mandala-computer-mcp 0.1.1 → 0.4.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +139 -24
- package/dist/api.d.ts +19 -6
- package/dist/api.d.ts.map +1 -1
- package/dist/api.js +355 -76
- package/dist/api.js.map +1 -1
- package/dist/cli.d.ts.map +1 -1
- package/dist/cli.js +113 -8
- package/dist/cli.js.map +1 -1
- package/dist/errors.d.ts +100 -12
- package/dist/errors.d.ts.map +1 -1
- package/dist/errors.js +169 -29
- package/dist/errors.js.map +1 -1
- package/dist/events.d.ts +45 -4
- package/dist/events.d.ts.map +1 -1
- package/dist/events.js +422 -114
- package/dist/events.js.map +1 -1
- package/dist/format.d.ts +48 -0
- package/dist/format.d.ts.map +1 -1
- package/dist/format.js +111 -4
- package/dist/format.js.map +1 -1
- package/dist/http-body.d.ts +17 -0
- package/dist/http-body.d.ts.map +1 -0
- package/dist/http-body.js +48 -0
- package/dist/http-body.js.map +1 -0
- package/dist/http.d.ts.map +1 -1
- package/dist/http.js +177 -51
- package/dist/http.js.map +1 -1
- package/dist/index.d.ts +1 -1
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +1 -1
- package/dist/index.js.map +1 -1
- package/dist/limits.d.ts +17 -0
- package/dist/limits.d.ts.map +1 -0
- package/dist/limits.js +17 -0
- package/dist/limits.js.map +1 -0
- package/dist/paths.d.ts +30 -20
- package/dist/paths.d.ts.map +1 -1
- package/dist/paths.js +89 -25
- package/dist/paths.js.map +1 -1
- package/dist/poll.d.ts +107 -0
- package/dist/poll.d.ts.map +1 -0
- package/dist/poll.js +233 -0
- package/dist/poll.js.map +1 -0
- package/dist/server.d.ts +1 -1
- package/dist/server.d.ts.map +1 -1
- package/dist/server.js +2 -1
- package/dist/server.js.map +1 -1
- package/dist/tools/agent.d.ts.map +1 -1
- package/dist/tools/agent.js +72 -5
- package/dist/tools/agent.js.map +1 -1
- package/dist/tools/computers.d.ts.map +1 -1
- package/dist/tools/computers.js +559 -233
- package/dist/tools/computers.js.map +1 -1
- package/dist/tools/events.d.ts.map +1 -1
- package/dist/tools/events.js +359 -69
- package/dist/tools/events.js.map +1 -1
- package/dist/tools/guest.d.ts.map +1 -1
- package/dist/tools/guest.js +234 -33
- package/dist/tools/guest.js.map +1 -1
- package/dist/tools/input.d.ts.map +1 -1
- package/dist/tools/input.js +92 -8
- package/dist/tools/input.js.map +1 -1
- package/dist/tools/snapshots.d.ts.map +1 -1
- package/dist/tools/snapshots.js +501 -33
- package/dist/tools/snapshots.js.map +1 -1
- package/dist/tools/templates.d.ts.map +1 -1
- package/dist/tools/templates.js +61 -26
- package/dist/tools/templates.js.map +1 -1
- package/dist/tools/webhooks.d.ts.map +1 -1
- package/dist/tools/webhooks.js +116 -17
- package/dist/tools/webhooks.js.map +1 -1
- package/package.json +3 -2
package/README.md
CHANGED
|
@@ -19,7 +19,24 @@ the way you would treat a password.
|
|
|
19
19
|
Node 20.3 or newer. There is nothing else to install: `npx` fetches the server
|
|
20
20
|
the first time a client starts it.
