Merge pull request 'forge CLI pivot: rewrite gitea+gitlab clients to tea/glab transports' (#9) from feat/forge-cli-pivot into custom
This commit is contained in:
@@ -2,17 +2,17 @@
|
||||
|
||||
## Purpose
|
||||
|
||||
- This module owns Gitea/Forgejo auth (Personal Access Token), raw REST v1 client access, remote-URL repo resolution, and Gitea issue / pull-request (PR) APIs for OpenChamber, including issue create/update and PR create/update/merge writes.
|
||||
- This module owns Gitea/Forgejo auth (Personal Access Token), CLI-backed REST v1 client access, remote-URL repo resolution, and Gitea issue / pull-request (PR) APIs for OpenChamber, including issue create/update and PR create/update/merge writes.
|
||||
- From a user perspective, this is the layer that lets the app show Gitea issues and pull requests for a local project, including comments and per-file diffs, and create, edit, and merge pull requests.
|
||||
- Gitea and Forgejo share the same GitHub-style REST v1 API, so this module serves both. Gitea calls remote work **pull requests** (PR), not merge requests. Gitea repos are flat `owner/repo` — there are no multi-segment namespaces.
|
||||
- The module mirrors `packages/web/server/lib/gitlab/` (PAT auth + raw-fetch client) but uses a **Personal Access Token** against the `Authorization: token <pat>` header and a **user-supplied base URL** (Gitea is self-hosted; codeberg.org is the only built-in default).
|
||||
- The module mirrors `packages/web/server/lib/gitlab/` but uses the **`tea` CLI** as its transport instead of raw `fetch`. Auth is passed via the `GITEA_SERVER_TOKEN` environment variable (never on argv). The `tea` binary path is env-overridable via `TEA_BIN`, defaulting to `/home/user/.local/bin/tea`.
|
||||
|
||||
## Entrypoints and structure
|
||||
|
||||
- `packages/web/server/lib/gitea/index.js`: public server entrypoint re-exports.
|
||||
- `packages/web/server/lib/gitea/routes.js`: Express route registration for `/api/gitea/*` endpoints.
|
||||
- `packages/web/server/lib/gitea/auth.js`: PAT auth storage, multi-account support, base URL normalization.
|
||||
- `packages/web/server/lib/gitea/client.js`: raw `fetch` Gitea REST v1 client (timeout, ETag conditional GET, rate-limit cooldown, `Link`-header pagination, redirect handling).
|
||||
- `packages/web/server/lib/gitea/client.js`: CLI-backed `tea api` client (process spawn with 8s timeout, `--include` for HTTP status/headers, `--paginate` for list endpoints, `--header` for raw diff Accept). Token is passed via `GITEA_SERVER_TOKEN` env var.
|
||||
- `packages/web/server/lib/gitea/client.d.ts`: hand-written type declaration for `client.js` (the module is plain JS); consumed by the live-test harness.
|
||||
- `packages/web/server/lib/gitea/repo.js`: Gitea remote URL parsing (flat `owner/repo`) and directory-to-repo resolution.
|
||||
- `packages/web/server/lib/opencode/feature-routes-runtime.js`: API route layer that calls this module (via `registerGiteaRoutes`).
|
||||
@@ -56,13 +56,13 @@
|
||||
|
||||
## Client behavior
|
||||
|
||||
- Base URL joining: `{baseUrl}/api/v1{path}`. Gitea repos are flat `owner/repo`, so owner/repo segments are interpolated directly (single path segments, no encoding needed).
|
||||
- Per-request timeout: 8000 ms via `AbortSignal.timeout`, unless the caller passes its own signal.
|
||||
- ETag conditional-GET cache: keyed `token\nurl`, max 300 LRU entries; a `304` is replayed from cache as a `200`. GET only.
|
||||
- Pagination: Gitea list endpoints return a `Link` header (`rel="next"`) plus `X-Total-Count`; both are parsed into the returned `page` object (`hasMore` = a next page exists). List requests use `page` + `limit` query params (Gitea caps `limit` at 50).
|
||||
- Redirects: `301`/`302`/`308` with a `Location` header are followed exactly once with `redirect: 'manual'`, preserving the `Authorization` header across the hop.
|
||||
- Rate limits: a `429` records a module-level cooldown (honoring `Retry-After` seconds / `X-RateLimit-Reset` Unix seconds when present) and surfaces `{ status: 429, error: 'Gitea rate limited' }`. While the cooldown is active, requests short-circuit without hitting the network.
|
||||
- Transport: each call spawns a `tea api` process with `--include` for HTTP status/headers. Auth is passed via `GITEA_SERVER_TOKEN` env var (never on argv). Binary path: `TEA_BIN` env or `/home/user/.local/bin/tea`.
|
||||
- Base URL: the `baseUrl` parameter is passed to the client constructor for compatibility but `tea` resolves the instance from its own login config. The `--include` flag provides HTTP status codes and response headers.
|
||||
- Per-request timeout: 8000 ms via `AbortSignal.timeout` on the spawned process. The process is killed with `SIGKILL` on timeout.
