1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
//! `--daemon monorepo` mode: launch the daemon from a checked-out copy of
//! `econ-v1/node`. Lets platform contributors edit BOTH app code AND
//! daemon code in a single inner-loop iteration.
//!
//! The CLI verifies the path is a workspace containing the `node-server`
//! crate, generates a per-(monorepo, instance) env file under
//! `$XDG_CACHE_HOME/node-app/monorepo-<hash>/`, and spawns
//! `cargo run -p node-server -- --env <env>` as a child process.
//!
//! The hash covers both the monorepo path AND the instance name so that
//! running alice and bob simultaneously does not collide on socket or db.
//!
//! When `log_tx` is `Some` (TUI active), the daemon and UI dev server
//! stdout/stderr are captured, prefixed with `[<instance>]`, and piped
//! into the split-pane log view instead of being inherited.
use std::collections::hash_map::DefaultHasher;
use std::fs;
use std::hash::{Hash, Hasher};
use std::io::BufRead;
use std::os::unix::net::UnixStream;
#[cfg(unix)]
use std::os::unix::process::CommandExt;
use std::path::{Path, PathBuf};
use std::process::{Child, Command, Stdio};
use std::sync::Mutex;
use std::time::{Duration, Instant};
use anyhow::{anyhow, bail, Context, Result};
#[cfg(unix)]
use libc;
use serde_json;
use super::{DaemonHandle, DaemonHost, InstanceProfile};
use crate::tui::{self, LogSource, LogTx, ServiceStatus, TuiEvent};
// Bumped from 120s to 240s: cold cargo run can include a non-trivial link
// step + multi-app inotify scan, especially in multi-instance platform mode
// where alice and bob compete for the page cache.
const HEALTH_TIMEOUT: Duration = Duration::from_secs(240);
const HEALTH_POLL: Duration = Duration::from_secs(2);
/// Max time we'll wait for the daemon's HTTP server to answer /api/healthz
/// after the IPC socket comes up. Distinct from `HEALTH_TIMEOUT` because the
/// IPC socket can be live before axum binds the TCP port — a window where a
/// browser-loaded Vite proxy would fail with "connection refused".
const HTTP_READY_TIMEOUT: Duration = Duration::from_secs(60);
/// Max time we'll wait for Vite to bind its UI port and respond to a smoke
/// test through the proxy. Vite startup is typically 1–3s; 30s leaves margin
/// for cold node_modules / vite optimizer runs without hanging the TUI.
const VITE_READY_TIMEOUT: Duration = Duration::from_secs(30);
/// Return PIDs of processes listening on `port` (TCP, LISTEN state). Used to
/// reap stale Vite / daemon orphans before we try to bind. On non-unix or
/// when `lsof` is unavailable this returns an empty Vec (degrading to a
/// "best effort" cleanup — the subsequent spawn will simply fail loudly on
/// EADDRINUSE).
#[cfg(unix)]
fn pids_listening_on(port: u16) -> Vec<i32> {
let arg = format!("-iTCP:{port}");
Command::new("lsof")
.args(["-sTCP:LISTEN", "-t", "-P", "-n", &arg])
.stdin(Stdio::null())
.stdout(Stdio::piped())
.stderr(Stdio::null())
.output()
.ok()
.filter(|o| o.status.success())
.map(|o| {
String::from_utf8_lossy(&o.stdout)
.lines()
.filter_map(|l| l.trim().parse::<i32>().ok())
.collect()
})
.unwrap_or_default()
}
#[cfg(not(unix))]
fn pids_listening_on(_port: u16) -> Vec<i32> {
Vec::new()
}
/// Tee a child stdio stream three ways: persistent log file (if open), TUI
/// sink (if active), and the controlling terminal (when no TUI). Each output
/// line is prefixed with the instance name so multi-instance log files stay
/// distinguishable on shared display paths.
fn tee_daemon_stream<R: std::io::Read + Send + 'static>(
reader: R,
prefix: String,
log_file: Option<std::sync::Arc<std::sync::Mutex<std::fs::File>>>,
tui_tx: Option<LogTx>,
is_stderr: bool,
) {
use std::io::Write as _;
std::thread::spawn(move || {
for line in std::io::BufReader::new(reader).lines().map_while(Result::ok) {
let formatted = format!("{prefix}{line}");
if let Some(file) = &log_file {
if let Ok(mut f) = file.lock() {
let _ = writeln!(f, "{formatted}");
}
}
if let Some(tx) = &tui_tx {
let _ = tx.send(TuiEvent::Log(crate::tui::LogEntry {
source: LogSource::Daemon,
line: formatted,
}));
} else if is_stderr {
eprintln!("{formatted}");
} else {
println!("{formatted}");
}
}
});
}
pub struct MonorepoHost {
monorepo_path: PathBuf,
profile: InstanceProfile,
socket_override: Option<PathBuf>,
dev_dir_override: Option<PathBuf>,
log_tx: Option<LogTx>,
child: Mutex<Option<Child>>,
ui_child: Mutex<Option<Child>>,
/// Env file path saved by ensure_running() so restart() can respawn.
env_file: Mutex<Option<PathBuf>>,
}
impl MonorepoHost {
pub fn new(
monorepo_path: PathBuf,
profile: InstanceProfile,
socket_override: Option<PathBuf>,
dev_dir_override: Option<PathBuf>,
log_tx: Option<LogTx>,
) -> Self {
Self {
monorepo_path,
profile,
socket_override,
dev_dir_override,
log_tx,
child: Mutex::new(None),
ui_child: Mutex::new(None),
env_file: Mutex::new(None),
}
}
fn syslog(&self, msg: impl Into<String>) {
tui::sys_log(self.log_tx.as_ref(), msg);
}
/// Spawn daemon child with stdio wired to:
/// 1. A persistent log file at `<env_dir>/daemon.log` (always)
/// 2. The TUI sink (when active)
/// 3. stdout/stderr inheritance (when no TUI is attached)
///
/// Persistence is the important one — without it, daemon output is lost
/// the moment the TUI exits, which makes diagnosing startup failures
/// (the "daemon did not come up within Ns" error) almost impossible.
