let i = 0; i < entryPoints.length && allTraces.length < cfg.maxProcesses * 2; i++) {
         const entryId = entryPoints[i];
-        const traces = traceFromEntryPoint(entryId, callsEdges, cfg);
+        const traces = traceFromEntryPoint(entryId, callsEdges, cfg, isSink, truncation);
+        tracedEntryPoints = i + 1;
         // Filter out traces that are too short
         traces.filter((t) => t.length >= cfg.minSteps).forEach((t) => allTraces.push(t));
         if (i % 10 === 0) {
             onProgress?.(`Tracing entry point ${i + 1}/${entryPoints.length}...`, 20 + (i / entryPoints.length) * 40);
         }
     }
+    // The loop exits on the TRACE quota, not on running out of entry points, so
+    // the remainder are not "no flows found" — they were never looked at.
+    //
+    // Counted over the list `findEntryPoints` RETURNS, which is already capped at
+    // `ENTRY_POINT_CANDIDATE_LIMIT`; candidates beyond that cap are invisible here
+    // by construction and are reported separately as
+    // `entryPointCandidatesDropped`.
+    truncation.entryPointsUnexplored = entryPoints.length - tracedEntryPoints;
     onProgress?.(`Found ${allTraces.length} traces, deduplicating...`, 60);
     // Step 3: Deduplicate similar traces (subset removal)
-    const uniqueTraces = deduplicateTraces(allTraces);
+    const uniqueTraces = deduplicateTraces(allTraces, isSink);
     // Step 3b: Deduplicate by entry+terminal pair (keep longest path per pair)
     const endpointDeduped = deduplicateByEndpoints(uniqueTraces);
     onProgress?.(`Deduped ${uniqueTraces.length} → ${endpointDeduped.length} unique endpoint pairs`, 70);
-    // Step 4: Limit to max processes (prioritize longer traces)
-    const limitedTraces = endpointDeduped
-        .sort((a, b) => b.length - a.length)
-        .slice(0, cfg.maxProcesses);
+    // Step 4: Limit to max processes — deepest first, but ROUND-ROBIN across
+    // TERMINALS (R2-3).
+    //
+    // Ranking was `sort by length` alone, and the top of that list was one
+    // behaviour described many ways: measured on the reporting repo, eleven of
+    // the top fourteen processes were four entry points crossed with three
+    // terminals of the same date-window utility cluster — `Handle ->
+    // AlignWindowEnd`, `Main -> AlignWindowStart`, `ProcessSymbol ->
+    // ResolveGridIntervalMs`. Genuine call chains, but a reader learns one thing
+    // from fourteen entries.
+    //
+    // Keyed on the TERMINAL, not the entry point. Keying on the entry was tried
+    // first and made it worse — many files declare a `main`, so each was a
+    // distinct entry that round-robin then awarded its own slot, and
+    // `Main -> AlignWindowEnd` went from one row to eight. The repetition was
+    // never in where a flow starts; it is in where flows pile up.
+    //
+    // Depth still orders within a terminal and still leads the list, since
+    // insertion order here is deepest-first. What changes is that no terminal
+    // takes a second slot until every other has had a first. Measured: distinct
+    // terminals in the top 20 went 3 -> 20, and the repo's own domain flows
+    // (`ReconcilePositions -> ...`) moved into the top 4%.
+    //
+    // What this does NOT do, recorded so it does not read as settled: the
+    // reported cause — ranking rewarding fan-in, promoting chains ending in
+    // widely-called helpers — measured FALSE. Those terminals have one caller
+    // each (`alignWindowStart` 1, `validateSymbol` 1). A fan-in discount was
+    // implemented against that hypothesis and moved nothing.
+    //
+    // R3-6 adds one rule ahead of depth: a SINK-terminated trace outranks a
+    // leaf-terminated one. A flow that ends where the program does something —
+    // places an order, writes a row — is what a reader came for; a chain that
+    // ends in a date helper is where control happened to stop. Depth still orders
+    // within each group.
+    //
+    // This is what closed the gap that used to be described here as out of reach:
+    // a business flow could not be a process in its own right, because the walk
+    // emitted only at a leaf, at max depth, or on a cycle, so a flow whose
+    // meaningful endpoint calls onward survived only as whatever leaf it bottomed
+    // out in. Ranking could never fix it — the flow was not a candidate to rank.