|
||||
- Pagination: `--paginate` is passed for GET requests with query params, causing `tea` to fetch all pages in a single call. The returned `page` object is `null` since pagination is handled by the CLI.
|
||||
- Raw diffs: `--header 'Accept: text/plain'` is passed for the `.diff` endpoint to get raw text output.
|
||||
- `request` never throws for HTTP error statuses — callers branch on `status`. The `raw: true` option returns the response body as text (used for the `.diff` endpoint).
|
||||
- ETag cache and rate-limit cooldown have been dropped with the CLI pivot. Each call spawns a fresh process, so there is no persistent connection for conditional requests or shared rate-limit state. `isGiteaRateLimited()` always returns `false`; `noteGiteaRateLimit()` is a no-op.
|
||||
|
||||
## API integration overview
|
||||
|
||||
@@ -149,6 +149,6 @@ Conventions mirror `github/routes.js` and `gitlab/routes.js`:
|
||||
|
||||
- Keep the response shapes in lockstep with `Gitea*` types in `packages/ui/src/lib/api/types.ts`.
|
||||
- Never log tokens. Error messages must not include the access token.
|
||||
- The ETag cache and rate-limit cooldown are module-level and per-instance — they are NOT shared with the GitHub or GitLab modules.
|
||||
- The `tea` CLI handles authentication, base URL resolution, and pagination internally. The client does not maintain its own ETag cache or rate-limit cooldown — each call spawns a fresh process.
|
||||
- Gitea `GET /user` returns `login`/`full_name`/`html_url`; the route mappers accept the GitHub-style `username`/`name`/`web_url` variants too, so Forgejo versions that differ still map.
|
||||
- To add further Gitea write operations, add the endpoint in `routes.js`, add a convenience method in `client.js`, and extend the shared types — mirror the existing issue/PR write routes and the GitHub PR write routes.
|
||||
|
||||
@@ -1,197 +1,147 @@
|
||||
import { spawn } from 'child_process';
|
||||
import { getGiteaAuth } from './auth.js';
|
||||
import { getProviderApiBaseUrl } from '../git-providers/config.js';
|
||||
import { getEffectiveProviderApiBaseUrl } from '../git-providers/project-config.js';
|
||||
|
||||
// Per-request timeout for every Gitea call. Self-hosted instances can hang
|
||||
// under load; bounding each request lets the caller fail fast and serve
|
||||
// cached/last-known state instead of holding a socket open.
|
||||
const TEA_BIN = process.env.TEA_BIN || '/home/user/.local/bin/tea';
|
||||
const REQUEST_TIMEOUT_MS = 8000;
|
||||
|
||||
const timeoutFetch = (url, options = {}) => {
|
||||
// Respect a caller-provided signal if present; otherwise attach our timeout.
|
||||
if (options.signal) {
|
||||
return fetch(url, options);
|
||||
}
|
||||
return fetch(url, { ...options, signal: AbortSignal.timeout(REQUEST_TIMEOUT_MS) });
|
||||
};
|
||||
|
||||
// Conditional-request cache for GET calls: Gitea serves 304 Not Modified for
|
||||
// matching If-None-Match, so polling unchanged issues/PRs stays cheap. Keyed by
|
||||
// token+URL so different identities never share responses.
|
||||
const ETAG_CACHE_MAX_ENTRIES = 300;
|
||||
const etagCache = new Map();
|
||||
|
||||
const rememberEtag = (key, etag, body, headers) => {
|
||||
etagCache.delete(key);
|
||||
etagCache.set(key, { etag, body, headers });
|
||||
if (etagCache.size > ETAG_CACHE_MAX_ENTRIES) {
|
||||
const oldest = etagCache.keys().next().value;
|
||||
if (oldest !== undefined) {
|
||||
etagCache.delete(oldest);
|
||||
}
|
||||
}
|
||||
};
|
||||
|
||||
const createConditionalFetch = (token) => async (url, options = {}) => {
|
||||
const method = (options.method || 'GET').toUpperCase();
|
||||
if (method !== 'GET') {
|
||||
return timeoutFetch(url, options);
|
||||
}
|
||||
|
||||
const cacheKey = `${token}\n${url}`;
|
||||
const cached = etagCache.get(cacheKey);
|
||||
const headers = { ...(options.headers || {}) };
|
||||
if (cached?.etag) {
|
||||
headers['if-none-match'] = cached.etag;
|
||||
}
|
||||
|
||||
const response = await timeoutFetch(url, { ...options, headers });
|
||||
|
||||
if (response.status === 304 && cached) {
|
||||
// Touch for LRU and replay the cached success response.
