fix(apiserver): job_status notnull filter matches every job (bug-1504)

`xcecli jobs list -a` returned "no job found" even when jobs existed.
FilterRows routes every condition on a bitmask field (job_status /
host_status) to MatchesBitmaskCondition, which only implemented OpEQ/OpNE
bitmask semantics — the `-a` filter compiles to `job_status notnull`
(OpNOTNULL), which fell through to `return false` and filtered out every
job. (The default list's `notin(EXIT,DONE)` is compiled to OpEQ(mask), so
it was unaffected.)

Add OpISNULL/OpNOTNULL branches to MatchesBitmaskCondition that test the
raw field value. New query_test.go: 3 regression tests (notnull matches
all, null matches none, bitmask notin path unaffected), RED->GREEN.
Verified live: `xcecli jobs list -a` now lists all jobs including DONE.

Co-Authored-By: Claude <noreply@anthropic.com>

.wolf/buglog.json

@@ -14797,5 +14797,24 @@
     ],
     "occurrences": 1,
     "last_seen": "2026-08-05"
+  },
+  {
+    "id": "bug-1504",
+    "timestamp": "2026-08-05",
+    "error_message": "xcecli jobs list -a \u8fd4\u56de \"no job found\"\uff08\u5e94\u663e\u793a\u5168\u90e8\u4f5c\u4e1a\u542b\u5df2\u5b8c\u6210\uff09\uff1b\u9ed8\u8ba4\u5217\u8868\u53ea\u663e\u793a RUN \u4f5c\u4e1a",
+    "file": "src/frontend/apiserver/store/query.go (MatchesBitmaskCondition) + query_test.go",
+    "root_cause": "FilterRows \u628a bitmask \u5b57\u6bb5 (job_status/host_status) \u7684\u6240\u6709\u6761\u4ef6\u90fd\u8def\u7531\u5230 MatchesBitmaskCondition\uff0c\u4f46\u8be5\u51fd\u6570\u53ea\u5b9e\u73b0 OpEQ/OpNE \u7684\u4f4d\u63a9\u7801\u8bed\u4e49\u3002 \u751f\u6210\u7684  (OpNOTNULL) \u843d\u5165  -> \u6240\u6709\u4f5c\u4e1a\u88ab\u8fc7\u6ee4\u3002\u9ed8\u8ba4\u5217\u8868\u7684 notin(EXIT,DONE) \u88ab\u7f16\u8bd1\u6210 OpEQ(mask) \u6240\u4ee5\u4e0d\u53d7\u5f71\u54cd\u3002",
+    "fix": "MatchesBitmaskCondition \u589e\u52a0 OpISNULL/OpNOTNULL \u5206\u652f\uff0c\u76f4\u63a5\u5728\u539f\u59cb\u5b57\u6bb5\u503c\u4e0a\u5224\u7a7a\u3002\u65b0\u589e query_test.go 3 \u4e2a\u56de\u5f52\u6d4b\u8bd5\uff08notnull \u5168\u5339\u914d / null \u7a7a\u5339\u914d / notin \u4f4d\u63a9\u7801\u8def\u5f84\u4e0d\u56de\u5f52\uff09\uff0cRED->GREEN \u9a8c\u8bc1\u3002",
+    "tags": [
+      "apiserver",
+      "xcecli",
+      "jobs-list",
+      "filter",
+      "bitmask",
+      "notnull"
+    ],
+    "related_bugs": [],
+    "occurrences": 1,
+    "last_seen": "2026-08-05"
   }
 ]

.wolf/memory.md

@@ -7370,3 +7370,4 @@
 | 05:50 | Session end: 9 writes across 5 files (frpcdbg_mon.sh, frpc_prod.toml, frpc_mon.sh, frpc_watchdog.sh, frpc-watchdog.sh) | 2 reads | ~2219 tok |
 | 09:39 | 修复 controller 反复崩溃 bug-1502: persist_job_hook_fun 把 RUSAGE_UPDATE hook 的第2参 (xce_job_rusages_t*) 当 job_t* 用, 读 offset 56 越界 56字节结构 -> ASAN heap-buffer-overflow. 加专用 case 从 jobmgr 取当前 state | src/controller/persist.c, testsuite/test_persist_hooks.c | RED(旧代码 stack-buffer-overflow)/GREEN(修复后 PASS), 全套458测试通过 | ~350 |
 | 01:08 | 验证完成: bug-1502(controller ASAN崩溃)修复生产生效(sleep 666跑满666秒, controller 0崩溃); 作业PENDING根因=配置正确但scheduler先于worker join拿不到slots(操作时序, bug-1503); 干净controller部署, 集群全链路(apiserver提交->scheduler->controller->worker)冒烟通过 | persist.c, sched.c, worker.c | 验证通过 | ~300 |
+| 01:42 | 修复 bug-1504: xcecli jobs list -a 返回空, 因 FilterRows 把 job_status notnull 路由到 MatchesBitmaskCondition(仅实现OpEQ/OpNE) -> 全过滤. 加 OpISNULL/OpNOTNULL 分支 + 3回归测试, 部署后 -a 正常显示全部作业 | src/frontend/apiserver/store/query.go + query_test.go | RED->GREEN, 生产验证 -a 列出全部 | ~250 |