///
/// We exec the built binary at `target/debug/node-server` directly
/// instead of `cargo run -p node-server`. The earlier `cargo build`
/// step already produced that binary, and `cargo run` would otherwise
/// re-acquire the build lock — which contends with stray cargo
/// processes (rust-analyzer, leftover daemons from cancelled runs) and
/// is exactly what caused the silent "Blocking waiting for file lock
/// on build directory" hangs that manifested as health-check timeouts.
fn spawn_daemon(&self, env_file: &Path) -> Result<Child> {
let binary_path = self.monorepo_path.join("target/debug/node-server");
if !binary_path.exists() {
bail!(
"daemon binary missing at {} — the earlier `cargo build -p node-server` \
step should have produced it. Run `cargo build -p node-server \
--features agentic_payments` in {} to recover.",
binary_path.display(),
self.monorepo_path.display()
);
}
let mut cmd = Command::new(&binary_path);
cmd.args(["--env"])
.arg(env_file)
.current_dir(&self.monorepo_path)
.stdin(Stdio::null())
.stdout(Stdio::piped())
.stderr(Stdio::piped());
#[cfg(unix)]
cmd.process_group(0);
let mut child = cmd
.spawn()
.with_context(|| format!("spawn {}", binary_path.display()))?;
// Open daemon.log next to the env file. Best-effort: failure to open
// (e.g. permission denied) only loses persistence; we still tee to
// TUI/terminal. Truncate on each spawn so the file always reflects
// the *current* run rather than accumulating across restarts.
let log_file_path = env_file
.parent()
.map(|d| d.join("daemon.log"))
.unwrap_or_else(|| PathBuf::from("daemon.log"));
let log_file: Option<std::sync::Arc<std::sync::Mutex<std::fs::File>>> =
std::fs::OpenOptions::new()
.create(true)
.write(true)
.truncate(true)
.open(&log_file_path)
.ok()
.map(|f| std::sync::Arc::new(std::sync::Mutex::new(f)));
if log_file.is_some() {
self.syslog(format!("→ daemon log: {}", log_file_path.display()));
} else {
self.syslog(format!(
"⚠could not open daemon log file at {} — output will only show in TUI/terminal",
log_file_path.display()
));
}
let prefix = format!("[{}] ", self.profile.name);
if let Some(stdout) = child.stdout.take() {
tee_daemon_stream(
stdout,
prefix.clone(),
log_file.clone(),
self.log_tx.clone(),
false,
);
}
if let Some(stderr) = child.stderr.take() {
tee_daemon_stream(stderr, prefix, log_file, self.log_tx.clone(), true);
}
Ok(child)
}
/// Spawn UI dev server with stdio wired to TUI or terminal.
fn spawn_ui(&self) -> Option<Child> {
let ui_dir = self.monorepo_path.join("system/ui");
if !ui_dir.join("package.json").exists() {
return None;
}
let yarn = if which_bin("yarn") { "yarn" } else { "npm" };
let ui_port = self.profile.ui_port;
let api_port = self.profile.http_port;
self.syslog(format!(
"→ starting UI dev server ({yarn} dev --port {ui_port}) in {}",
ui_dir.display()
));
let mut cmd = Command::new(yarn);
cmd.args(["dev", "--port", &ui_port.to_string()])
.env("VITE_BACKEND_PORT", api_port.to_string())
.current_dir(&ui_dir)
.stdin(Stdio::null());
if self.log_tx.is_some() {
cmd.stdout(Stdio::piped()).stderr(Stdio::piped());
} else {
cmd.stdout(Stdio::inherit()).stderr(Stdio::inherit());
}
#[cfg(unix)]
cmd.process_group(0);
match cmd.spawn() {
Ok(mut child) => {
if let Some(tx) = &self.log_tx {
let prefix = format!("[{}] ", self.profile.name);
if let Some(stdout) = child.stdout.take() {
let tx = tx.clone();
let p = prefix.clone();
std::thread::spawn(move || {
for line in std::io::BufReader::new(stdout).lines().map_while(Result::ok) {
let _ = tx.send(TuiEvent::Log(crate::tui::LogEntry {
source: LogSource::UiServer,
line: format!("{p}{line}"),
}));
}
});
}
if let Some(stderr) = child.stderr.take() {
let tx = tx.clone();
let p = prefix.clone();
std::thread::spawn(move || {
for line in std::io::BufReader::new(stderr).lines().map_while(Result::ok) {
let _ = tx.send(TuiEvent::Log(crate::tui::LogEntry {
source: LogSource::UiServer,
line: format!("{p}{line}"),
}));
}
});
}
}
tui::update_status(
self.log_tx.as_ref(),
LogSource::UiServer,
ServiceStatus::Ready,
Some(format!("[{}] http://localhost:{ui_port}", self.profile.name)),
);
self.syslog(format!(
"✓ UI available at http://localhost:{ui_port} (Vite, hot reload) \
or http://localhost:{api_port} (backend-served, requires `yarn build` in system/ui)"
));
Some(child)
}
Err(e) => {
self.syslog(format!(
"⚠could not start UI dev server: {e} — run manually: \
cd system/ui && VITE_BACKEND_PORT={api_port} {yarn} dev --port {ui_port}"
));
tui::update_status(
self.log_tx.as_ref(),
LogSource::UiServer,
ServiceStatus::Disabled,
Some(format!("[{}]", self.profile.name)),
);
None
}
}
}
fn shutdown_daemon_only(&self) {
if let Some(mut child) = self.child.lock().unwrap().take() {
let pid = child.id() as i32;
#[cfg(unix)]
unsafe {
libc::kill(-pid, libc::SIGTERM);
}
let t = Instant::now();
loop {
match child.try_wait() {
Ok(Some(_)) => break,
Ok(None) if t.elapsed() > Duration::from_secs(10) => {
#[cfg(unix)]
unsafe {
libc::kill(-pid, libc::SIGKILL);
}
let _ = child.wait();
break;
}
Ok(None) => std::thread::sleep(Duration::from_millis(200)),
Err(_) => break,
}
}
}
}
}
impl MonorepoHost {
/// Reap any process listening on `port` (best-effort SIGTERM, escalating
/// to SIGKILL after `grace`). Logged as "freeing <label>" so the TUI
/// makes it obvious *which* port was contested.