|
||||
rememberEtag(cacheKey, cached.etag, cached.body, cached.headers);
|
||||
return new Response(cached.body, { status: 200, headers: cached.headers });
|
||||
}
|
||||
|
||||
if (response.ok) {
|
||||
const etag = response.headers.get('etag');
|
||||
if (etag) {
|
||||
const body = await response.arrayBuffer();
|
||||
rememberEtag(cacheKey, etag, body, response.headers);
|
||||
return new Response(body, { status: response.status, headers: response.headers });
|
||||
}
|
||||
}
|
||||
|
||||
return response;
|
||||
};
|
||||
|
||||
// ---- Own rate-limit cooldown (deliberately NOT shared with github/gitlab) ----
|
||||
const MAX_COOLDOWN_MS = 15 * 60 * 1000;
|
||||
const DEFAULT_COOLDOWN_MS = 60 * 1000;
|
||||
let rateLimitedUntil = 0;
|
||||
|
||||
const headerValue = (headers, name) => {
|
||||
if (!headers) return undefined;
|
||||
if (typeof headers.get === 'function') return headers.get(name);
|
||||
return headers[name];
|
||||
};
|
||||
// NOTE: ETag conditional-GET cache and rate-limit cooldown have been dropped
|
||||
// with the pivot to CLI transports. Each call spawns a fresh `tea` process,
|
||||
// so there is no persistent connection to attach conditional headers to, and
|
||||
// rate-limit state cannot be shared across invocations. The tradeoff is higher
|
||||
// latency per call (process spawn overhead) and no 304 short-circuit, but
|
||||
// simpler state management and no module-level mutable cache. Callers that
|
||||
// relied on `isGiteaRateLimited()` will always see false (no cooldown active).
|
||||
|
||||
/**
|
||||
* Record a cooldown after a Gitea 429. Accepts a fetch Response or any object
|
||||
* carrying headers, honoring `Retry-After` (seconds) or `X-RateLimit-Reset`
|
||||
* (Unix seconds) when present.
|
||||
* Spawn a CLI binary and return { stdout, stderr, exitCode }.
|
||||
* Rejects if the process does not finish within REQUEST_TIMEOUT_MS.
|
||||
*/
|
||||
export function noteGiteaRateLimit(error) {
|
||||
const headers = error?.headers;
|
||||
let retryMs = null;
|
||||
const retryAfter = headerValue(headers, 'retry-after');
|
||||
if (retryAfter !== undefined && retryAfter !== null) {
|
||||
const secs = Number(retryAfter);
|
||||
if (Number.isFinite(secs) && secs > 0) retryMs = secs * 1000;
|
||||
function spawnCli(bin, args, env, timeoutMs = REQUEST_TIMEOUT_MS) {
|
||||
return new Promise((resolve, reject) => {
|
||||
const child = spawn(bin, args, {
|
||||
env: { ...process.env, ...env },
|
||||
stdio: ['ignore', 'pipe', 'pipe'],
|
||||
});
|
||||
let stdout = '';
|
||||
let stderr = '';
|
||||
child.stdout.on('data', (chunk) => { stdout += chunk; });
|
||||
child.stderr.on('data', (chunk) => { stderr += chunk; });
|
||||
const timer = setTimeout(() => {
|
||||
child.kill('SIGKILL');
|
||||
reject(new Error(`CLI ${bin} timed out after ${timeoutMs}ms`));
|
||||
}, timeoutMs);
|
||||
child.on('close', (code) => {
|
||||
clearTimeout(timer);
|
||||
resolve({ stdout, stderr, exitCode: code ?? 1 });
|
||||
});
|
||||
child.on('error', (err) => {
|
||||
clearTimeout(timer);
|
||||
reject(err);
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
/**
|
||||
* Run a `tea api` call and parse the response envelope.
|
||||
*
|
||||
* `tea api --include` outputs:
|
||||
* <status line: HTTP/1.1 200 OK>
|
||||
* <headers, one per line>
|
||||
* <empty line>
|
||||
* <JSON body>
|
||||
*
|
||||
* Without `--include`, stdout is just the JSON body on success.
|
||||
*/
|
||||
async function teaApiCall(endpoint, { method = 'GET', body, raw, paginate, teaBin, token }) {
|
||||
const args = ['api', '--include'];
|
||||
if (method !== 'GET') args.push('-X', method);
|
||||
if (paginate) args.push('--paginate');
|
||||
if (raw) args.push('--header', 'Accept: text/plain');
|
||||
if (body !== undefined) args.push('--header', 'Content-Type: application/json', '-d', JSON.stringify(body));
|
||||
args.push(endpoint);
|
||||
|
||||
let result;
|
||||
try {
|
||||
result = await spawnCli(teaBin, args, { GITEA_SERVER_TOKEN: token });
|
||||
} catch (err) {
|
||||
return { status: 500, headers: {}, data: null, page: null, error: err.message };
|
||||
}
|
||||
if (retryMs === null) {
|
||||
// Gitea sends `X-RateLimit-Reset`; check the generic name too for robustness.