|
|
21
21
|
|
|
22
|
-
**Claude Code**
|
|
22
|
+
**Claude Code** — as a plugin, which installs the server and a skill together:
|
|
23
|
+
|
|
24
|
+
```sh
|
|
25
|
+
export MANDALA_API_KEY=com_… # in the shell Claude Code starts from
|
|
26
|
+
/plugin marketplace add mandalacomputer/mcp
|
|
27
|
+
/plugin install mandala-computer@mandala
|
|
28
|
+
```
|
|
29
|
+
|
|
30
|
+
The skill — [`plugin/skills/mandala-computer/SKILL.md`](plugin/skills/mandala-computer/SKILL.md)
|
|
31
|
+
— is the part the tools cannot say for themselves: when a cloud desktop is the
|
|
32
|
+
right answer at all, that it costs money until it is suspended or stopped, that
|
|
33
|
+
`run_agent` is usually the right level and a screenshot per click is not, and
|
|
34
|
+
which refusals are worth a second try. It is a description of *when and how*,
|
|
35
|
+
not a second client; once the server is installed it stays out of the way.
|
|
36
|
+
`MANDALA_MODEL_KEY`, if exported alongside, is passed through and turns on
|
|
37
|
+
`run_agent`.
|
|
38
|
+
|
|
39
|
+
Or the server on its own, with the key inline:
|
|
23
40
|
|
|
24
41
|
```sh
|
|
25
42
|
claude mcp add mandala -e MANDALA_API_KEY=com_… -- npx -y mandala-computer-mcp
|
|
@@ -122,8 +139,14 @@ already supports for when they are.
|
|
|
122
139
|
**A running computer costs money, and a forgotten one keeps costing it.**
|
|
123
140
|
`create_computer` says so in its own description, and so does everything that
|
|
124
141
|
starts a machine by a side door — `restore_snapshot` boots a stopped computer,
|
|
125
|
-
`write_clipboard`
|
|
126
|
-
|
|
142
|
+
and `write_clipboard`, `read_file` and `cursor_position` each resume a suspended
|
|
143
|
+
one, because each has to reach the guest agent to do its job. All of them are
|
|
144
|
+
charged like any other start, and on a plan at its limit can come back 402
|
|
145
|
+
rather than a result. The two reads are the surprising half of that list, and
|
|
146
|
+
are why neither is annotated `readOnlyHint`: a host that auto-approves
|
|
147
|
+
read-only tools would otherwise wake and bill a machine with nobody asked.
|
|
148
|
+
`screenshot`, `read_clipboard` and `list_windows` genuinely do not start
|
|
149
|
+
anything and keep the hint. When a stretch of work is over, `suspend_computer` (a pause:
|
|
127
150
|
`start_computer` brings the same session back in about a second) or
|
|
128
151
|
`stop_computer` (a shutdown: the disk is kept, the session is not). Idle
|
|
129
152
|
suspend catches the ones a model forgets, but only after 30 minutes untouched.
|
|
@@ -134,9 +157,11 @@ time — `screenshot`, `click`, `screenshot` — puts an image in the calling
|
|
|
134
157
|
model's context for every step. `run_agent` hands a task in plain language to
|
|
135
158
|
the platform's own loop instead, which screenshots, decides and clicks inside
|
|
136
159
|
the platform and answers with a sentence and the list of what it did. It is
|
|
137
|
-
registered only when a model key is present (see [Configuration](#configuration))
|
|
138
|
-
bills that key for
|
|
139
|
-
the
|
|
160
|
+
registered only when a model key is present (see [Configuration](#configuration))
|
|
161
|
+
and bills that key for the run. `max_steps` bounds the WORK rather than the
|
|
162
|
+
bill: a step is one action on the desktop, one model reply can ask for several
|
|
163
|
+
and spends a step on each, a paused turn costs tokens and no step, and not every
|
|
164
|
+
step takes a screenshot. It defaults to 20 and is capped at 100 here.
|
|
140
165
|
|
|
141
166
|
**A screenshot is how you find out what the screen looks like.** A click that
|
|
142
167
|
landed and a click that did nothing produce the same tool result, so a model
|
|
@@ -258,6 +283,44 @@ they are doing — it does not receive them and has no `verify`, because a
|
|
|
258
283
|
server with no endpoint has nothing to verify. `list_webhook_deliveries` is
|
|
259
284
|
where a delivery that ran out of retries shows up; nothing is dropped silently.
