@sema-agent/core 5.40.0 → 5.42.0

package/dist/tools/fs/search.js

@@ -1,9 +1,11 @@
 import { delimitUntrusted } from "../../core/untrusted-text.js";
+import { MAX_EXEC_OUTPUT_BYTES } from "../../core/exec-output-tail.js";
 import { hasBinaryExtension, isAbsolutePathForm } from "./safety.js";
 const WALK_MAX_FILES = 5000;
 const WALK_MAX_DEPTH = 32;
 const GREP_DEFAULT_CAP = 250;
 const FILE_MAX_BYTES = 5 * 1024 * 1024;
+export const JS_GREP_OUTPUT_MAX_BYTES = MAX_EXEC_OUTPUT_BYTES;
 export const DEFAULT_IGNORE_DIRS = new Set([
     "node_modules", ".git", ".hg", ".svn", ".bzr", ".jj", ".sl", "build", "dist", "out", "target",
     ".dart_tool", ".pub-cache", ".next", ".nuxt", ".gradle", ".idea", ".vscode", "Pods", ".venv", "venv",
@@ -840,6 +842,27 @@ export async function jsGrep(env, root, p, signal, guards, deny, denyOut) {
     const off = Math.min(100_000, Math.max(0, Math.floor(p.offset ?? 0)));
     const collectCap = cap + off;
     const out = [];
+    const outputMaxBytes = guards?.outputMaxBytes ?? JS_GREP_OUTPUT_MAX_BYTES;
+    let outBytes = 0;
+    let outputTruncated = false;
+    let rowsBeforeWindow = 0;
+    const pushRow = (row) => {
+        if (outputTruncated)
+            return;
+        if (rowsBeforeWindow < off) {
+            rowsBeforeWindow++;
+            return;
+        }
+        if (out.length >= cap)
+            return;
+        const size = Buffer.byteLength(row, "utf8") + 1;
+        if (outBytes + size > outputMaxBytes) {
+            outputTruncated = true;
+            return;
+        }
+        outBytes += size;
+        out.push(row);
+    };
     const fileMatches = [];
     const counts = [];
     let totalContent = 0;
@@ -888,14 +911,12 @@ export async function jsGrep(env, root, p, signal, guards, deny, denyOut) {
                     if (p.only_matching && text !== undefined) {
                         const parts = text.split("\n");
                         for (let k = 0; k < parts.length; k++) {
-                            if (out.length < collectCap)
-                                out.push(`${relOut(f)}:${s + k}:${clipLine(parts[k])}`);
+                            pushRow(`${relOut(f)}:${s + k}:${clipLine(parts[k])}`);
                         }
                         continue;
                     }
                     for (let j = Math.max(0, s - 1 - ctxB); j <= Math.min(lines.length - 1, eL - 1 + ctxA); j++) {
-                        if (out.length < collectCap)
-                            out.push(`${relOut(f)}:${j + 1}:${clipLine(lines[j])}`);
+                        pushRow(`${relOut(f)}:${j + 1}:${clipLine(lines[j])}`);
                     }
                 }
             }
@@ -924,18 +945,16 @@ export async function jsGrep(env, root, p, signal, guards, deny, denyOut) {
                 if (mode === "content") {
                     if (p.only_matching) {
                         for (const part of matchedParts(lines[i], p)) {
-                            if (out.length < collectCap)
-                                out.push(`${relOut(f)}:${i + 1}:${clipLine(part)}`);
+                            pushRow(`${relOut(f)}:${i + 1}:${clipLine(part)}`);
                         }
                     }
                     else if (ctx > 0) {
                         for (let j = Math.max(0, i - ctxB); j <= Math.min(lines.length - 1, i + ctxA); j++) {
-                            if (out.length < collectCap)
-                                out.push(`${relOut(f)}:${j + 1}:${clipLine(lines[j])}`);
+                            pushRow(`${relOut(f)}:${j + 1}:${clipLine(lines[j])}`);
                         }
                     }
-                    else if (out.length < collectCap) {
-                        out.push(`${relOut(f)}:${i + 1}:${clipLine(lines[i])}`);
+                    else {
+                        pushRow(`${relOut(f)}:${i + 1}:${clipLine(lines[i])}`);
                     }
                 }
             }
@@ -990,15 +1009,16 @@ export async function jsGrep(env, root, p, signal, guards, deny, denyOut) {
             ? NO_MATCHES + offNote + typeNote + caveat