|
||||
const reset = headerValue(headers, 'x-ratelimit-reset') ?? headerValue(headers, 'ratelimit-reset');
|
||||
if (reset !== undefined && reset !== null) {
|
||||
const delta = Number(reset) * 1000 - Date.now();
|
||||
if (Number.isFinite(delta) && delta > 0) retryMs = delta;
|
||||
|
||||
const { stdout, stderr, exitCode } = result;
|
||||
|
||||
if (exitCode !== 0 && !stdout.trim()) {
|
||||
return { status: 500, headers: {}, data: null, page: null, error: stderr.trim() || `tea exited with code ${exitCode}` };
|
||||
}
|
||||
|
||||
// Parse --include output: status line, headers, blank line, body.
|
||||
const lines = stdout.split('\n');
|
||||
let status = 200;
|
||||
const headers = {};
|
||||
let bodyStart = 0;
|
||||
|
||||
const statusMatch = lines[0]?.match(/HTTP\/\S+\s+(\d+)/);
|
||||
if (statusMatch) {
|
||||
status = Number(statusMatch[1]);
|
||||
for (let i = 1; i < lines.length; i++) {
|
||||
if (lines[i].trim() === '') {
|
||||
bodyStart = i + 1;
|
||||
break;
|
||||
}
|
||||
const colonIdx = lines[i].indexOf(':');
|
||||
if (colonIdx > 0) {
|
||||
headers[lines[i].slice(0, colonIdx).trim().toLowerCase()] = lines[i].slice(colonIdx + 1).trim();
|
||||
}
|
||||
}
|
||||
}
|
||||
if (retryMs === null) retryMs = DEFAULT_COOLDOWN_MS;
|
||||
retryMs = Math.min(retryMs, MAX_COOLDOWN_MS);
|
||||
const until = Date.now() + retryMs;
|
||||
if (until > rateLimitedUntil) {
|
||||
rateLimitedUntil = until;
|
||||
console.warn(`[gitea] rate limited — pausing Gitea calls for ~${Math.round(retryMs / 1000)}s`);
|
||||
|
||||
const bodyText = lines.slice(bodyStart).join('\n').trim();
|
||||
if (!bodyText) {
|
||||
return { status, headers, data: null, page: null };
|
||||
}
|
||||
|
||||
if (raw) {
|
||||
return { status, headers, data: bodyText, page: null };
|
||||
}
|
||||
|
||||
try {
|
||||
return { status, headers, data: JSON.parse(bodyText), page: null };
|
||||
} catch {
|
||||
return { status, headers, data: bodyText, page: null };
|
||||
}
|
||||
}
|
||||
|
||||
export function isGiteaRateLimited() {
|
||||
return Date.now() < rateLimitedUntil;
|
||||
}
|
||||
// ---- Rate-limit helpers (no-ops with CLI transport) ----
|
||||
export function noteGiteaRateLimit() { /* no-op: CLI processes are stateless */ }
|
||||
export function isGiteaRateLimited() { return false; }
|
||||
|
||||
// ---- Response helpers ----
|
||||
|
||||
const joinApiUrl = (baseUrl, path) => {
|
||||
const base = String(baseUrl || '').replace(/\/+$/, '');
|
||||
// Build the relative API path for tea CLI. tea resolves the base URL from its
|
||||
// own login config, so we pass only the /api/v1/... portion.
|
||||
const apiPath = (path) => {
|
||||
const p = typeof path === 'string' && path ? (path.startsWith('/') ? path : `/${path}`) : '';
|
||||
return `${base}/api/v1${p}`;
|
||||
};
|
||||
|
||||
const headersToObject = (headers) => {
|
||||
const out = {};
|
||||
if (!headers) return out;
|
||||
if (typeof headers.forEach === 'function') {
|
||||
headers.forEach((value, key) => {
|
||||
out[key] = value;
|
||||
});
|
||||
} else if (typeof headers === 'object') {
|
||||
for (const [key, value] of Object.entries(headers)) {
|
||||
out[key] = value;
|
||||
}
|
||||
}
|
||||
return out;
|
||||
};
|
||||
|
||||
const parsePageInfo = (headers) => {
|
||||
const get = (name) => {
|
||||
const value = headerValue(headers, name);
|
||||
return typeof value === 'string' ? value : '';
|
||||
};
|
||||
// Gitea paginates list endpoints via the `Link` header (rel="next") and
|
||||
// reports the total via `X-Total-Count`.
|
||||
const linkHeader = get('link');
|
||||
const relNextMatch = linkHeader.match(/<([^>]+)>\s*;\s*rel="next"/);
|
||||
const totalRaw = get('x-total-count');
|
||||
const total = totalRaw ? Number(totalRaw) : null;
|
||||
const parsed = {
|
||||
page: null,
|
||||
next: null,
|
||||
total: total !== null && Number.isFinite(total) ? total : null,
|
||||
hasMore: Boolean(relNextMatch),
|
||||
};
|
||||
if (relNextMatch) {
|
||||
parsed.nextUrl = relNextMatch[1];
|
||||
}
|
||||
return parsed;
|
||||
};
|
||||
|
||||
const parseData = async (response, raw) => {
|
||||
const text = await response.text();
|
||||
if (raw) {
|
||||
return text;
|
||||
}
|
||||
if (!text) {
|
||||
return null;
|
||||
}
|
||||
try {
|
||||
return JSON.parse(text);
|
||||
} catch {
|
||||
return null;
|
||||
}
|
||||
return `/api/v1${p}`;
|
||||
};
|
||||
|
||||
/**
|
||||
* Create a raw-fetch Gitea/Forgejo REST v1 client. `request` never throws for
|
||||
* HTTP error statuses — it returns `{ status, headers, data, page }` so callers
|
||||
* can branch on status codes. On 429 it also sets `error: 'Gitea rate limited'`
|
||||
* and records a module-level cooldown.