+    // It is bounded honestly rather than fully closed: sinks are exactly where
+    // fetch/ORM extraction fires (see `buildSinkFunctionSet`), so a codebase whose
+    // outward calls are not detected as such still sees leaf-terminated traces
+    // only.
+    // DETERMINISM. The comparator below ranks by sink-ness then by depth, and for
+    // two flows equal on both it returned 0. `Array.prototype.sort` is stable, so
+    // a 0 preserves INPUT order — which traces back to `graph.iterNodes()`, i.e.
+    // the order files happened to be inserted. Under `maxProcesses` capping that
+    // decided which `Process` and `STEP_IN_PROCESS` nodes were persisted at all.
+    //
+    // Reproduced: two equal three-step flows with `maxProcesses: 1` select
+    // `handleAlpha`; inserting the identical nodes and CALLS edges in reverse
+    // select `handleBeta`. Same repository, same commit, different persisted
+    // graph — so an incremental run that reorders assembly, or a filesystem that
+    // enumerates differently, silently changes what the tool reports.
+    //
+    // The id is the tiebreak because it is the only totally-ordered, content-derived
+    // key available here; comparing the whole path keeps it stable when two traces
+    // share a terminal.
+    //
+    // MUTATION STATUS, recorded so nobody mistakes this for a verified guard:
+    // removing THIS tiebreak alone fails nothing, because the two dedup sorts
+    // below already impose a total order on the list that reaches here. The
+    // entry-point sort and the dedup sorts each ARE individually verified. This
+    // one is kept as defence in depth — it cannot misbehave (it only makes an
+    // already-deterministic order explicit) and it is what stops a future change
+    // to dedup ordering from silently re-opening the defect.
+    //
+    // Decorated once and sorted on the precomputed key rather than joining inside
+    // the comparator — see `traceOrderKey` for why the key exists at all and
+    // `sortByDepthThenPath` for why it is built exactly once per trace. The sink
+    // test is hoisted for the same reason: it ran twice per comparison.
+    const tracesByTerminal = new Map();
+    const rankedByInterest = (() => {
+        const decorated = endpointDeduped.map((trace) => ({
+            trace,
+            sink: isSink(trace[trace.length - 1] ?? '') ? 1 : 0,
+            key: traceOrderKey(trace),
+        }));
+        decorated.sort((a, b) => b.sink - a.sink || b.trace.length - a.trace.length || compareOrderKeys(a.key, b.key));
+        return decorated.map((d) => d.trace);
+    })();
+    for (const trace of rankedByInterest) {
+        const terminalId = trace[trace.length - 1];
+        if (terminalId === undefined)
+            continue;
+        const existing = tracesByTerminal.get(terminalId);
+        if (existing === undefined)
+            tracesByTerminal.set(terminalId, [trace]);
+        else
+            existing.push(trace);
+    }
+    // Insertion order is deepest-trace-first, so the round-robin visits terminals
+    // in that order too and depth still leads the list.
+    const limitedTraces = [];
+    for (let round = 0; limitedTraces.length < cfg.maxProcesses; round++) {
+        let addedAny = false;
+        for (const traces of tracesByTerminal.values()) {
+            const trace = traces[round];
+            if (trace === undefined)
+                continue;
+            limitedTraces.push(trace);
+            addedAny = true;
+            if (limitedTraces.length >= cfg.maxProcesses)
+                break;
+        }
+        if (!addedAny)
+            break;
+    }
+    // Counted against the DEDUPED input, not `allTraces`: the difference between
+    // those two is deduplication doing its job, which is not truncation.
+    const dedupedAvailable = [...tracesByTerminal.values()].reduce((n, t) => n + t.length, 0);
+    truncation.processesDropped = dedupedAvailable - limitedTraces.length;
     onProgress?.(`Creating ${limitedTraces.length} process nodes...`, 80);
     // Step 5: Create process nodes
     const processes = [];
@@ -113,6 +253,24 @@ export const processProcesses = async (knowledgeGraph, memberships, onProgress,
     const avgStepCount = processes.length > 0
         ? processes.reduce((sum, p) => sum + p.stepCount, 0) / processes.length
         : 0;
+    // Derived last, from the counters the walk accumulated, so a new ceiling added
+    // later cannot be forgotten here — it only has to increment its own counter.