             : body.join("\n") + (totalFiles > off + cap ? `\n…[capped at ${cap} of ${totalFiles}]` : "") + offNote + typeNote + caveat;
     }
-    const body = paged(out);
+    const body = out;
+    const byteNote = outputTruncated ? `\n…[output truncated at ${outputMaxBytes} bytes — narrow the pattern or set a smaller head_limit]` : "";
     if (body.length === 0)
-        return NO_MATCHES + offNote + typeNote + caveat;
-    if (out.length < collectCap)
-        return body.join("\n") + offNote + typeNote + caveat;
+        return NO_MATCHES + byteNote + offNote + typeNote + caveat;
+    if (out.length < cap)
+        return body.join("\n") + byteNote + offNote + typeNote + caveat;
     const marker = ctx > 0 || p.multiline || p.only_matching
         ? `\n…[capped at ${cap}; ${totalContent}+ match(es) found, more output omitted]`
         : `\n…[capped at ${cap} of ${totalContent}]`;
-    return body.join("\n") + marker + offNote + typeNote + caveat;
+    return body.join("\n") + marker + byteNote + offNote + typeNote + caveat;
 }
 const rgCache = new WeakMap();
 export function detectRipgrep(env) {
@@ -1012,79 +1032,89 @@ export function detectRipgrep(env) {
     }
     return cached;
 }
-async function sortRgFilesByMtime(env, root, stdout, signal) {
-    const paths = stdout.split("\n").filter((l) => l.length > 0);
-    if (paths.length === 0)
-        return stdout;
+async function sortRgFilesByMtime(env, root, records, signal) {
+    if (records.length === 0)
+        return [...records];
     const rootPrefix = root.replace(/[\\/]+$/, "") + (root.includes("\\") ? "\\" : "/");
     const toAbs = (p) => {
         const stripped = p.startsWith("./") ? p.slice(2) : p;
         return isAbsolutePathForm(stripped) ? stripped : `${rootPrefix}${stripped}`;
     };
+    const paths = records.map((r) => r.path ?? r.text);
     const infos = await Promise.all(paths.map((p) => env.fileInfo(toAbs(p), signal)));
-    const withMtime = paths.map((p, i) => {
-        const r = infos[i];
-        return { p, m: r.ok && typeof r.value.mtimeMs === "number" ? r.value.mtimeMs : undefined };
+    const withMtime = records.map((r, i) => {
+        const info = infos[i];
+        return { r, p: paths[i], m: info.ok && typeof info.value.mtimeMs === "number" ? info.value.mtimeMs : undefined };
     });
     const sorted = withMtime.every((x) => x.m !== undefined)
         ? [...withMtime].sort((a, b) => b.m - a.m || (a.p < b.p ? -1 : a.p > b.p ? 1 : 0))
         : [...withMtime].sort((a, b) => (a.p < b.p ? -1 : a.p > b.p ? 1 : 0));
-    return sorted.map((x) => x.p).join("\n");
+    return sorted.map((x) => x.r);
+}
+function rgRecordSeparator(mode, rest) {
+    if (mode === "count")
+        return /^\d+$/.test(rest) ? ":" : null;
+    const m = /^\d+([:\-])/.exec(rest);
+    return m === null ? null : m[1];
+}
+function parseRgRecords(stdout, mode, dropIncompleteTail = false) {
+    if (stdout.length === 0)
+        return [];
+    if (mode === "files_with_matches") {
+        if (!stdout.includes("\0")) {
+            if (dropIncompleteTail)
+                return [];
+            return stdout.split("\n").filter((l) => l.length > 0).map((l) => ({ path: l, text: l }));
+        }
+        const parts = stdout.split("\0");
+        const tail = parts.pop() ?? "";
+        const records = parts.filter((p) => p.length > 0).map((p) => ({ path: p, text: p }));
+        if (tail.length > 0 && !dropIncompleteTail)
+            records.push({ text: tail });
+        return records;
+    }
+    const lines = stdout.split("\n");
+    if (!stdout.endsWith("\n") && dropIncompleteTail)
+        lines.pop();
+    const records = [];
+    for (const line of lines) {
+        if (line.length === 0)
+            continue;
+        const nul = line.indexOf("\0");
+        const rest = nul < 0 ? "" : line.slice(nul + 1);
+        const sep = nul < 0 ? null : rgRecordSeparator(mode, rest);
+        if (nul < 0 || sep === null) {
+            records.push({ text: line });
+            continue;
+        }
+        const path = line.slice(0, nul);
+        records.push({ path, text: `${path}${sep}${rest}` });