|
|
260
285
|
|
|
286
|
+
**A capture outlives the request that starts it.** `POST /computers/:id/snapshots`
|
|
287
|
+
answers `202` with a placeholder row and copies the disk afterwards, which takes
|
|
288
|
+
minutes and scales with how much has been written — longer than any HTTP request
|
|
289
|
+
survives. So `create_snapshot` polls: it holds the id the platform allocated
|
|
290
|
+
before the copy, watches `list_snapshots` for that row to stop reading
|
|
291
|
+
`capturing`, and answers with the finished snapshot. It waits for *not
|
|
292
|
+
capturing* rather than for `pending`, because replication can carry a small
|
|
293
|
+
snapshot straight on to `durable` between two polls; and it matches on the id
|
|
294
|
+
rather than on the newest row for the computer, because a scheduled capture
|
|
295
|
+
finishing in the same window makes that wrong on exactly the long captures where
|
|
296
|
+
it matters. `wait: false` hands back the id instead, for a caller that would
|
|
297
|
+
rather poll on its own schedule.
|
|
298
|
+
|
|
299
|
+
The three answers it can end on are kept apart deliberately. A capture that
|
|
300
|
+
lands is the snapshot. A capture that FAILS mid-copy leaves no snapshot and no
|
|
301
|
+
row — the `capturing` row simply disappears, and that absence is the only signal
|
|
302
|
+
there is, which is why an incomplete listing is never allowed to decide it. A
|
|
303
|
+
wait that runs out says the capture is still running and names the id to follow,
|
|
304
|
+
because that one asks for a look rather than for another attempt.
|
|
305
|
+
|
|
306
|
+
**A deletion outlives its request too, and reads the other way round.**
|
|
307
|
+
`DELETE /snapshots/:id` answers `202` and then detaches the dependent snapshots
|
|
308
|
+
and removes the stored objects. There is no state that means deleted, so
|
|
309
|
+
`delete_snapshot` polls for the row to **go** — the mirror of a capture, which
|
|
310
|
+
polls for a row to stay and change. A row that stays is one that STALLED, and it
|
|
311
|
+
sits in `deleting`, a state a bare listing hides: the poll asks with
|
|
312
|
+
`include=unfinished` for exactly that reason, since without it a half-deleted
|
|
313
|
+
snapshot is indistinguishable from a deleted one. The platform retries a stalled
|
|
314
|
+
deletion itself every fifteen minutes, so the give-up sentence says to watch
|
|
315
|
+
rather than to repeat.
|
|
316
|
+
|
|
317
|
+
Because absence is the success signal here, "Deleted" is said only off a listing
|
|
318
|
+
read **whole**. A row missing because a hypervisor did not answer is not a row
|
|
319
|
+
that is gone, and that mistake is unrecoverable in a way the others are not:
|
|
320
|
+
nobody goes looking for a snapshot they have been told was destroyed. A `409`
|
|
321
|
+
saying the snapshot is already being deleted is progress, not a fault — the
|
|
322
|
+
answer is to watch that deletion finish, never to go and delete something else.
|
|
323
|
+
|
|
261
324
|
**A schedule says when, not how long.** `snapshot_schedule` sets the window a
|
|
262
325
|
computer's automatic snapshot is taken in; `get_retention` is what says how many
|
|
263
326
|
of them survive, and it takes no computer because the window belongs to the
|
|
@@ -304,9 +367,12 @@ nothing; `background: true` is the only thing that works. The abandoned command
|
|
|
304
367
|
keeps running, so the call after one of these often reports the guest agent as
|
|
305
368
|
busy — that is the first failure continuing, not a second one.