///
/// This is the antidote to the most common "node-app dev doesn't work"
/// failure mode: a stale Vite or orphaned `node-server` from a previous
/// run is still bound to the port the new launch wants, and the new
/// process silently picks a different port (Vite) or fails to bind
/// (axum) — either way the UI ends up calling a daemon that isn't there.
#[cfg(unix)]
fn free_port(&self, port: u16, label: &str) {
let pids = pids_listening_on(port);
if pids.is_empty() {
return;
}
self.syslog(format!(
"→ freeing {label} port {port} from stale pid(s) {}…",
pids.iter()
.map(i32::to_string)
.collect::<Vec<_>>()
.join(",")
));
for pid in &pids {
unsafe {
libc::kill(*pid, libc::SIGTERM);
}
}
let deadline = Instant::now() + Duration::from_secs(5);
while Instant::now() < deadline {
if pids_listening_on(port).is_empty() {
self.syslog(format!("✓ {label} port {port} freed"));
return;
}
std::thread::sleep(Duration::from_millis(200));
}
let remaining = pids_listening_on(port);
if !remaining.is_empty() {
self.syslog(format!(
"⚠{label} port {port} still held after SIGTERM — escalating to SIGKILL"
));
for pid in &remaining {
unsafe {
libc::kill(*pid, libc::SIGKILL);
}
}
}
}
#[cfg(not(unix))]
fn free_port(&self, _port: u16, _label: &str) {}
/// Block until `http://127.0.0.1:<port>/api/healthz` answers (any status
/// below 500 counts — we only care that the listener is up and routing).
/// Returns Err on timeout so the caller can refuse to spawn Vite and
/// surface a useful failure rather than letting the browser hit a dead
/// proxy target.
fn wait_http_ready(&self, port: u16) -> Result<()> {
let url = format!("http://127.0.0.1:{port}/api/healthz");
let started = Instant::now();
self.syslog(format!("→ waiting for daemon HTTP on {url}…"));
loop {
// ureq treats any HTTP status >= 400 as an Err — but for a
// readiness check we want to accept those too (the listener is
// up, that's what matters). Match both Ok and Err(Status).
let ready = match ureq::get(&url)
.timeout(Duration::from_secs(2))
.call()
{
Ok(_) => true,
Err(ureq::Error::Status(code, _)) => code < 500,
Err(_) => false,
};
if ready {
self.syslog(format!(
"✓ daemon HTTP ready ({}ms)",
started.elapsed().as_millis()
));
return Ok(());
}
if let Some(child) = self.child.lock().unwrap().as_mut() {
if let Ok(Some(status)) = child.try_wait() {
bail!(
"daemon exited with {status} while waiting for HTTP readiness on {url}"
);
}
}
if started.elapsed() >= HTTP_READY_TIMEOUT {
bail!(
"daemon HTTP server not responding on {} after {}s",
url,
HTTP_READY_TIMEOUT.as_secs()
);
}
std::thread::sleep(Duration::from_millis(300));
}
}
/// Verify the Vite WebSocket proxy by speaking raw HTTP upgrade and
/// checking for `HTTP/1.1 101`. This catches path-mismatch bugs
/// (frontend `/ws` vs backend `/api/ws`) that otherwise only surface as
/// noisy ECONNRESET loops in the Vite log.
fn verify_ws_proxy(&self, ui_port: u16) -> Result<()> {
use std::io::{Read, Write};
use std::net::TcpStream;
// Vite usually binds IPv6 on macOS; resolve via the OS so we hit
// whichever family it picked.
let addr_candidates: Vec<std::net::SocketAddr> = std::net::ToSocketAddrs::to_socket_addrs(
&format!("localhost:{ui_port}"),
)
.map(|iter| iter.collect())
.unwrap_or_default();
if addr_candidates.is_empty() {
bail!("could not resolve localhost:{ui_port} for WS probe");
}
let req = format!(
"GET /api/ws HTTP/1.1\r\n\
Host: localhost:{ui_port}\r\n\
Upgrade: websocket\r\n\
Connection: Upgrade\r\n\
Sec-WebSocket-Key: dGhlIHNhbXBsZSBub25jZQ==\r\n\
Sec-WebSocket-Version: 13\r\n\
\r\n"
);
let mut last_err = String::from("no candidate addresses tried");
for addr in &addr_candidates {
match TcpStream::connect_timeout(addr, Duration::from_secs(2)) {
Ok(mut sock) => {
let _ = sock.set_read_timeout(Some(Duration::from_secs(2)));
let _ = sock.write_all(req.as_bytes());
let mut buf = [0u8; 256];
match sock.read(&mut buf) {
Ok(n) if n > 0 => {
let response = String::from_utf8_lossy(&buf[..n]);
let first_line =
response.lines().next().unwrap_or("").to_string();
if first_line.contains("101") {
self.syslog(format!(
"✓ ws proxy verified: localhost:{ui_port}/api/ws → backend ({first_line})"
));
return Ok(());
}
last_err = format!(
"{addr} replied {first_line} (expected 101 Switching Protocols)"
);
}
Ok(_) => last_err = format!("{addr} closed without responding"),
Err(e) => last_err = format!("{addr} read: {e}"),
}
}
Err(e) => last_err = format!("{addr} connect: {e}"),
}
}
bail!("ws proxy probe failed: {last_err}");
}
/// Smoke-test the Vite proxy by issuing `/api/healthz` against the UI
/// port and verifying we get a backend status (anything < 500 — the
/// proxy forwarded successfully). Emits a clear ✓ or ⚠line so the
/// developer can tell at a glance whether the `/api` lane works.
