mirror of
https://github.com/stablyai/orca.git
synced 2026-10-03 00:02:19 +00:00
Jinjing
5b8977ffa2
Extracts the built-in/custom agent merge logic from NewWorkspaceComposerCard into buildWorkspaceAgentOptions, so the composer card and the selection resolver share one source of truth for which agents are choosable. Adds e2e coverage for custom-agent launch surfaces (composer, annotations, git actions, settings authoring, SSH) to close gaps left by the earlier seededCustomAgents harness truncation bug.
71 lines
3.0 KiB
TypeScript
71 lines
3.0 KiB
TypeScript
import type { CustomTuiAgentId } from '../../src/shared/types'
|
|
import { test } from './helpers/orca-app'
|
|
import { ensureTerminalVisible, waitForActiveWorktree, waitForSessionReady } from './helpers/store'
|
|
import {
|
|
sendToTerminal,
|
|
waitForActivePanePtyId,
|
|
waitForActiveTerminalManager,
|
|
waitForPaneCount,
|
|
waitForTerminalOutput
|
|
} from './helpers/terminal'
|
|
import {
|
|
buildCustomAgent,
|
|
ECHO_PREFIX,
|
|
expectCustomAgentSeeded,
|
|
launchCustomAgentViaBoundary,
|
|
READY_MARKER
|
|
} from './helpers/custom-agent-e2e'
|
|
|
|
// Annotation surfaces (diff notes, browser annotations) launch a NEW custom
|
|
// agent through QuickLaunchAgentMenuItems → launchAgentInNewTab with
|
|
// launch_source 'notes_send' and promptDelivery 'submit-after-ready': the tab
|
|
// launches bare (allowEmptyPromptLaunch) and the annotation markdown is pasted
|
|
// and submitted once the agent is ready. This spec drives that launch boundary
|
|
// and then proves the annotation content reaches THIS process.
|
|
//
|
|
// Not covered here (deliberately): the "Send notes to <running agent>" followup
|
|
// and the terminal-context-menu fork. Both gate on agent-status readiness the
|
|
// deterministic echo fixture cannot emit (it is not a real TUI agent), and the
|
|
// fork path nulls custom ids in resolveTuiAgent — those are covered by unit
|
|
// tests, not this launch-surface suite.
|
|
const CUSTOM_AGENT = buildCustomAgent({
|
|
uuid: '4d5e6f7a-8b9c-4d0e-8f1a-2b3c4d5e6f70',
|
|
label: 'E2E Annotation Agent'
|
|
})
|
|
const CUSTOM_AGENT_ID = CUSTOM_AGENT.id as CustomTuiAgentId
|
|
|
|
test.use({ seededCustomAgents: { agents: [CUSTOM_AGENT] } })
|
|
|
|
test('launches a custom agent from an annotation send and delivers the note content', async ({
|
|
orcaPage
|
|
}) => {
|
|
await waitForSessionReady(orcaPage)
|
|
const worktreeId = await waitForActiveWorktree(orcaPage)
|
|
await expectCustomAgentSeeded(orcaPage, CUSTOM_AGENT_ID, CUSTOM_AGENT.label)
|
|
|
|
// notes_send launches bare (allowEmptyPromptLaunch): the annotation markdown is
|
|
// delivered post-ready, not folded into the launch. Match that shape by
|
|
// launching without a prompt, then delivering the note once the agent is ready.
|
|
await launchCustomAgentViaBoundary(orcaPage, {
|
|
worktreeId,
|
|
agentId: CUSTOM_AGENT_ID,
|
|
launchSource: 'notes_send'
|
|
})
|
|
|
|
await ensureTerminalVisible(orcaPage)
|
|
await waitForActiveTerminalManager(orcaPage, 30_000)
|
|
await waitForPaneCount(orcaPage, 1, 30_000)
|
|
const ptyId = await waitForActivePanePtyId(orcaPage, 30_000)
|
|
await waitForTerminalOutput(orcaPage, READY_MARKER, 30_000)
|
|
|
|
// The annotation content is delivered post-ready (submit-after-ready). Prove it
|
|
// reaches the launched process: the fixture echoes the submitted line back with
|
|
// a prefix the terminal itself never produces (raw mode, no local echo).
|
|
const ANNOTATION_NOTE = 'review-annotation-payload'
|
|
await sendToTerminal(orcaPage, ptyId, `${ANNOTATION_NOTE}\r`)
|
|
await waitForTerminalOutput(orcaPage, `${ECHO_PREFIX}${ANNOTATION_NOTE}`, 15_000)
|
|
|
|
// Clean shutdown so the daemon/PTY teardown does not race the app close.
|
|
await sendToTerminal(orcaPage, ptyId, '\x03')
|
|
})
|