+    //
+    // That claim was wrong when it was written: `findEntryPoints`' 200-candidate
+    // slice was an EXISTING ceiling with no counter, so it was not merely
+    // forgettable, it had already been forgotten. Adding a counter is only half
+    // the discipline; the other half is checking, when you write a line like this,
+    // that every cap in the file actually has one.
+    truncation.truncated =
+        truncation.entryPointCandidatesDropped > 0 ||
+            truncation.entryPointsUnexplored > 0 ||
+            truncation.walksCutByBudget > 0 ||
+            truncation.tracesDepthCapped > 0 ||
+            truncation.calleesDropped > 0 ||
+            truncation.processesDropped > 0;
+    if (truncation.truncated) {
+        logger.debug({ truncation }, 'process-processor: detection was truncated by one or more caps');
+    }
     return {
         processes,
         steps,
@@ -121,6 +279,7 @@ export const processProcesses = async (knowledgeGraph, memberships, onProgress,
             crossCommunityCount,
             avgStepCount: Math.round(avgStepCount * 10) / 10,
             entryPointsFound: entryPoints.length,
+            truncation,
         },
     };
 };
@@ -132,28 +291,37 @@ export const processProcesses = async (knowledgeGraph, memberships, onProgress,
 const MIN_TRACE_CONFIDENCE = 0.5;
 const buildCallsGraph = (graph) => {
     const adj = new Map();
-    for (const rel of graph.iterRelationships()) {
-        if (rel.type === 'CALLS' && rel.confidence >= MIN_TRACE_CONFIDENCE) {
-            if (!adj.has(rel.sourceId)) {
-                adj.set(rel.sourceId, []);
-            }
-            adj.get(rel.sourceId).push(rel.targetId);
-        }
-    }
+    // Field-wise scan (#2680) — whole-graph walk, four fields, no object needed.
+    graph.forEachRelationshipFields((sourceId, targetId, type, confidence) => {
+        if (type !== 'CALLS' || confidence < MIN_TRACE_CONFIDENCE)
+            return;
+        const existing = adj.get(sourceId);
+        if (existing === undefined)
+            adj.set(sourceId, [targetId]);
+        else
+            existing.push(targetId);
+    });
     return adj;
 };
 const buildReverseCallsGraph = (graph) => {
     const adj = new Map();
-    for (const rel of graph.iterRelationships()) {
-        if (rel.type === 'CALLS' && rel.confidence >= MIN_TRACE_CONFIDENCE) {
-            if (!adj.has(rel.targetId)) {
-                adj.set(rel.targetId, []);
-            }
-            adj.get(rel.targetId).push(rel.sourceId);
-        }
-    }
+    graph.forEachRelationshipFields((sourceId, targetId, type, confidence) => {
+        if (type !== 'CALLS' || confidence < MIN_TRACE_CONFIDENCE)
+            return;
+        const existing = adj.get(targetId);
+        if (existing === undefined)
+            adj.set(targetId, [sourceId]);
+        else
+            existing.push(sourceId);
+    });
     return adj;
 };
+/**
+ * How many ranked candidates survive to be traced. Everything below this line
+ * is discarded — see `ProcessTruncationStats.entryPointCandidatesDropped`, the
+ * counter that exists because this cap spent a release being invisible.
+ */
+const ENTRY_POINT_CANDIDATE_LIMIT = 200;
 /**
  * Find functions/methods that are good entry points for tracing.
  *
@@ -164,7 +332,15 @@ const buildReverseCallsGraph = (graph) => {
  *
  * Test files are excluded entirely.
  */
-const findEntryPoints = (graph, reverseCallsEdges, callsEdges) => {
+const findEntryPoints = (graph, reverseCallsEdges, callsEdges, 
+/**
+ * Mutated in place when the candidate cap fires. Optional and reported
+ * through an out-parameter rather than a richer return value, matching
+ * `traceFromEntryPoint`: the `string[]` contract every caller already uses is
+ * unchanged, and a caller that does not care about completeness does not have
+ * to unwrap a counter to ask for entry points.
+ */
+truncation) => {
     const symbolTypes = new Set(['Function', 'Method']);
     const entryPointCandidates = [];
     for (const node of graph.iterNodes()) {
@@ -191,8 +367,12 @@ const findEntryPoints = (graph, reverseCallsEdges, callsEdges) => {
             entryPointCandidates.push({ id: node.id, score, reasons });
         }
     }
-    // Sort by score descending and return top candidates
-    const sorted = entryPointCandidates.sort((a, b) => b.score - a.score);
+    // Sort by score descending, then by node id. Ties on score are common — most
+    // candidates share a heuristic bucket — and a stable sort resolves them by
+    // `iterNodes()` order, so which entry points survive the `slice` below became
+    // a function of file insertion order. See the determinism note on
+    // `rankedByInterest`.