src/frontend/apiserver/store/query.go

@@ -78,8 +78,18 @@ func MatchesCondition(row utils.OrderMap, cond utils.Condition) bool {
 // MatchesBitmaskCondition handles bitmask fields (job_status, host_status).
 // For bitmask EQ: (val & mask) != 0 means the status is set.
 // For bitmask NE: (val & mask) == 0 means the status is NOT set.
+// null/notnull are handled on the raw field value (bug-1504: job_status
+// notnull previously fell through to `return false`, so `jobs list -a`
+// filtered out every job).
 func MatchesBitmaskCondition(row utils.OrderMap, field string, bitmask int64, op utils.OperatorType) bool {
-	val := toInt64(row.Data[field])
+	fieldVal := row.Data[field]
+	switch op {
+	case utils.OpISNULL:
+		return fieldVal == nil
+	case utils.OpNOTNULL:
+		return fieldVal != nil
+	}
+	val := toInt64(fieldVal)
 	result := val & bitmask
 	switch op {
 	case utils.OpNE: // bitmask "eq" is stored as OpNE (val & mask != 0)

src/frontend/apiserver/store/query_test.go

@@ -0,0 +1,85 @@
+package store
+
+import (
+	"testing"
+
+	"apiserver/utils"
+)
+
+// makeJobRow builds a row in the OrderMap shape QueryData produces for a job.
+func makeJobRow(jobid int64, status int) utils.OrderMap {
+	return utils.OrderMap{
+		Fields: []string{"job_id", "job_status", "submit_user", "queue_name"},
+		Data: map[string]interface{}{
+			"job_id":      jobid,
+			"job_status":  status,
+			"submit_user": "admin",
+			"queue_name":  "normal",
+		},
+	}
+}
+
+// bug-1504: `xcecli jobs list -a` (job_status notnull) filtered out EVERY job.
+// job_status is a bitmask field, so FilterRows routed the OpNOTNULL condition
+// to MatchesBitmaskCondition, which only implements OpEQ/OpNE bitmask
+// semantics and fell through to `return false` — the "all jobs" query always
+// returned "no job found".
+func TestFilterRows_JobStatusNotNull_MatchesAllJobs(t *testing.T) {
+	rows := []utils.OrderMap{
+		makeJobRow(1, 0x02), // RUN
+		makeJobRow(2, 0x04), // EXIT
+		makeJobRow(3, 0x08), // DONE
+	}
+
+	cond := utils.Condition{
+		Column:   "job_status",
+		Operator: utils.OpNOTNULL,
+		Value:    utils.Value{Type: utils.TypeInt, Value: int64(0)},
+	}
+
+	got := FilterRows(rows, []utils.Condition{cond})
+	if len(got) != len(rows) {
+		t.Fatalf("job_status notnull: got %d rows, want %d (all jobs have a status)", len(got), len(rows))
+	}
+}
+
+func TestFilterRows_JobStatusIsNull_MatchesNothing(t *testing.T) {
+	rows := []utils.OrderMap{
+		makeJobRow(1, 0x02),
+		makeJobRow(2, 0x04),
+	}
+
+	cond := utils.Condition{
+		Column:   "job_status",
+		Operator: utils.OpISNULL,
+		Value:    utils.Value{Type: utils.TypeInt, Value: int64(0)},
+	}
+
+	got := FilterRows(rows, []utils.Condition{cond})
+	if len(got) != 0 {
+		t.Fatalf("job_status null: got %d rows, want 0", len(got))
+	}
+}
+
+// Regression guard: the default list filter (job_status notin EXIT/DONE →
+// bitmask OpEQ with mask 0x0C) must keep working after the null/notnull fix.
+func TestFilterRows_JobStatusNotinExitDone(t *testing.T) {
+	rows := []utils.OrderMap{
+		makeJobRow(1, 0x02), // RUN — should survive
+		makeJobRow(2, 0x04), // EXIT — filtered
+		makeJobRow(3, 0x0C), // EXIT|DONE — filtered
+	}
+
+	// notin (EXIT,DONE) on a bitmask field is compiled to OpEQ with mask
+	// EXIT|DONE = 0x0C: keep rows where (val & mask) == 0.
+	cond := utils.Condition{
+		Column:   "job_status",
+		Operator: utils.OpEQ,
+		Value:    utils.Value{Type: utils.TypeInt, Value: int64(0x0C)},
+	}
+
+	got := FilterRows(rows, []utils.Condition{cond})
+	if len(got) != 1 || got[0].Data["job_status"] != 0x02 {
+		t.Fatalf("job_status notin(EXIT,DONE): got %d rows, want exactly the RUN job", len(got))
+	}
+}