///
/// We try `localhost` first (which resolves to whichever family Vite
/// bound — Vite 5+ defaults to IPv6 `::1` on macOS) and fall back to
/// `127.0.0.1` then `[::1]` so a misconfigured /etc/hosts doesn't make
/// us report a false negative.
fn verify_proxy(&self, ui_port: u16, http_port: u16) -> Result<()> {
let urls = [
format!("http://localhost:{ui_port}/api/healthz"),
format!("http://127.0.0.1:{ui_port}/api/healthz"),
format!("http://[::1]:{ui_port}/api/healthz"),
];
let started = Instant::now();
let mut last_err = String::new();
loop {
for url in &urls {
match ureq::get(url).timeout(Duration::from_secs(2)).call() {
Ok(resp) => {
self.syslog(format!(
"✓ proxy verified: {url} → 127.0.0.1:{http_port} (status {})",
resp.status()
));
return Ok(());
}
Err(ureq::Error::Status(code, _)) if code < 500 => {
self.syslog(format!(
"✓ proxy verified: {url} → 127.0.0.1:{http_port} (status {code})"
));
return Ok(());
}
Err(e) => last_err = format!("{url}: {e}"),
}
}
if started.elapsed() >= VITE_READY_TIMEOUT {
bail!(
"vite proxy unreachable on port {ui_port} after {}s (last: {last_err}) — \
check that VITE_BACKEND_PORT was set when vite started, \
and that the daemon is listening on 127.0.0.1:{http_port}",
VITE_READY_TIMEOUT.as_secs()
);
}
std::thread::sleep(Duration::from_millis(400));
}
}
/// Run a build command, streaming stderr to the TUI log (or to the
/// terminal when no TUI is active). Returns the exit status and all
/// stderr lines collected — callers include them in error messages.
fn run_build_cmd(
&self,
cmd: &mut Command,
log_source: LogSource,
) -> Result<(std::process::ExitStatus, Vec<String>)> {
use std::sync::mpsc;
cmd.stdout(Stdio::null()).stderr(Stdio::piped());
let mut child = cmd.spawn()?;
let stderr = child.stderr.take().expect("piped");
let log_tx = self.log_tx.clone();
let (tx, rx) = mpsc::channel::<Vec<String>>();
std::thread::spawn(move || {
let mut lines = Vec::new();
for line in std::io::BufReader::new(stderr).lines().map_while(Result::ok) {
if let Some(ref ltx) = log_tx {
let _ = ltx.send(TuiEvent::Log(crate::tui::LogEntry {
source: log_source,
line: line.clone(),
}));
} else {
eprintln!("{line}");
}
lines.push(line);
}
let _ = tx.send(lines);
});
let status = loop {
match child.try_wait()? {
Some(s) => break s,
None => {
if crate::commands::dev::is_cancelled() {
let _ = child.kill();
let _ = child.wait();
let _lines = rx.recv().unwrap_or_default();
bail!("build cancelled");
}
std::thread::sleep(Duration::from_millis(50));
}
}
};
let lines = rx.recv().unwrap_or_default();
Ok((status, lines))
}
}
impl DaemonHost for MonorepoHost {
fn ensure_running(&self) -> Result<DaemonHandle> {
// 1. Validate the path looks like the node monorepo.
let cargo_toml = self.monorepo_path.join("Cargo.toml");
if !cargo_toml.exists() {
bail!(
"{} doesn't look like a monorepo root (no Cargo.toml found)",
self.monorepo_path.display()
);
}
let manifest = fs::read_to_string(&cargo_toml)
.with_context(|| format!("read {}", cargo_toml.display()))?;
let has_server = manifest.contains("\"system/server\"")
|| manifest.contains("system/server")
|| manifest.contains("\"apps/server\"")
|| manifest.contains("apps/server");
if !has_server {
bail!(
"{}/Cargo.toml does not include 'system/server' — is this the right path?",
self.monorepo_path.display()
);
}
// 2. Resolve cache dir (namespaced by monorepo path + instance name).
let env_dir = monorepo_env_dir(&self.monorepo_path, &self.profile.name)?;
fs::create_dir_all(&env_dir).ok();
let pid_file = env_dir.join("daemon.pid");
// 2a. Kill any previous instance.
if let Ok(contents) = fs::read_to_string(&pid_file) {
if let Ok(old_pid) = contents.trim().parse::<i32>() {
let alive = unsafe { libc::kill(old_pid, 0) == 0 };
if alive {
self.syslog(format!(
"→ previous instance found (PID {old_pid}), sending SIGTERM…"
));
unsafe {
libc::kill(-old_pid, libc::SIGTERM);
}
let deadline = Instant::now();
loop {
std::thread::sleep(Duration::from_millis(200));
if unsafe { libc::kill(old_pid, 0) } != 0 {
self.syslog(format!(
"✓ previous instance exited ({}ms)",
deadline.elapsed().as_millis()
));
break;
}
if deadline.elapsed() > Duration::from_secs(15) {
self.syslog("âš previous instance did not exit after 15s, SIGKILL");
unsafe {
libc::kill(-old_pid, libc::SIGKILL);
}
break;
}
}
}
let _ = fs::remove_file(&pid_file);
}
}
// 2b. Belt-and-suspenders: reap any stray process still holding our
// ports. Covers daemons started without writing a pid file (e.g.