|
|
306
369
|
|
|
307
|
-
|
|
308
|
-
`timeout_s`
|
|
309
|
-
|
|
370
|
+
That roughly two-minute ceiling belongs to the hosted proxy. The server accepts
|
|
371
|
+
integer foreground `timeout_s` values from 1 through 600 seconds (default 30),
|
|
372
|
+
and the tool exposes that range for a `MANDALA_BASE_URL` reached without the
|
|
373
|
+
proxy. The HTTP client allows 630 seconds for response headers so the server can
|
|
374
|
+
report a 600-second timeout. Use `background: true` and poll the handle for
|
|
375
|
+
longer work.
|
|
310
376
|
|
|
311
377
|
**`list_windows` sees what a screenshot cannot.** It is how you tell an
|
|
312
378
|
application that failed to start from one that has not painted yet. Match on
|
|
@@ -362,11 +428,25 @@ simply has fewer rows, an unknown number missing and nothing marking the gap.
|
|
|
362
428
|
The other two append a row marked `unreachable` for each thing they could not
|
|
363
429
|
reach — but only for a key that spans the account. A WORKSPACE-SCOPED key gets
|
|
364
430
|
no marked rows either, because naming the missing ids would mean reading them
|
|
365
|
-
out of a
|
|
431
|
+
out of a host cache that has no workspace column, and handing a confined
|
|
366
432
|
credential ids from the workspaces it is confined away from. For such a key all
|
|
367
433
|
three listings are the `INCOMPLETE:` line and nothing else, which is why that
|
|
368
434
|
line is written first and in prose.
|
|
369
435
|
|
|
436
|
+
**A computer has a lifecycle of its own, separate from what its guest is
|
|
437
|
+
doing.** `state` is the platform's record of whether the machine exists —
|
|
438
|
+
`live`, `deleting`, `deleted`, `lost`, or `unreachable` when a listing could not
|
|
439
|
+
confirm the row against its host — while `status` is the host's answer about the
|
|
440
|
+
guest. A row served from the record has the first and not the second, so
|
|
441
|
+
`list_computers` prints both when both are there: a computer can be running and
|
|
442
|
+
being deleted at once.
|
|
443
|
+
|
|
444
|
+
An unfiltered listing is `live`, `unreachable` and `deleting`. The two terminal
|
|
445
|
+
states are withheld from it, so `list_computers(state: 'deleted')` is the only
|
|
446
|
+
way a computer that has gone is ever shown — and an empty answer to a filtered
|
|
447
|
+
listing is a fact about the filter rather than about the account, which is what
|
|
448
|
+
it says rather than inviting you to create one.
|
|
449
|
+
|
|
370
450
|
**Snapshots mid-deletion are billed but hidden.** A deletion that began and did
|
|
371
451
|
not finish still holds objects and still counts against storage, and the default
|
|
372
452
|
listing leaves it out — every ordinary caller is asking "what can I restore".
|
|
@@ -382,6 +462,10 @@ by default — the platform drops input on that socket, so it is safe to hand to
|
|
|
382
462
|
somebody. `control: true` returns the full-control one, which is root-equivalent
|
|
383
463
|
on that machine. Neither appears in any other tool's output, deliberately: a
|
|
384
464
|
tool result lands in a model's context and from there in whatever captured it.
|
|
465
|
+
It is also why `get_desktop_url` carries no `readOnlyHint` even though its route
|
|
466
|
+
neither writes nor spends: hosts treat that hint as licence to call without
|
|
467
|
+
asking, so keeping it would let a model pass out control of a desktop with
|
|
468
|
+
nobody prompted.
|
|
385
469
|
|
|
386
470
|
**Retiring a template cannot be undone, and takes more than it looks.**
|
|
387
471
|
`retire_template` without a `version` retires **every** version of the name —
|
|
@@ -411,18 +495,47 @@ default request timeout is 60 seconds and only a progress notification can reset
|
|
|
411
495
|
it — but the SDK resets it only for a caller that passed that option, so a client
|
|
412
496
|
which merely accepts progress is still cancelled a minute into a fifteen-minute
|
|
413
497
|
build. `get_build` is the answer for a client that cannot hold a request open:
|
|
414
|
-
it reads once and returns.
|
|
498
|
+
it reads once and returns.