+    }
+    return records;
 }
-function formatRgStdout(stdout, p, caveat = "") {
+function formatRgRecords(records, p, caveat = "") {
     const cap = p.head_limit === 0 ? Infinity : Math.max(1, Math.floor(p.head_limit ?? GREP_DEFAULT_CAP));
     const off = Math.max(0, Math.floor(p.offset ?? 0));
-    const lines = stdout.split("\n").filter((l) => l.length > 0);
-    const capped = lines.slice(off, off + cap);
+    const capped = records.slice(off, off + cap);
     if (capped.length === 0)
         return NO_MATCHES + (off > 0 ? `\n[offset ${off}]` : "") + caveat;
-    return (capped.join("\n") +
-        (lines.length > off + cap ? `\n…[capped at ${cap} of ${lines.length}]` : "") +
+    return (capped.map((r) => r.text).join("\n") +
+        (records.length > off + cap ? `\n…[capped at ${cap} of ${records.length}]` : "") +
         (off > 0 ? `\n[offset ${off}]` : "") +
         caveat);
 }
-export function rgOutputDenyTripwire(stdout, mode, judge) {
+export function rgOutputDenyTripwire(stdout, mode, judge, opts = {}) {
     if (stdout.length === 0)
         return { trip: false };
-    const judgeBoundaries = (line, boundary) => {
-        let sawBoundary = false;
-        for (let m = boundary.exec(line); m !== null; m = boundary.exec(line)) {
-            sawBoundary = true;
-            const prefix = line.slice(0, m.index);
-            if (prefix.length === 0)
+    for (const record of parseRgRecords(stdout, mode, opts.dropIncompleteTail === true)) {
+        if (record.path === undefined) {
+            if (record.text === "--")
                 continue;
-            const hit = judge.matchPath(prefix);
-            if (hit !== null)
-                return hit.pattern;
-        }
-        return sawBoundary ? null : "";
-    };
-    for (const line of stdout.split("\n")) {
-        if (line.length === 0 || line === "--")
-            continue;
-        if (mode === "files_with_matches") {
-            const h = judge.matchPath(line);
-            if (h !== null)
-                return { trip: true, reason: "deny-hit", pattern: h.pattern };
-        }
-        else if (mode === "count") {
-            const m = /^(.*):\d+$/.exec(line);
-            if (m === null)
-                return { trip: true, reason: "ambiguous-record" };
-            const h = judge.matchPath(m[1]);
-            if (h !== null)
-                return { trip: true, reason: "deny-hit", pattern: h.pattern };
-        }
-        else {
-            const colon = judgeBoundaries(line, /:\d+:/g);
-            if (colon !== null && colon !== "")
-                return { trip: true, reason: "deny-hit", pattern: colon };
-            const dash = judgeBoundaries(line, /-\d+-/g);
-            if (dash !== null && dash !== "")
-                return { trip: true, reason: "deny-hit", pattern: dash };
-            if (colon === "" && dash === "")
-                return { trip: true, reason: "ambiguous-record" };
+            return { trip: true, reason: "ambiguous-record" };
         }
+        const hit = judge.matchPath(record.path);
+        if (hit !== null)
+            return { trip: true, reason: "deny-hit", pattern: hit.pattern };
     }
     return { trip: false };
 }
@@ -1113,7 +1143,7 @@ async function jsGrepFallback(env, root, p, signal, reason, deny) {
 }
 export async function rgGrepDetailed(env, root, p, signal, deny) {
     const mode = p.output_mode ?? "files_with_matches";
-    const flags = ["--no-messages", "--no-require-git", "--hidden"];
+    const flags = ["--null", "--no-messages", "--no-require-git", "--hidden"];
     for (const d of VCS_DIRS)
         flags.push("--glob", `!${d}`);
     flags.push("--max-columns", "500", "--max-columns-preview");
@@ -1155,12 +1185,15 @@ export async function rgGrepDetailed(env, root, p, signal, deny) {
             const partial = err.partialStdout ?? "";
             if (partial.trim().length > 0) {
                 if (deny !== undefined) {
-                    const trip = rgOutputDenyTripwire(partial, mode, deny);
+                    const trip = rgOutputDenyTripwire(partial, mode, deny, { dropIncompleteTail: true });