// `cargo run` from a separate shell), Vite servers left behind by a
// killed `node-app dev`, and old `make dev-*` runs.
self.free_port(self.profile.http_port, "backend");
self.free_port(self.profile.ui_port, "ui");
// 2c. Stale Vite pid file (mirrors the daemon.pid path above but
// tracks the UI dev server, which we spawn in step 6).
let vite_pid_file = env_dir.join("vite.pid");
if let Ok(contents) = fs::read_to_string(&vite_pid_file) {
if let Ok(old_pid) = contents.trim().parse::<i32>() {
let alive = unsafe { libc::kill(old_pid, 0) == 0 };
if alive {
self.syslog(format!(
"→ previous vite found (PID {old_pid}), sending SIGTERM…"
));
unsafe {
libc::kill(-old_pid, libc::SIGTERM);
}
}
}
let _ = fs::remove_file(&vite_pid_file);
}
let socket_path = self
.socket_override
.clone()
.unwrap_or_else(|| env_dir.join("control.sock"));
let dev_dir = self
.dev_dir_override
.clone()
.unwrap_or_else(|| env_dir.join("dev-apps"));
let db_path = env_dir.join("dev.db");
fs::create_dir_all(&dev_dir).ok();
// 3. Generate env file.
// The instance-specific base env (alice.env / bob.env) is the
// source of truth for everything except the per-instance isolation
// paths and ports — bail clearly if the file is missing rather than
// silently fall back to a near-empty env that leads to confusing
// downstream failures.
let env_file = env_dir.join("daemon.env");
let base_env_path = self
.monorepo_path
.join("system/server")
.join(&self.profile.env_file_name);
if !base_env_path.exists() {
bail!(
"missing base env file for instance '{}' — expected {}.\n\
Create it (typically a copy of an example env), or pick a different \
instance via --instances.",
self.profile.name,
base_env_path.display()
);
}
let base_env = fs::read_to_string(&base_env_path)
.with_context(|| format!("read base env {}", base_env_path.display()))?;
let log_dir = env_dir.join("logs");
fs::create_dir_all(&log_dir).ok();
let ldk_dir = env_dir.join("ldk_data");
fs::create_dir_all(&ldk_dir).ok();
let api_port = self.profile.http_port;
// Point APT_APPS_DIR at the monorepo's `modules/` so the daemon loads
// the built-in apps (ldk-node, core-storage, …) whose .dylib/.so was
// staged there by `make builtin-apps`. Without this, the daemon falls
// back to the production `/usr/lib/node/apps` path which is absent in
// dev (especially on macOS), so no builtin capabilities ever register
// and dep apps like `discovery` hit `capability.not_found` on every
// `core.lightning.*` call.
// Overrides we *must* inject because they depend on runtime values
// the base env file can't know (cache-dir hash, monorepo path,
// per-profile ports). Anything else — TLS settings, log verbosity,
// bitcoin RPC, JWT secrets — comes from the base env file as
// written. dotenvy::from_path (the server's env loader at
// core/foundation/src/config.rs:170) is first-wins and never
// overrides an already-set key, so writing overrides BEFORE the
// base content is the only way they take effect.
let apt_apps_dir = self.monorepo_path.join("modules");
let static_dir = self.monorepo_path.join("system/ui/dist");
let pairs: [(&str, String, &str); 9] = [
("APT_APPS_DIR", apt_apps_dir.display().to_string(),
"load builtin apps from the monorepo's modules/ instead of /usr/lib/node/apps"),
("NODE_DEV_APPS_DIR", dev_dir.display().to_string(),
"stage dep apps into a per-instance cache dir so alice/bob don't collide"),
("DATABASE_URL", format!("sqlite://{}", db_path.display()),
"per-instance dev DB under cache dir"),
("NODE_IPC_SOCKET", socket_path.display().to_string(),
"per-instance IPC socket the orchestrator polls for readiness"),
("SERVER_ADDRESS", format!("127.0.0.1:{api_port}"),
"profile-specific HTTP port (alice=3001, bob=3002, …)"),
("APPLICATION_LOG_DIR_PATH", log_dir.display().to_string(),
"per-instance application logs under cache dir"),
("DATA_DIR_PATH", ldk_dir.display().to_string(),
"per-instance LDK data dir under cache dir"),
("SIGNER_SEED_PATH", format!("{}/signer_seed.hex", ldk_dir.display()),
"per-instance LDK signer seed under cache dir"),
("STATIC_DIR_PATH", static_dir.display().to_string(),
"absolute path to monorepo/system/ui/dist so backend serves UI on its port"),
];
let mut overrides = String::from(
"# ============================================================\n\
# node-app dev overrides (auto-generated — do not edit)\n\
#\n\
# These keys are computed at orchestrator startup and MUST win\n\
# over the base env file below for per-instance isolation. The\n\
# base env file (alice.env / bob.env) is the source of truth\n\
# for everything else — edit it directly if you need to change\n\
# other settings.\n\
#\n\
# dotenvy is first-wins, so this block has to come first.\n\
# ============================================================\n",
);
for (key, _, reason) in &pairs {
overrides.push_str(&format!("# {key}: {reason}\n"));
}
overrides.push('\n');
for (key, value, _) in &pairs {
overrides.push_str(&format!("{key}={value}\n"));
}
overrides.push_str(&format!(
"\n# ============================================================\n\
# Base config (system/server/{env_file_name})\n\
# ============================================================\n",
env_file_name = self.profile.env_file_name,
));
fs::write(&env_file, overrides + &base_env)
.with_context(|| format!("write {}", env_file.display()))?;
*self.env_file.lock().unwrap() = Some(env_file.clone());
// Publish the env file path to the TUI so developers can find the
// generated config the running daemon is actually using.