|
||||
* Create a CLI-backed Gitea/Forgejo REST v1 client. Spawns `tea api` for each
|
||||
* request. `request` never throws for HTTP error statuses — it returns
|
||||
* `{ status, headers, data, page }` so callers can branch on status codes.
|
||||
*/
|
||||
export function createGiteaClient({ token, baseUrl }) {
|
||||
const effectiveBaseUrl = typeof baseUrl === 'string' ? baseUrl.trim().replace(/\/+$/, '') : '';
|
||||
const teaBin = process.env.TEA_BIN || '/home/user/.local/bin/tea';
|
||||
|
||||
const request = async (path, options = {}) => {
|
||||
const method = (typeof options.method === 'string' ? options.method : 'GET').toUpperCase();
|
||||
const query = options.query && typeof options.query === 'object' ? options.query : {};
|
||||
const body = options.body;
|
||||
const callerSignal = options.signal;
|
||||
const raw = options.raw === true;
|
||||
|
||||
if (isGiteaRateLimited()) {
|
||||
return { status: 429, headers: {}, data: null, page: null, error: 'Gitea rate limited' };
|
||||
}
|
||||
|
||||
let url = joinApiUrl(effectiveBaseUrl, path);
|
||||
// Build the relative endpoint path with query params baked in.
|
||||
let endpoint = apiPath(path);
|
||||
const qs = new URLSearchParams();
|
||||
let hasQuery = false;
|
||||
for (const [key, value] of Object.entries(query)) {
|
||||
@@ -200,61 +150,11 @@ export function createGiteaClient({ token, baseUrl }) {
|
||||
hasQuery = true;
|
||||
}
|
||||
if (hasQuery) {
|
||||
url += `${url.includes('?') ? '&' : '?'}${qs.toString()}`;
|
||||
endpoint += `${endpoint.includes('?') ? '&' : '?'}${qs.toString()}`;
|
||||
}
|
||||
|
||||
// Gitea/Forgejo PAT auth: `Authorization: token <pat>`.
|
||||
const headers = {
|
||||
Authorization: `token ${token}`,
|
||||
accept: raw ? 'text/plain' : 'application/json',
|
||||
};
|
||||
const fetchOptions = {
|
||||
method,
|
||||
headers,
|
||||
redirect: 'manual',
|
||||
};
|
||||
if (body !== undefined) {
|
||||
headers['content-type'] = 'application/json';
|
||||
fetchOptions.body = JSON.stringify(body);
|
||||
}
|
||||
if (callerSignal) {
|
||||
fetchOptions.signal = callerSignal;
|
||||
}
|
||||
|
||||
const conditionalFetch = createConditionalFetch(token);
|
||||
|
||||
let response = await conditionalFetch(url, fetchOptions);
|
||||
|
||||
// Follow redirects (301/302/308) exactly once. Gitea serves them for moved
|
||||
// repos/users; a manual redirect keeps our Authorization header across the hop.
|
||||
// Only follow same-origin redirects to avoid leaking the token to a different host.
|
||||
let redirects = 0;
|
||||
const baseHost = new URL(url).host;
|
||||
while (
|
||||
(response.status === 301 || response.status === 302 || response.status === 308)
|
||||
&& headerValue(response.headers, 'location')
|
||||
&& redirects < 1
|
||||
) {
|
||||
const location = headerValue(response.headers, 'location');
|
||||
const nextUrl = new URL(location, url).toString();
|
||||
if (new URL(nextUrl).host !== baseHost) break;
|
||||
response = await conditionalFetch(nextUrl, fetchOptions);
|
||||
redirects += 1;
|
||||
}
|
||||
|
||||
const result = {
|
||||
status: response.status,
|
||||
headers: headersToObject(response.headers),
|
||||
data: await parseData(response, raw),
|
||||
page: parsePageInfo(response.headers),
|
||||
};
|
||||
|
||||
if (response.status === 429) {
|
||||
noteGiteaRateLimit(response);
|
||||
result.error = 'Gitea rate limited';
|
||||
}
|
||||
|
||||
return result;
|
||||
const paginate = method === 'GET' && hasQuery;
|
||||
return teaApiCall(endpoint, { method, body, raw, paginate, teaBin, token });
|
||||
};
|
||||
|
||||
return {
|
||||
@@ -302,8 +202,6 @@ export function createGiteaClient({ token, baseUrl }) {
|
||||
request(`/repos/${owner}/${repo}/pulls/${number}/merge`, { method: 'POST', body }),
|
||||
branches: (owner, repo, params = {}) =>
|
||||
request(`/repos/${owner}/${repo}/branches`, { query: params }),
|
||||
// Assignable users (collaborators with role access + org members) are the
|
||||
// mention/assign candidate set; Gitea mirrors the GitHub assignees route.