|
|
499
|
+
|
|
500
|
+
**The same applies to every tool here that waits.** `wait_for_computer`,
|
|
501
|
+
`move_computer`, `create_snapshot` and `delete_snapshot` all poll, and all report
|
|
502
|
+
progress on every poll — a changed line at once, an unchanged one on a ten-second
|
|
503
|
+
heartbeat, so the request stays open without flooding the client. A live capture
|
|
504
|
+
took 107 seconds and was cancelled at 60 before this existed, which turned every
|
|
505
|
+
carefully-worded answer about what had actually happened into a transport error.
|
|
506
|
+
Each of them has a way out for a client that cannot opt in: `wait: false` on the
|
|
507
|
+
two snapshot tools, `list_moves` after a move, a shorter `timeout_s` and a second
|
|
508
|
+
call on the wait. A build that *failed* is a normal answer from
|
|
415
509
|
`watch_build`, not an error — it names the step that stopped it, which is the
|
|
416
510
|
thing to fix. An `error` event is the *stream* failing and says nothing about the
|
|
417
511
|
build, and the tool says so rather than letting a model rewrite a document that
|
|
418
512
|
is fine.
|
|
419
513
|
|
|
420
|
-
|
|
421
|
-
|
|
422
|
-
|
|
423
|
-
|
|
424
|
-
|
|
425
|
-
|
|
514
|
+
A completed template build can launch with `create_computer` when eligible
|
|
515
|
+
capacity and its image are available. Publishing a template and completing its
|
|
516
|
+
build are separate steps; publication alone does not guarantee launch capacity.
|
|
517
|
+
|
|
518
|
+
When the API supports image preparation continuation, a `create_computer`
|
|
519
|
+
refusal with code `template_image_preparing` preserves `template_transfer` and
|
|
520
|
+
`preparation` (including `state` and `error`) in the JSON appended to the tool's
|
|
521
|
+
error text. `retry_after_ms`, when present, is the parsed `Retry-After` delay in
|
|
522
|
+
**milliseconds**, also available to embedders as `APIError.retryAfterMs`. Both
|
|
523
|
+
integer seconds and HTTP dates are accepted; missing or invalid delays are
|
|
524
|
+
omitted. Delays are capped at 2,147,483,647 milliseconds to fit a timer.
|
|
525
|
+
|
|
526
|
+
For `preparing`, `copying`, or `ready` with a usable token and delay, wait for
|
|
527
|
+
that delay, then repeat the original identical create arguments, including
|
|
528
|
+
`template`, adding the token exactly as returned. The token must be a nonblank
|
|
529
|
+
string, requires the original `template`, and cannot be combined with `size`.
|
|
530
|
+
A `failed` preparation reports its error for inspection before deciding what to
|
|
531
|
+
do next. Missing or unknown states and missing or invalid delays do not supply
|
|
532
|
+
automatic retry advice. The server never replays a create automatically, and
|
|
533
|
+
`isTransient` returns false for this refusal code because unchanged replay is
|
|
534
|
+
not the continuation protocol.
|
|
535
|
+
|
|
536
|
+
The token selects the image build; it is **not a create idempotency key**. Stop
|
|
537
|
+
after success. After a lost or ambiguous response, inspect `list_computers`
|
|
538
|
+
before deciding what to do; never automatically replay the create.
|
|
426
539
|
|
|
427
540
|
## Running it as a service
|
|
428
541
|
|
|
@@ -472,7 +585,7 @@ naming the fix.