+    const sorted = entryPointCandidates.sort((a, b) => b.score - a.score || (a.id < b.id ? -1 : a.id > b.id ? 1 : 0));
     // DEBUG: Log top candidates with new scoring details
     if (sorted.length > 0 && isDev) {
         logger.info(`[Process] Top 10 entry point candidates (new scoring):`);
@@ -204,24 +384,66 @@ const findEntryPoints = (graph, reverseCallsEdges, callsEdges) => {
             logger.info(`     score: ${c.score.toFixed(2)} = [${c.reasons.join(' × ')}]`);
         });
     }
-    return sorted
-        .slice(0, 200) // Limit to prevent explosion
-        .map((c) => c.id);
+    // Limit to prevent explosion — and SAY SO. This is the ceiling that decides
+    // how much of a repository process detection ever looks at: on anything with
+    // more than 200 scoring candidates the reported flows are a sample of the
+    // top-ranked ones, and every downstream count (`entryPointsFound`,
+    // `entryPointsUnexplored`) is computed over the survivors, so none of them can
+    // see what was cut here.
+    if (truncation !== undefined && sorted.length > ENTRY_POINT_CANDIDATE_LIMIT) {
+        truncation.entryPointCandidatesDropped = sorted.length - ENTRY_POINT_CANDIDATE_LIMIT;
+    }
+    return sorted.slice(0, ENTRY_POINT_CANDIDATE_LIMIT).map((c) => c.id);
 };
 // ============================================================================
-// HELPER: Trace from entry point (BFS)
+// HELPER: Trace from entry point (DFS)
 // ============================================================================
 /**
- * Trace forward from an entry point using BFS.
+ * Trace forward from an entry point using DEPTH-first search.
  * Returns all distinct paths up to maxDepth.
  */
-const traceFromEntryPoint = (entryId, callsEdges, config) => {
+// Exported for tests ONLY: traversal order is the whole behaviour here, and it
+// is unobservable through `processProcesses` because `findEntryPoints` supplies
+// several starting points — a deep chain gets traced from inside it regardless
+// of order, so a test at that level passes under either traversal and pins
+// nothing.
+export const traceFromEntryPoint = (entryId, callsEdges, config, 
+/**
+ * Functions that reach outward (R3-6). A trace also ENDS here, even though
+ * the walk continues past it: a flow whose meaningful endpoint calls onward
+ * was otherwise never a candidate, only ever surviving as whatever leaf it
+ * bottomed out in.
+ */
+isSink = () => false, 
+/**
+ * Mutated in place when a ceiling fires. Optional so the many direct callers
+ * in tests are unchanged, and because a caller that does not care about
+ * completeness should not have to invent a counter to ask for a trace.
+ */
+truncation) => {
     const traces = [];
-    // BFS with path tracking
-    // Each queue item: [currentNodeId, pathSoFar]
-    const queue = [[entryId, [entryId]]];
-    while (queue.length > 0 && traces.length < config.maxBranching * 3) {
-        const [currentId, path] = queue.shift();
+    // DEPTH-first, not breadth-first. Each stack item: [currentNodeId, pathSoFar].
+    //
+    // The walk stops after a fixed NUMBER of traces, so the traversal order
+    // decides which traces those are. Breadth-first reaches every shallow
+    // terminal before any deep one, so the quota filled with the shortest paths
+    // in the graph and the walk stopped — `maxTraceDepth` was never approached.
+    // Measured on a 75k-node repo: of 300 processes, none exceeded 7 steps and
+    // 90% were 3-4, so a multi-hop business flow (signal → order → exit) had no
+    // process that could represent it, and `query` could only rank the mechanical
+    // pairs that did exist.
+    //
+    // Depth-first descends to a terminal first, so the same quota is spent on
+    // paths worth keeping. Cost is unchanged — same budget, same cycle guard,
+    // same `maxTraceDepth` ceiling — only the ORDER of exploration differs, and
+    // the caller already sorts by length and dedupes by endpoint, so it was
+    // always asking for the deepest traces this walk could give it.