                     if (trip.trip)
                         return jsGrepDenyTripFallback(env, root, p, signal, trip, deny);
                 }
+                const records = parseRgRecords(partial, mode, true);
+                if (records.length === 0)
+                    return jsGrepFallback(env, root, p, signal, "timed out before completing a result", deny);
                 return {
-                    text: `${delimitUntrusted("ripgrep partial output", formatRgStdout(partial, p))}\n…[ripgrep timed out after producing partial output — results may be incomplete]`,
+                    text: `${delimitUntrusted("ripgrep partial output", formatRgRecords(records, p))}\n…[ripgrep timed out after producing partial output — results may be incomplete]`,
                     degraded: { partial: true, reason: "ripgrep timed out" },
                 };
             }
@@ -1187,12 +1220,15 @@ export async function rgGrepDetailed(env, root, p, signal, deny) {
     if (exitCode >= 2) {
         if (stdout.trim().length > 0) {
             if (deny !== undefined) {
-                const trip = rgOutputDenyTripwire(stdout, mode, deny);
+                const trip = rgOutputDenyTripwire(stdout, mode, deny, { dropIncompleteTail: true });
                 if (trip.trip)
                     return jsGrepDenyTripFallback(env, root, p, signal, trip, deny);
             }
+            const records = parseRgRecords(stdout, mode, true);
+            if (records.length === 0)
+                return jsGrepFallback(env, root, p, signal, `exited with code ${exitCode} and produced no complete result`, deny);
             return {
-                text: `${delimitUntrusted("ripgrep partial output", formatRgStdout(stdout, p))}\n…[ripgrep exited with an error after producing partial output — results may be incomplete]`,
+                text: `${delimitUntrusted("ripgrep partial output", formatRgRecords(records, p))}\n…[ripgrep exited with an error after producing partial output — results may be incomplete]`,
                 degraded: { partial: true, reason: `ripgrep exited with code ${exitCode}` },
             };
         }
@@ -1203,9 +1239,10 @@ export async function rgGrepDetailed(env, root, p, signal, deny) {
         if (trip.trip)
             return jsGrepDenyTripFallback(env, root, p, signal, trip, deny);
     }
-    const orderedStdout = mode === "files_with_matches" ? await sortRgFilesByMtime(env, root, stdout, signal) : stdout;
+    const parsed = parseRgRecords(stdout, mode);
+    const ordered = mode === "files_with_matches" ? await sortRgFilesByMtime(env, root, parsed, signal) : parsed;
     const d = await denyDisclosure();
-    return { text: formatRgStdout(orderedStdout, p) + d.note, ...(d.withheld !== undefined ? { withheld: d.withheld } : {}) };
+    return { text: formatRgRecords(ordered, p) + d.note, ...(d.withheld !== undefined ? { withheld: d.withheld } : {}) };
 }
 async function rgDenyExistenceProbe(env, root, deny, target, signal) {
     const probeFlags = ["--files", "--hidden", "--no-require-git", "--no-messages"];

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@sema-agent/core",
-  "version": "5.40.0",
+  "version": "5.42.0",
   "description": "Stateless, task-oriented AI agent core",
   "type": "module",
   "license": "BUSL-1.1",

package/test/export-surface.snapshot.json

@@ -1,6 +1,6 @@
 {
   "_comment": "design/87 L3 — frozen public export surface of src/index.ts (name -> kind). DO NOT edit by hand to silence a red test. A removed/changed entry = a SemVer-BREAKING change; bump MAJOR and update this fixture in the SAME commit (design/87 §4.2 / §5.2). Regenerate via REGEN in test/export-surface.test.ts.",
-  "count": 1605,
+  "count": 1606,
   "exports": {
     "A2ATaskState": "type",
     "A2ATaskStateReversal": "type",
@@ -1390,6 +1390,7 @@
     "mergeRecallHits": "function",
     "mergeWorkflowArgs": "function",
     "migrateScope": "function",
+    "mintCheckpointId": "function",
     "mintCheckpointToken": "function",
     "mintRuleTicket": "function",
     "missingRestoreSurface": "function",