|
||||
assignees: (owner, repo, params = {}) =>
|
||||
request(`/repos/${owner}/${repo}/assignees`, { query: params }),
|
||||
tags: (owner, repo, params = {}) =>
|
||||
@@ -320,9 +218,6 @@ export function getGiteaClientOrNull(directory) {
|
||||
let baseUrl = auth.baseUrl;
|
||||
if (directory) {
|
||||
const effectiveBaseUrl = getEffectiveProviderApiBaseUrl('gitea', directory);
|
||||
// Only a per-project override replaces the account's base URL; without one
|
||||
// the effective value is just the global default, which stored accounts
|
||||
// (an explicit baseUrl is required) already outrank.
|
||||
if (effectiveBaseUrl !== null && effectiveBaseUrl !== getProviderApiBaseUrl('gitea')) {
|
||||
baseUrl = effectiveBaseUrl;
|
||||
}
|
||||
|
||||
@@ -1,6 +1,7 @@
|
||||
import fs from 'fs';
|
||||
import os from 'os';
|
||||
import path from 'path';
|
||||
import { EventEmitter } from 'events';
|
||||
import { afterAll, afterEach, describe, expect, test, vi } from 'vitest';
|
||||
|
||||
// Isolate auth storage so getGiteaClientOrNull never reads a real account.
|
||||
@@ -11,6 +12,14 @@ afterAll(() => {
|
||||
fs.rmSync(TEMP_DATA_DIR, { recursive: true, force: true });
|
||||
});
|
||||
|
||||
// Mock child_process.spawn to simulate `tea api --include` output.
|
||||
const originalSpawn = (await import('child_process')).spawn;
|
||||
let spawnMock = null;
|
||||
|
||||
vi.mock('child_process', () => ({
|
||||
spawn: (...args) => (spawnMock ? spawnMock(...args) : originalSpawn(...args)),
|
||||
}));
|
||||
|
||||
const {
|
||||
createGiteaClient,
|
||||
getGiteaClientOrNull,
|
||||
@@ -18,74 +27,99 @@ const {
|
||||
noteGiteaRateLimit,
|
||||
} = await import('./client.js');
|
||||
|
||||
const jsonResponse = (data, { status = 200, headers = {} } = {}) =>
|
||||
new Response(JSON.stringify(data), { status, headers: { 'content-type': 'application/json', ...headers } });
|
||||
/**
|
||||
* Build the `tea api --include` output format:
|
||||
* <status line>\n<header: value>\n...\n\n<body>
|
||||
*/
|
||||
const cliOutput = (data, { status = 200, headers = {} } = {}) => {
|
||||
const lines = [`HTTP/1.1 ${status} OK`];
|
||||
for (const [k, v] of Object.entries(headers)) {
|
||||
lines.push(`${k}: ${v}`);
|
||||
}
|
||||
lines.push('');
|
||||
lines.push(typeof data === 'string' ? data : JSON.stringify(data));
|
||||
return lines.join('\n');
|
||||
};
|
||||
|
||||
const originalFetch = globalThis.fetch;
|
||||
/**
|
||||
* Create a vi.fn() mock spawn function that calls `on('close')` with the given
|
||||
* exit code and delivers `output` on stdout.