+    // A LIFO stack, not a queue — the name followed the traversal when this was
+    // breadth-first and was left behind by the change to depth-first.
+    const stack = [[entryId, [entryId]]];
+    const traceBudget = config.maxBranching * 3;
+    while (stack.length > 0 && traces.length < traceBudget) {
+        const [currentId, path] = stack.pop();
         // Get outgoing calls
         const callees = callsEdges.get(currentId) || [];
         if (callees.length === 0) {
@@ -231,19 +453,41 @@ const traceFromEntryPoint = (entryId, callsEdges, config) => {
             }
         }
         else if (path.length >= config.maxTraceDepth) {
-            // Max depth reached - save what we have
+            // Max depth reached - save what we have. The trace is kept, but it is a
+            // PREFIX of a longer flow rather than a flow that ended, and only this
+            // counter distinguishes the two downstream.
+            if (truncation !== undefined)
+                truncation.tracesDepthCapped++;
             if (path.length >= config.minSteps) {
                 traces.push([...path]);
             }
         }
         else {
+            // A SINK ends a trace without ending the walk (R3-6). Emitting here is
+            // what lets `placeOrder` be an endpoint while `placeOrder -> formatDate`
+            // still exists as its own longer trace; the two answer different
+            // questions and neither should suppress the other.
+            if (isSink(currentId) && path.length >= config.minSteps) {
+                traces.push([...path]);
+            }
             // Continue tracing - limit branching
             const limitedCallees = callees.slice(0, config.maxBranching);
+            if (truncation !== undefined && callees.length > limitedCallees.length) {
+                truncation.calleesDropped += callees.length - limitedCallees.length;
+            }
             let addedBranch = false;
-            for (const calleeId of limitedCallees) {
+            // PUSHED IN REVERSE so the stack POPS them in source order. `slice`
+            // selects the first N callees while `pop()` takes the last pushed, so
+            // without this the walk spends its trace budget on the LAST-declared
+            // branch first: for `main() { init(); loadConfig(); run(); shutdown(); }`
+            // it explores `shutdown` first and can exhaust the quota before reaching
+            // `init` — dropping the earliest steps of a flow, which is the opposite
+            // of what a process is meant to describe. Selecting the first N and then
+            // exploring them last-first was simply inconsistent.
+            for (const calleeId of [...limitedCallees].reverse()) {
                 // Avoid cycles
                 if (!path.includes(calleeId)) {
-                    queue.push([calleeId, [...path, calleeId]]);
+                    stack.push([calleeId, [...path, calleeId]]);
                     addedBranch = true;
                 }
             }
@@ -253,8 +497,130 @@ const traceFromEntryPoint = (entryId, callsEdges, config) => {
             }
         }
     }
+    // A silently truncating cap reads as "this is everything", which is the same
+    // class of confident-empty answer this work is about. The repo already sets
+    // this precedent for `dispatchFanoutSkipped` and
+    // `propertyDispatch.skippedKeys`.
+    //
+    // The debug line stays for the per-entry-point detail (which entry, how many
+    // branches); the counter is what escapes to a CONSUMER. A log nobody has
+    // enabled is not a disclosure.
+    if (stack.length > 0) {
+        if (truncation !== undefined)
+            truncation.walksCutByBudget++;
+        logger.debug({ entryId, traceBudget, unexploredBranches: stack.length }, 'process-processor: trace budget exhausted; unexplored branches remain for this entry point');
+    }
     return traces;
 };
+const CALLABLE_SINK_LABELS = new Set([
+    'Function',
+    'Method',
+    'Constructor',
+]);
+/**
+ * Functions that DO something outward — issue a request, run a query (R3-6).
+ *
+ * The missing layer behind "business flows are never processes". A trace is
+ * only emitted at a node with NO outgoing calls, so a real flow — scan, score,
+ * arm, place the order — is always a PREFIX of some longer chain that runs on
+ * into date helpers and formatters, and can never be a process in its own
+ * right. Ending a trace somewhere meaningful needs a notion of an endpoint that
+ * is not a leaf, and the walk had none.
+ *
+ * GitNexus already knew where the program reaches outward: the parse phase
+ * collects fetch calls and ORM queries carrying `filePath` + `lineNumber`. Those
+ * facts only ever produced FILE-level edges (`File -[FETCHES]-> Route`), which
+ * is too coarse to end a trace on — every function in a file containing one
+ * would qualify. Attributing each site to the function whose range CONTAINS it
+ * turns the same facts into the function-level signal the walk needs, with no
+ * new extraction, no new relation pair, and no schema change.