tui::update_daemon_env_file(
self.log_tx.as_ref(),
&self.profile.name,
env_file.clone(),
);
// 3b. Remove stale socket.
let _ = fs::remove_file(&socket_path);
// 4a. Build mandatory builtin apps.
let has_make = Command::new("make")
.arg("--version")
.stdout(Stdio::null())
.stderr(Stdio::null())
.status()
.map(|s| s.success())
.unwrap_or(false);
if has_make {
self.syslog("→ building mandatory builtin apps: make builtin-apps CARGO_PROFILE=debug");
let t0 = Instant::now();
let (status, _) = self.run_build_cmd(
Command::new("make")
.args(["builtin-apps", "CARGO_PROFILE=debug"])
.current_dir(&self.monorepo_path)
.stdin(Stdio::null()),
LogSource::System,
)
.with_context(|| "spawn `make builtin-apps`")?;
let builtins_ms = t0.elapsed().as_millis() as u64;
if !status.success() {
self.syslog(format!(
"⚠`make builtin-apps` failed (exit {}) — some platform apps may not load",
status
));
}
// Timing-only status update — no log lines go to the build tab for builtin-apps.
tui::update_status(
self.log_tx.as_ref(),
LogSource::Build,
ServiceStatus::Ready,
Some(format!("builtins:{}ms", builtins_ms)),
);
} else {
self.syslog("⚠`make` not found — skipping builtin-apps build");
}
// 4b. Delete old binary to avoid macOS linker-replace issue.
let binary_path = self.monorepo_path.join("target/debug/node-server");
if binary_path.exists() {
let _ = fs::remove_file(&binary_path);
}
self.syslog(format!(
"→ building daemon: cargo build -p node-server (cwd={})",
self.monorepo_path.display()
));
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Building,
Some(format!("cargo build (cwd={})", self.monorepo_path.display())),
);
let (build_status, build_errors) = self
.run_build_cmd(
Command::new("cargo")
.args(["build", "-p", "node-server", "--features", "agentic_payments"])
.current_dir(&self.monorepo_path)
.stdin(Stdio::null()),
LogSource::Daemon,
)
.with_context(|| "spawn `cargo build -p node-server`")?;
if !build_status.success() {
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Failed("cargo build failed".into()),
None,
);
let detail = if build_errors.is_empty() {
String::new()
} else {
format!("\n\n{}", build_errors.join("\n"))
};
bail!(
"cargo build -p node-server failed (exit {}){detail}",
build_status
);
}
// 4c. Launch daemon.
self.syslog(format!(
"→ spawning daemon (env={})",
env_file.display()
));
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Starting,
Some(format!("[{}] api=:{} ipc={}", self.profile.name, self.profile.http_port, socket_path.display())),
);
let child = self.spawn_daemon(&env_file)?;
let child_pid = child.id();
if let Err(e) = fs::write(&pid_file, child_pid.to_string()) {
self.syslog(format!("âš could not write PID file: {e}"));
}
*self.child.lock().unwrap() = Some(child);
// 5. Wait for IPC socket.
let started = Instant::now();
loop {
if UnixStream::connect(&socket_path).is_ok() {
self.syslog(format!(
"✓ IPC socket up at {} ({}s) — waiting for apps…",
socket_path.display(),
started.elapsed().as_secs()
));
break;
}
if started.elapsed() >= HEALTH_TIMEOUT {
self.shutdown();
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Failed(format!(
"socket not connectable after {}s",
HEALTH_TIMEOUT.as_secs()
)),
None,
);
return Err(anyhow!(
"daemon did not come up within {}s (socket at {} not connectable).",
HEALTH_TIMEOUT.as_secs(),
socket_path.display()
));
}
if let Some(child) = self.child.lock().unwrap().as_mut() {
if let Ok(Some(status)) = child.try_wait() {
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Failed(format!("exited {}", status)),
None,
);
return Err(anyhow!("daemon exited with {} before becoming ready", status));
}
}
std::thread::sleep(HEALTH_POLL);
}
// 5b. Wait for the HTTP listener to actually answer. The IPC socket
// can come up before axum binds its TCP port; spawning Vite before
// that point produces "ECONNREFUSED" through the proxy and looks
// exactly like a broken proxy config.
self.wait_http_ready(self.profile.http_port)?;
// 6. Start UI dev server.
let ui_child = self.spawn_ui();
if let Some(ref c) = ui_child {
// Persist the Vite PID so the next `node-app dev` run can reap
// this process even if we get killed without running our
// shutdown handler (kill -9, panic, OOM).
let _ = fs::write(&vite_pid_file, c.id().to_string());
// Smoke-test the proxy: hit /api/healthz through Vite and verify
// we got a response from the backend. A failure here is far more
// useful than letting the user discover the broken proxy at
// first page load.
if let Err(e) = self.verify_proxy(self.profile.ui_port, self.profile.http_port) {
self.syslog(format!("âš proxy verification failed: {e:#}"));
tui::update_status(
self.log_tx.as_ref(),
LogSource::UiServer,
ServiceStatus::Failed("proxy not forwarding /api".into()),
None,
);
} else if let Err(e) = self.verify_ws_proxy(self.profile.ui_port) {
// /api proxy works but WS upgrade doesn't — most likely a
// path mismatch between vite.config.ts and the backend
// route. Don't fail the whole launch (the UI will still
// mostly work without realtime), but flag it loudly.