|
||||
*/
|
||||
const mockSpawn = (output, { exitCode = 0, stderr = '' } = {}) => {
|
||||
return vi.fn((...args) => {
|
||||
const child = new EventEmitter();
|
||||
child.stdout = new EventEmitter();
|
||||
child.stderr = new EventEmitter();
|
||||
child.kill = vi.fn();
|
||||
queueMicrotask(() => {
|
||||
child.stdout.emit('data', output);
|
||||
if (stderr) child.stderr.emit('data', stderr);
|
||||
child.emit('close', exitCode);
|
||||
});
|
||||
return child;
|
||||
});
|
||||
};
|
||||
|
||||
afterEach(() => {
|
||||
globalThis.fetch = originalFetch;
|
||||
spawnMock = null;
|
||||
});
|
||||
|
||||
describe('createGiteaClient request basics', () => {
|
||||
test('calls {baseUrl}/api/v1{path} and sends the token Authorization header', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 42, login: 'alice' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('spawns tea api --include and sends GITEA_SERVER_TOKEN env', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ id: 42, login: 'alice' }));
|
||||
|
||||
const client = createGiteaClient({ token: 'gitea-token', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.user();
|
||||
|
||||
expect(fetchMock).toHaveBeenCalledTimes(1);
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/user');
|
||||
expect(options.headers.Authorization).toBe('token gitea-token');
|
||||
expect(spawnMock).toHaveBeenCalledTimes(1);
|
||||
const [bin, args, opts] = spawnMock.mock.calls[0];
|
||||
expect(bin).toBe('/home/user/.local/bin/tea');
|
||||
expect(args).toEqual(['api', '--include', '/api/v1/user']);
|
||||
expect(opts.env.GITEA_SERVER_TOKEN).toBe('gitea-token');
|
||||
expect(result).toMatchObject({ status: 200, data: { id: 42, login: 'alice' } });
|
||||
expect(result.error).toBeUndefined();
|
||||
});
|
||||
|
||||
test('joins a custom base URL with a path without duplicating /api/v1', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput([]));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com/gitea/' });
|
||||
await client.issues('owner', 'repo', { state: 'open' });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/gitea/api/v1/repos/owner/repo/issues?state=open');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('/api/v1/repos/owner/repo/issues?state=open');
|
||||
});
|
||||
|
||||
test('serializes query params and omits empty ones', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput([]));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.pullRequests('owner', 'repo', { state: 'open', limit: 50, page: 2, q: '', sort: null });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
const query = String(url).split('?')[1];
|
||||
expect(query).toContain('state=open');
|
||||
expect(query).toContain('limit=50');
|
||||
expect(query).toContain('page=2');
|
||||
expect(query).not.toContain('q');
|
||||
expect(query).not.toContain('sort');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
expect(endpoint).toContain('state=open');
|
||||
expect(endpoint).toContain('limit=50');
|
||||
expect(endpoint).toContain('page=2');
|
||||
expect(endpoint).not.toContain('q');
|
||||
expect(endpoint).not.toContain('sort');
|
||||
});
|
||||
|
||||
test('POST requests send a JSON body', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ ok: true }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('POST requests pass the method flag and a JSON body', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ ok: true }, { status: 201 }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.request('/some/action', { method: 'POST', body: { hello: 'world' } });
|
||||
|
||||
const [, options] = fetchMock.mock.calls[0];
|
||||
expect(options.method).toBe('POST');
|
||||
expect(options.headers['content-type']).toBe('application/json');
|
||||
expect(options.body).toBe(JSON.stringify({ hello: 'world' }));
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain('-d');
|
||||
expect(args).toContain(JSON.stringify({ hello: 'world' }));
|
||||
});
|
||||
|
||||
test('surfaces error statuses without throwing', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ message: 'nope' }, { status: 401 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ message: 'nope' }, { status: 401 }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.user();
|
||||
@@ -93,124 +127,61 @@ describe('createGiteaClient request basics', () => {
|
||||
expect(result.data).toEqual({ message: 'nope' });
|
||||
});
|
||||
|
||||
test('attaches a caller signal when provided, else a timeout signal', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const controller = new AbortController();
|
||||
await client.branches('owner', 'repo', { limit: 50 });
|
||||
await client.request('/user', { signal: controller.signal });
|
||||
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(fetchMock.mock.calls[0][1].signal).toEqual(expect.any(AbortSignal));
|
||||
expect(fetchMock.mock.calls[1][1].signal).toBe(controller.signal);
|
||||
});
|
||||
|
||||
test('raw requests return the body as text', async () => {
|
||||
const fetchMock = vi.fn(async () => new Response('diff --git a/src/a.ts b/src/a.ts\n', { status: 200 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('raw requests pass Accept: text/plain header', async () => {
|
||||
spawnMock = mockSpawn(cliOutput('diff --git a/src/a.ts b/src/a.ts\n'));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.pullRequestDiff('owner', 'repo', 5);
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/pulls/5.diff');
|
||||
expect(options.headers.accept).toBe('text/plain');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('--header');
|
||||
expect(args).toContain('Accept: text/plain');
|
||||
expect(args).toContain('/api/v1/repos/owner/repo/pulls/5.diff');
|
||||
expect(result.status).toBe(200);
|
||||
expect(result.data).toBe('diff --git a/src/a.ts b/src/a.ts\n');
|
||||
expect(result.data).toBe('diff --git a/src/a.ts b/src/a.ts');
|
||||
});
|
||||
|
||||
test('returns 500 on CLI process error', async () => {
|
||||
spawnMock = (...args) => {
|
||||
const child = new EventEmitter();
|
||||
child.stdout = new EventEmitter();
|
||||
child.stderr = new EventEmitter();
|
||||
child.kill = vi.fn();
|