|
|
|
472
585
|
| `MANDALA_BASE_URL` | Defaults to `https://app.mandala.computer/api/v1`. |
|
|
473
586
|
| `MANDALA_COMPUTER_ID` | Bind a computer at startup, so `use_computer` is not needed. **stdio only** — under `--http` it is ignored rather than bound into every caller's session, since it names a machine on the operator's account. |
|
|
474
587
|
| `MANDALA_MODEL_KEY` | An Anthropic key. Enables `run_agent`, which runs the platform's own loop on that key. **stdio only** — under `--http` each caller sends their own as `X-Model-Key`, and this variable is ignored. |
|
|
475
|
-
| `MANDALA_NO_LIFECYCLE` | `1` withholds `create_computer`, `clone_computer`, `clone_snapshot`, `delete_computer` and `delete_snapshot` — every tool that makes a computer or destroys one. |
|
|
588
|
+
| `MANDALA_NO_LIFECYCLE` | `1`, `true`, `yes` or `on` withholds `create_computer`, `clone_computer`, `clone_snapshot`, `delete_computer` and `delete_snapshot` — every tool that makes a computer or destroys one. `0`, `false`, `no`, `off` or unset leaves them registered. Any other value is **refused at startup** rather than read as off: a typo here would otherwise leave those tools in place on a server whose operator believes they are gone. The `--no-lifecycle` flag reads the same vocabulary and refuses the same way, except that it has no spelling for *unset*: `--no-lifecycle=` is refused rather than ignored, so a launcher template whose variable did not expand stops instead of quietly leaving the tools registered. |
|
|
476
589
|
| `PORT`, `HOST` | For `--http`. Default `3000`, `127.0.0.1`. |
|
|
477
590
|
| `MANDALA_ALLOWED_HOSTS`, `MANDALA_ALLOWED_ORIGINS` | Comma-separated. Which `Host` and `Origin` values this server answers to. On a loopback bind the host list defaults to the address it was given, so DNS-rebinding protection is on without configuration; set this when serving under a name. |
|
|
478
591
|
|
|
@@ -519,14 +632,16 @@ written down in `UNIMPLEMENTED`. A route added upstream becomes a failing test
|
|
|
519
632
|
here rather than a feature nobody noticed.
|
|
520
633
|
|
|
521
634
|
`npm run check:surface` goes further and diffs the mirror against the platform's
|
|
522
|
-
|
|
523
|
-
|
|
524
|
-
|
|
525
|
-
|
|
526
|
-
|
|
635
|
+
published surface manifest — a file the platform generates from its own tables
|
|
636
|
+
and commits like a lockfile — whenever the platform repository happens to be
|
|
637
|
+
checked out next door, or wherever `MANDALA_PLATFORM_REPO` points. Without it
|
|
638
|
+
the script says it is skipping and exits 0, which is what it does for anyone
|
|
639
|
+
outside the platform team; a manifest it cannot read is a failure, never a
|
|
640
|
+
comparison of nothing. The diff is enforced from the platform's own CI, which
|
|
641
|
+
checks this repository out beside itself and runs the same script.
|
|
527
642
|
|
|
528
643
|
```
|
|
529
|
-
check:surface — the mirror matches the platform (
|
|
644
|
+
check:surface — the mirror matches the platform (N routes, N parameters, from …).
|
|
530
645
|
```
|
|
531
646
|
|
|
532
647
|
## See also
|
package/dist/api.d.ts
CHANGED
|
@@ -1,3 +1,4 @@
|
|
|
1
|
+
import { Agent } from 'undici';
|
|
1
2
|
export declare const DEFAULT_BASE_URL = "https://app.mandala.computer/api/v1";
|
|
2
3
|
/** Anthropic's own key, forwarded for the one route that runs a model. */
|
|
3
4
|
export declare const MODEL_KEY_HEADER = "X-Model-Key";
|
|
@@ -54,10 +55,10 @@ export type Bytes = {
|
|
|
54
55
|
unrangeable: boolean;
|
|
55
56
|
};
|
|
56
57
|
/**
|
|
57
|
-
* The longest guest exec waits
|
|
58
|
-
* fetch
|
|
59
|
-
* lose that race while the command is still finishing in the guest.
|
|
60
|
-
*
|
|
58
|
+
* The longest foreground guest exec waits 600 seconds before it answers.
|
|
59
|
+
* Node's bundled fetch gives response headers 300 seconds by default, so the
|
|
60
|
+
* client can lose that race while the command is still finishing in the guest.
|
|
61
|
+
* Allow 30 seconds beyond the public exec limit for the timeout response.