self.syslog(format!("âš ws proxy verification failed: {e:#}"));
tui::update_status(
self.log_tx.as_ref(),
LogSource::UiServer,
ServiceStatus::Failed("ws proxy /api/ws not upgrading".into()),
None,
);
}
}
*self.ui_child.lock().unwrap() = ui_child;
// 7. Wait for critical apps.
let apps_started = Instant::now();
self.syslog("→ waiting for critical apps (device-registry, core-storage)…");
loop {
let all_ready = ipc_list_app_statuses(&socket_path)
.map(|statuses| {
["device-registry", "core-storage"].iter().all(|name| {
statuses
.get(*name)
.map(|s| s == "active" || s == "running" || s == "lazy")
.unwrap_or(false)
})
})
.unwrap_or(false);
if all_ready {
self.syslog(format!(
"✓ critical apps ready ({}s)",
apps_started.elapsed().as_secs()
));
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Ready,
Some(format!(
"[{}] api=http://localhost:{} ui=http://localhost:{}",
self.profile.name, self.profile.http_port, self.profile.ui_port
)),
);
break;
}
if apps_started.elapsed() >= Duration::from_secs(60) {
self.syslog("⚠critical apps not fully loaded after 60s — proceeding anyway");
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Ready,
Some("api ready (some apps slow)".into()),
);
break;
}
if let Some(child) = self.child.lock().unwrap().as_mut() {
if let Ok(Some(status)) = child.try_wait() {
return Err(anyhow!("daemon exited with {} while waiting for apps", status));
}
}
std::thread::sleep(Duration::from_secs(2));
}
Ok(DaemonHandle {
name: self.profile.name.clone(),
banner: format!(
"monorepo daemon [{}] (cargo run from {}, api=http://localhost:{}, \
ui=http://localhost:{}, socket={})",
self.profile.name,
self.monorepo_path.display(),
self.profile.http_port,
self.profile.ui_port,
socket_path.display()
),
socket_path,
dev_dir,
})
}
fn tail_logs(&self, app_name: &str) {
let log_tx = match &self.log_tx {
Some(tx) => tx.clone(),
None => return,
};
let env_dir = match monorepo_env_dir(&self.monorepo_path, &self.profile.name) {
Ok(d) => d,
Err(_) => return,
};
// App SDK logs: {DATA_DIR_PATH}/logs/apps/{app_name}/app.log
let log_file = env_dir
.join("ldk_data")
.join("logs")
.join("apps")
.join(app_name)
.join("app.log");
// Prefix every log line with `[<node>][<app>] ` so AppState::push_log
// routes it into both the per-node and per-app buffers. The outer
// node prefix lets multi-instance platform mode show alice/bob's
// logs separately; the inner app prefix preserves the existing
// per-app filter. App-developer mode (no tracked_nodes) ignores
// the node prefix since no name matches.
let prefix = format!("[{node}][{app}] ", node = self.profile.name, app = app_name);
std::thread::spawn(move || {
use std::io::{BufRead, BufReader, Seek, SeekFrom};
// Wait up to 5s for the log file to appear (created when daemon loads the app).
let deadline = std::time::Instant::now();
loop {
if log_file.exists() {
break;
}
if deadline.elapsed() > std::time::Duration::from_secs(5) {
return;
}
std::thread::sleep(std::time::Duration::from_millis(200));
}
let file = match fs::File::open(&log_file) {
Ok(f) => f,
Err(_) => return,
};
let mut reader = BufReader::new(file);
// Start from the beginning — app startup logs are already written by the time
// tail_logs() is called (ipc_call for app.dev_load completes after init()).
let _ = reader.seek(SeekFrom::Start(0));
loop {
let mut line = String::new();
match reader.read_line(&mut line) {
Ok(0) => {
// EOF — file hasn't grown yet, poll.
std::thread::sleep(std::time::Duration::from_millis(100));
}
Ok(_) => {
let trimmed =
line.trim_end_matches('\n').trim_end_matches('\r');
if !trimmed.is_empty()
&& log_tx
.send(TuiEvent::Log(crate::tui::LogEntry {
source: LogSource::App,
line: format!("{prefix}{trimmed}"),
}))
.is_err()
{
return; // TUI channel closed, exit cleanly
}
}
Err(_) => return,
}
}
});
}
fn shutdown(&self) {
if let Ok(env_dir) = monorepo_env_dir(&self.monorepo_path, &self.profile.name) {
let _ = fs::remove_file(env_dir.join("daemon.pid"));
}
if let Some(mut child) = self.child.lock().unwrap().take() {
let pid = child.id() as i32;
self.syslog(format!("→ shutting down monorepo daemon (PID {pid})…"));
#[cfg(unix)]
unsafe {
libc::kill(-pid, libc::SIGTERM);
}
let t = Instant::now();
loop {
match child.try_wait() {
Ok(Some(_)) => {
self.syslog(format!(
"✓ daemon exited ({}ms)",
t.elapsed().as_millis()
));
break;
}
Ok(None) if t.elapsed() >= Duration::from_secs(10) => {
self.syslog("âš daemon did not exit after 10s, SIGKILL");
#[cfg(unix)]
unsafe {
libc::kill(-pid, libc::SIGKILL);
}
let _ = child.wait();
break;
}
Ok(None) => std::thread::sleep(Duration::from_millis(200)),
Err(e) => {
self.syslog(format!("âš daemon wait error: {e}"));
break;
}
}
}
}
if let Some(mut ui) = self.ui_child.lock().unwrap().take() {
let pid = ui.id() as i32;
self.syslog(format!("→ shutting down UI dev server (PID {pid})…"));
#[cfg(unix)]
unsafe {
libc::kill(-pid, libc::SIGTERM);
}
let t = Instant::now();
loop {
match ui.try_wait() {
Ok(Some(_)) => break,
Ok(None) if t.elapsed() >= Duration::from_secs(5) => {
#[cfg(unix)]
unsafe {
libc::kill(-pid, libc::SIGKILL);
}
let _ = ui.wait();
break;
}
Ok(None) => std::thread::sleep(Duration::from_millis(200)),
Err(_) => break,
}
}
}
}
fn restart(&self) -> Result<()> {
let env_file = self
.env_file
.lock()
.unwrap()
.clone()
.ok_or_else(|| anyhow!("daemon was never started — cannot restart"))?;
self.syslog("→ restart: stopping daemon…");
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Starting,
Some("restarting…".into()),
);
self.shutdown_daemon_only();
// Reap any stale process still holding the backend port. Without
// this, a daemon child that died unclean (panic, OOM) can leave its
// socket in TIME_WAIT and the respawned daemon fails to bind.