+ *
+ * Innermost wins: a nested closure that performs the call is the sink, not the
+ * outer function that merely spans it.
+ */
+export function buildSinkFunctionSet(graph, sites) {
+    const sinks = new Set();
+    if (sites.length === 0)
+        return sinks;
+    const byFile = new Map();
+    for (const node of graph.iterNodes()) {
+        if (!CALLABLE_SINK_LABELS.has(node.label))
+            continue;
+        const props = node.properties;
+        if (typeof props.filePath !== 'string')
+            continue;
+        if (typeof props.startLine !== 'number' || typeof props.endLine !== 'number')
+            continue;
+        const entry = { id: node.id, start: props.startLine, end: props.endLine };
+        const list = byFile.get(props.filePath);
+        if (list === undefined)
+            byFile.set(props.filePath, [entry]);
+        else
+            list.push(entry);
+    }
+    for (const site of sites) {
+        const candidates = byFile.get(site.filePath);
+        if (candidates === undefined)
+            continue;
+        let best;
+        for (const c of candidates) {
+            if (site.lineNumber < c.start || site.lineNumber > c.end)
+                continue;
+            if (best === undefined || c.end - c.start < best.end - best.start)
+                best = c;
+        }
+        if (best !== undefined)
+            sinks.add(best.id);
+    }
+    return sinks;
+}
+// ============================================================================
+// HELPER: Deterministic trace ordering
+// ============================================================================
+/**
+ * Total-order key for a trace — the TIEBREAK every trace sort in this file uses.
+ *
+ * NUL is the separator, and that is load-bearing rather than cosmetic. Node ids
+ * embed file paths and a path may contain a SPACE, so a space-joined key is
+ * ambiguous in exactly the way `traceKey`'s unpadded `->` join was (#2894):
+ * `['A B', 'C']` and `['A', 'B C']` both render as `A B C`, the comparator
+ * returns 0, and `Array.prototype.sort` — being stable — falls straight back to
+ * the input order the tiebreak exists to remove. Two of the three trace sorts
+ * here joined on a space and had that hole; all three now share this key.
+ *
+ * NUL cannot occur in a node id (not in a POSIX path, not in a source
+ * identifier) and sorts below every character that can, so joining on it is
+ * order-equivalent to comparing the two arrays element by element. That
+ * equivalence is what makes it a drop-in for the space-joined keys: on any
+ * corpus without the collision above the resulting order is IDENTICAL, which is
+ * asserted directly in `process-processor.test.ts`.
+ */
+const traceOrderKey = (trace) => trace.join('\u0000');
+/** Lexicographic compare of two `traceOrderKey` results. */
+const compareOrderKeys = (a, b) => (a < b ? -1 : a > b ? 1 : 0);
+/**
+ * Sort traces deepest-first, breaking ties on the path key.
+ *
+ * A Schwartzian transform: the key is built ONCE per trace and the comparator
+ * only compares two strings. Written the obvious way — `a.join(sep) <
+ * b.join(sep)` inside the comparator — each comparison allocates up to FOUR
+ * joined strings, so an O(n log n) sort performs O(n log n) joins of
+ * O(depth x id-length) characters each.
+ *
+ * `n` is bounded here (`ENTRY_POINT_CANDIDATE_LIMIT` entry points x the
+ * per-entry trace budget), so the cost is small and once-per-analyze: measured
+ * at the ceiling, 23,851 comparisons performed 70,524 joins, and end-to-end
+ * `processProcesses` at 80,000 functions / 640k CALLS went 456 -> 555 ms. It is
+ * fixed anyway because it is the same allocation-in-the-comparator shape that
+ * was just removed from `deduplicateTraces` below (deep_chain 1233 -> 102 ms),
+ * and leaving one instance of it standing next to that comment invites the next
+ * one.
+ */
+const sortByDepthThenPath = (traces) => {
+    const decorated = traces.map((trace) => ({ trace, key: traceOrderKey(trace) }));
+    decorated.sort((a, b) => b.trace.length - a.trace.length || compareOrderKeys(a.key, b.key));
+    return decorated.map((d) => d.trace);
+};
 // ============================================================================
 // HELPER: Deduplicate traces
 // ============================================================================
@@ -262,21 +628,65 @@ const traceFromEntryPoint = (entryId, callsEdges, config) => {
  * Merge traces that are subsets of other traces.