|
|
61
62
|
*
|
|
62
63
|
* The body is a different clock. undici's default `bodyTimeout` is 300 seconds
|
|
63
64
|
* of silence *between chunks*, and `run_agent` SSE (or a long exec that has
|
|
@@ -66,9 +67,10 @@ export type Bytes = {
|
|
|
66
67
|
* disables it: a quiet gap is not a dead connection, and the caller's
|
|
67
68
|
* AbortSignal is what ends a request nobody is waiting for.
|
|
68
69
|
*/
|
|
69
|
-
export declare const PLATFORM_HEADERS_TIMEOUT_MS =
|
|
70
|
+
export declare const PLATFORM_HEADERS_TIMEOUT_MS = 630000;
|
|
70
71
|
/** Disabled. A finite idle limit is what used to kill a quiet SSE stream. */
|
|
71
72
|
export declare const PLATFORM_BODY_TIMEOUT_MS = 0;
|
|
73
|
+
export declare const PLATFORM_DISPATCHER: Agent;
|
|
72
74
|
/**
|
|
73
75
|
* The fetch a platform request actually goes through, and why it is not simply
|
|
74
76
|
* `fetch`.
|
|
@@ -160,7 +162,7 @@ export declare class Api {
|
|
|
160
162
|
* short 200 — but a caller that opts in gets the list plus `X-GC-Incomplete`,
|
|
161
163
|
* and a header is only a warning if something reads it.
|
|
162
164
|
*
|
|
163
|
-
* It is the count of what the
|
|
165
|
+
* It is the count of what the host cache could account for, and it is
|
|
164
166
|
* legitimately `0`: a computer created during the outage was never cached
|
|
165
167
|
* against the host now holding it. So presence is the signal and the number is
|
|
166
168
|
* detail, which is why this returns `null` versus a number rather than a
|
|
@@ -181,6 +183,17 @@ export declare class Api {
|
|
|
181
183
|
*/
|
|
182
184
|
sse(method: string, path: string, opts?: RequestOptions): AsyncGenerator<SSEEvent>;
|
|
183
185
|
}
|
|
186
|
+
/**
|
|
187
|
+
* Every error under one, including the ones a fetch hides two levels down.
|
|
188
|
+
*
|
|
189
|
+
* A rejected fetch is a `TypeError: fetch failed` whose `cause` is what
|
|
190
|
+
* actually went wrong, and on a dual-stack host that cause is an
|
|
191
|
+
* `AggregateError` holding one attempt per address. Neither the top error nor
|
|
192
|
+
* its immediate cause carries the code the classifiers below read, so both
|
|
193
|
+
* links have to be followed. Bounded, because a cause chain is user-reachable
|
|
194
|
+
* data and nothing here needs to be robust to a cycle.
|
|
195
|
+
*/
|
|
196
|
+
export declare function causes(err: unknown, depth?: number): Generator<Record<string, unknown>>;
|
|
184
197
|
/** The filename the platform put on a download, if it put one there. */
|
|
185
198
|
export declare function filenameFrom(disposition: string | null): string | undefined;
|
|
186
199
|
//# sourceMappingURL=api.d.ts.map
|