self.free_port(self.profile.http_port, "backend");
// Remove stale socket.
let socket_path = self
.socket_override
.clone()
.unwrap_or_else(|| {
monorepo_env_dir(&self.monorepo_path, &self.profile.name)
.ok()
.map(|d| d.join("control.sock"))
.unwrap_or_else(|| PathBuf::from("/tmp/node-control.sock"))
});
let _ = fs::remove_file(&socket_path);
self.syslog("→ restart: spawning daemon…");
let child = self.spawn_daemon(&env_file)?;
let child_pid = child.id();
*self.child.lock().unwrap() = Some(child);
// Update PID file.
if let Ok(env_dir) = monorepo_env_dir(&self.monorepo_path, &self.profile.name) {
let _ = fs::write(env_dir.join("daemon.pid"), child_pid.to_string());
}
// Wait for socket.
let t = Instant::now();
loop {
if UnixStream::connect(&socket_path).is_ok() {
// Socket up — also confirm the HTTP listener answers before
// declaring the restart successful. Otherwise the next /api
// call from the browser races a half-booted daemon.
if let Err(e) = self.wait_http_ready(self.profile.http_port) {
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Failed("HTTP not ready after restart".into()),
None,
);
bail!("daemon restart: HTTP readiness failed: {e:#}");
}
self.syslog(format!(
"✓ daemon restarted ({}s)",
t.elapsed().as_secs()
));
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Ready,
Some("restarted".into()),
);
return Ok(());
}
if t.elapsed() > HEALTH_TIMEOUT {
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Failed("socket not connectable after restart".into()),
None,
);
bail!("daemon did not come up after restart");
}
if let Some(child) = self.child.lock().unwrap().as_mut() {
if let Ok(Some(status)) = child.try_wait() {
tui::update_status(
self.log_tx.as_ref(),
LogSource::Daemon,
ServiceStatus::Failed(format!("exited {}", status)),
None,
);
bail!("daemon exited with {} during restart", status);
}
}
std::thread::sleep(HEALTH_POLL);
}
}
fn pre_start_dev_dir(&self) -> Option<PathBuf> {
monorepo_dev_dir(&self.monorepo_path, &self.profile.name, self.dev_dir_override.as_deref()).ok()
}
}
fn monorepo_env_dir(monorepo_path: &Path, instance_name: &str) -> Result<PathBuf> {
let cache_dir = cache_root()?;
let path_hash = {
let mut h = DefaultHasher::new();
monorepo_path
.canonicalize()
.unwrap_or_else(|_| monorepo_path.to_path_buf())
.hash(&mut h);
instance_name.hash(&mut h);
h.finish()
};
Ok(cache_dir.join(format!("monorepo-{:x}", path_hash)))
}
fn monorepo_dev_dir(monorepo_path: &Path, instance_name: &str, dev_dir_override: Option<&Path>) -> Result<PathBuf> {
if let Some(override_path) = dev_dir_override {
return Ok(override_path.to_path_buf());
}
Ok(monorepo_env_dir(monorepo_path, instance_name)?.join("dev-apps"))
}
fn ipc_list_app_statuses(
socket_path: &std::path::Path,
) -> Result<std::collections::HashMap<String, String>> {
use std::io::{BufRead, BufReader, Write};
let mut stream = UnixStream::connect(socket_path)
.with_context(|| "connect to IPC socket")?;
stream.set_read_timeout(Some(Duration::from_secs(5))).ok();
stream.set_write_timeout(Some(Duration::from_secs(5))).ok();
let request = serde_json::json!({
"jsonrpc": "2.0",
"id": 1,
"method": "app.list",
"params": {}
});
let mut line = serde_json::to_string(&request).unwrap();
line.push('\n');
stream.write_all(line.as_bytes())?;
let reader = BufReader::new(&stream);
let response_line = reader
.lines()
.next()
.ok_or_else(|| anyhow!("no response"))??;
let v: serde_json::Value = serde_json::from_str(&response_line)?;
let mut map = std::collections::HashMap::new();
if let Some(apps) = v.pointer("/result/apps").and_then(|a| a.as_array()) {
for app in apps {
if let (Some(name), Some(status)) = (
app.get("name").and_then(|n| n.as_str()),
app.get("status").and_then(|s| s.as_str()),
) {
map.insert(name.to_string(), status.to_string());
}
}
}
Ok(map)
}
fn which_bin(bin: &str) -> bool {
std::env::var_os("PATH")
.map(|path| std::env::split_paths(&path).any(|dir| dir.join(bin).is_file()))
.unwrap_or(false)
}
fn cache_root() -> Result<PathBuf> {
if let Ok(c) = std::env::var("XDG_CACHE_HOME") {
if !c.is_empty() {
return Ok(PathBuf::from(c).join("node-app"));
}
}
let home = std::env::var_os("HOME").ok_or_else(|| anyhow!("$HOME not set"))?;
Ok(PathBuf::from(home).join(".cache/node-app"))
}