  * Keep longer traces, remove redundant shorter ones.
  */
-const deduplicateTraces = (traces) => {
+export const deduplicateTraces = (traces, 
+/** See `buildSinkFunctionSet` — a sink-terminated trace survives subsumption. */
+isSink = () => false) => {
     if (traces.length === 0)
         return [];
-    // Sort by length descending
-    const sorted = [...traces].sort((a, b) => b.length - a.length);
+    // Sort by length descending, then by path, so equal-length traces have a
+    // total order instead of inheriting graph-traversal order (determinism note
+    // on `rankedByInterest`). Which of two equal traces is kept as the
+    // representative is otherwise decided by insertion order.
+    const sorted = sortByDepthThenPath(traces);
     const unique = [];
+    // Keys for `unique`, built ONCE per surviving trace rather than once per
+    // COMPARISON. The join used to sit inside the `some()` callback below, so
+    // every already-kept trace had its key rebuilt from scratch against every
+    // candidate — O(T*U) joins of O(depth * id-length) characters, and it is that
+    // allocation, not the substring scan, that dominates the pass.
+    //
+    // Nothing about breadth-first search made that safe; it only kept the cost
+    // small by keeping traces short. Measured on this repo, the walk went from an
+    // average of 4.3 steps to 9.4 when D1 made it depth-first, which roughly
+    // doubles both the number of surviving traces and the length of every key —
+    // so the same quadratic that was affordable under BFS is ~6x the work under
+    // DFS. Hoisting the join removes the multiplication entirely.
+    const uniqueKeys = [];
     for (const trace of sorted) {
+        // A SINK-TERMINATED trace is never redundant (R3-6), even though it is by
+        // definition a prefix of the longer chain that runs on past the sink into
+        // helpers. That is the whole shape of a business flow — scan, score, arm,
+        // PLACE THE ORDER — so subsuming it here is exactly what kept such flows
+        // from ever being processes. Emitting one at the walk and deleting it one
+        // step later would have been a no-op fix.
+        const terminal = trace[trace.length - 1];
+        // PADDED with the separator on both ends, so `includes` can only match whole
+        // steps (#2894). Unpadded, the test is not anchored to a step boundary and a
+        // match may begin in the MIDDLE of a node id:
+        //
+        //   'X->AA->B'.includes('A->B')   ->  true
+        //
+        // which discards `A -> B` as redundant against a chain `A` is not a step of
+        // at all. Measured inert on every corpus tried — the collision needs one id
+        // to be a strict suffix of another, which real ids
+        // (`Function:<path>:<name>`) do not produce — but the predicate did not mean
+        // what the surrounding code says it means, and this is a function whose
+        // entire job is deciding what to delete.
+        //
+        // Note the encoding assumes `->` never appears IN a node id. A C++
+        // `operator->` would defeat the join regardless of padding; out of scope
+        // here, but the assumption is real.
+        const traceKey = `->${trace.join('->')}->`;
+        if (terminal !== undefined && isSink(terminal)) {
+            unique.push(trace);
+            uniqueKeys.push(traceKey);
+            continue;
+        }
         // Check if this trace is a subset of any already-added trace
-        const traceKey = trace.join('->');
-        const isSubset = unique.some((existing) => {
-            const existingKey = existing.join('->');
-            return existingKey.includes(traceKey);
-        });
+        const isSubset = uniqueKeys.some((existingKey) => existingKey.includes(traceKey));
         if (!isSubset) {
             unique.push(trace);
+            uniqueKeys.push(traceKey);
         }
     }
     return unique;
@@ -292,8 +702,10 @@ const deduplicateByEndpoints = (traces) => {
     if (traces.length === 0)
         return [];
     const byEndpoints = new Map();
-    // Sort longest first so the first seen per key is the longest
-    const sorted = [...traces].sort((a, b) => b.length - a.length);
+    // Sort longest first so the first seen per key is the longest; the path
+    // tiebreak makes "which of two equal-length traces represents this endpoint
+    // pair" independent of insertion order.
+    const sorted = sortByDepthThenPath(traces);
     for (const trace of sorted) {
         const key = `${trace[0]}::${trace[trace.length - 1]}`;
         if (!byEndpoints.has(key)) {