package/dist/api.d.ts.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"api.d.ts","sourceRoot":"","sources":["../src/api.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"api.d.ts","sourceRoot":"","sources":["../src/api.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAmE,MAAM,QAAQ,CAAC;AAYhG,eAAO,MAAM,gBAAgB,wCAAwC,CAAC;AAEtE,0EAA0E;AAC1E,eAAO,MAAM,gBAAgB,gBAAgB,CAAC;AAE9C,MAAM,MAAM,cAAc,GAAG;IAC3B,KAAK,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,GAAG,OAAO,GAAG,SAAS,CAAC,CAAC;IAC9D,IAAI,CAAC,EAAE,OAAO,CAAC;IACf,0FAA0F;IAC1F,GAAG,CAAC,EAAE,UAAU,CAAC;IACjB,uEAAuE;IACvE,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACjC,MAAM,CAAC,EAAE,WAAW,CAAC;CACtB,CAAC;AAEF,MAAM,MAAM,KAAK,GAAG;IAClB,KAAK,EAAE,UAAU,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,kEAAkE;IAClE,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,2EAA2E;IAC3E,SAAS,EAAE,OAAO,CAAC;IACnB;;;;;;OAMG;IACH,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB;;;;;;;;;;;;OAYG;IACH,MAAM,CAAC,EAAE;QAAE,KAAK,EAAE,MAAM,CAAC;QAAC,GAAG,EAAE,MAAM,CAAC;QAAC,KAAK,CAAC,EAAE,MAAM,CAAA;KAAE,CAAC;IACxD;;;;;;;OAOG;IACH,WAAW,EAAE,OAAO,CAAC;CACtB,CAAC;AA6BF;;;;;;;;;;;;GAYG;AACH,eAAO,MAAM,2BAA2B,SAAU,CAAC;AACnD,6EAA6E;AAC7E,eAAO,MAAM,wBAAwB,IAAI,CAAC;AAC1C,eAAO,MAAM,mBAAmB,OAG9B,CAAC;AAWH;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AACH,eAAO,MAAM,aAAa,QAAO,OAAO,UAAU,CAAC,KAG7B,CAAC;AAEvB,iDAAiD;AACjD,MAAM,MAAM,QAAQ,GAAG;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,IAAI,EAAE,OAAO,CAAA;CAAE,CAAC;AAExD;;;;;;;;;;;GAWG;AACH,qBAAa,GAAG;;IACd,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;gBAQb,MAAM,EAAE,MAAM,EAAE,OAAO,GAAE,MAAyB,EAAE,MAAM,CAAC,EAAE,WAAW;IAoDpF;;;;;;;;;;;OAWG;IACH,IAAI,CAAC,MAAM,EAAE,WAAW,GAAG,SAAS,GAAG,GAAG;IAiP1C;;;;;;;;;;;;OAYG;IACG,IAAI,CAAC,CAAC,GAAG,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,GAAE,cAAmB,GAAG,OAAO,CAAC,CAAC,CAAC;IAW5F;;;;;;OAMG;IACG,IAAI,CAAC,CAAC,GAAG,OAAO,EACpB,MAAM,EAAE,MAAM,EACd,IAAI,EAAE,MAAM,EACZ,IAAI,GAAE,cAAmB,GACxB,OAAO,CAAC,CAAC,GAAG,SAAS,CAAC;IAKzB;;;;;;;;;;;;;;OAcG;IACG,OAAO,CAAC,CAAC,EACb,IAAI,EAAE,MAAM,EACZ,IAAI,GAAE,cAAmB,GACxB,OAAO,CAAC;QAAE,KAAK,EAAE,CAAC,GAAG,SAAS,CAAC;QAAC,UAAU,EAAE,MAAM,GAAG,IAAI,CAAA;KAAE,CAAC;IAY/D,kFAAkF;IAC5E,KAAK,CACT,MAAM,EAAE,MAAM,EACd,IAAI,EAAE,MAAM,EACZ,IAAI,GAAE,cAAmB,EACzB,QAAQ,CAAC,EAAE,MAAM,GAAG,CAAC,CAAC,WAAW,EAAE,MAAM,KAAK,MAAM,CAAC,GACpD,OAAO,CAAC,KAAK,CAAC;IAoFjB;;;;;;OAMG;IACI,GAAG,CAAC,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,GAAE,cAAmB,GAAG,cAAc,CAAC,QAAQ,CAAC;CAsF9F;AAyCD;;;;;;;;;GASG;AACH,wBAAiB,MAAM,CAAC,GAAG,EAAE,OAAO,EAAE,KAAK,SAAI,GAAG,SAAS,CAAC,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAQnF;AAkgBD,wEAAwE;AACxE,wBAAgB,YAAY,CAAC,WAAW,EAAE,MAAM,GAAG,IAAI,GAAG,MAAM,GAAG,SAAS,CAmD3E"}
|