||||
queueMicrotask(() => {
|
||||
child.emit('error', new Error('ENOENT'));
|
||||
});
|
||||
return child;
|
||||
};
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.user();
|
||||
expect(result.status).toBe(500);
|
||||
expect(result.error).toContain('ENOENT');
|
||||
});
|
||||
|
||||
test('returns 500 on non-zero exit with no stdout', async () => {
|
||||
spawnMock = mockSpawn('', { exitCode: 1, stderr: 'not logged in' });
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.user();
|
||||
expect(result.status).toBe(500);
|
||||
expect(result.error).toBe('not logged in');
|
||||
});
|
||||
});
|
||||
|
||||
describe('pagination', () => {
|
||||
test('parses the Link rel=next header into the page object', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([], {
|
||||
headers: {
|
||||
link: '<https://gitea.example.com/api/v1/repos/o/r/issues?page=3>; rel="next", <...>; rel="last"',
|
||||
'x-total-count': '57',
|
||||
},
|
||||
}));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('page object is null (CLI handles pagination)', async () => {
|
||||
spawnMock = mockSpawn(cliOutput([], { headers: {} }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.issues('o', 'r', { page: 2 });
|
||||
expect(result.page.hasMore).toBe(true);
|
||||
expect(result.page.nextUrl).toBe('https://gitea.example.com/api/v1/repos/o/r/issues?page=3');
|
||||
expect(result.page.total).toBe(57);
|
||||
});
|
||||
|
||||
test('reports hasMore=false on the last page', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([], { headers: {} }));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.issues('o', 'r', { page: 1 });
|
||||
expect(result.page.hasMore).toBe(false);
|
||||
});
|
||||
});
|
||||
|
||||
describe('redirect handling', () => {
|
||||
test('follows a redirect exactly once, preserving the Authorization header', async () => {
|
||||
const movedUrl = 'https://gitea.example.com/api/v1/repos/newowner/home/issues';
|
||||
const fetchMock = vi.fn(async (url) => {
|
||||
if (String(url).includes('/repos/owner/repo/issues')) {
|
||||
return jsonResponse({}, { status: 301, headers: { location: '/api/v1/repos/newowner/home/issues' } });
|
||||
}
|
||||
if (String(url) === movedUrl) {
|
||||
return jsonResponse([{ number: 1 }]);
|
||||
}
|
||||
return jsonResponse({}, { status: 404 });
|
||||
});
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 'gitea-t', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.issues('owner', 'repo');
|
||||
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(result.status).toBe(200);
|
||||
expect(result.data).toEqual([{ number: 1 }]);
|
||||
const [, secondOptions] = fetchMock.mock.calls[1];
|
||||
expect(secondOptions.headers.Authorization).toBe('token gitea-t');
|
||||
});
|
||||
});
|
||||
|
||||
describe('etag conditional cache', () => {
|
||||
test('sends if-none-match and replays a 304 as a 200 with cached body', async () => {
|
||||
const fetchMock = vi.fn(async (_url, options) => {
|
||||
if (options.headers['if-none-match'] === '"v1"') {
|
||||
return new Response(null, { status: 304 });
|
||||
}
|
||||
return jsonResponse({ ok: true }, { headers: { etag: '"v1"' } });
|
||||
});
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 'gitea-t', baseUrl: 'https://gitea.example.com' });
|
||||
const first = await client.user();
|
||||
expect(first.status).toBe(200);
|
||||
expect(first.data).toEqual({ ok: true });
|
||||
|
||||
const second = await client.user();
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(fetchMock.mock.calls[1][1].headers['if-none-match']).toBe('"v1"');
|
||||
expect(second.status).toBe(200);
|
||||
expect(second.data).toEqual({ ok: true });
|
||||
});
|
||||
|
||||
test('does not cache POST responses', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ ok: true }));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 'gitea-t', baseUrl: 'https://gitea.example.com' });
|
||||
await client.request('/thing', { method: 'POST', body: {} });
|
||||
await client.request('/thing', { method: 'POST', body: {} });
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(result.page).toBeNull();
|
||||
});
|
||||
});
|
||||
|
||||
describe('pull request write methods', () => {
|
||||
test('createPullRequest POSTs title/head/base to the pulls endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ number: 5, title: 'New PR' }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('createPullRequest POSTs to the pulls endpoint', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ number: 5, title: 'New PR' }, { status: 201 }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.createPullRequest('owner', 'repo', {
|
||||
@@ -219,45 +190,42 @@ describe('pull request write methods', () => {
|
||||
base: 'main',
|
||||
});
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/pulls');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(options.headers['content-type']).toBe('application/json');
|
||||
expect(JSON.parse(options.body)).toEqual({ title: 'New PR', head: 'feat/x', base: 'main' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ title: 'New PR', head: 'feat/x', base: 'main' }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/repos/owner/repo/pulls'))).toBe(true);
|
||||
expect(result.status).toBe(201);
|
||||
expect(result.data).toEqual({ number: 5, title: 'New PR' });
|
||||
});
|
||||
|
||||
test('updatePullRequest PATCHes a JSON body to the pull request endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ number: 5, title: 'Updated' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('updatePullRequest PATCHes to the pull request endpoint', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ number: 5, title: 'Updated' }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.updatePullRequest('owner', 'repo', 5, { title: 'Updated', body: 'Body text' });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/pulls/5');
|
||||
expect(options.method).toBe('PATCH');
|
||||
expect(options.headers['content-type']).toBe('application/json');
|
||||
expect(JSON.parse(options.body)).toEqual({ title: 'Updated', body: 'Body text' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('PATCH');
|
||||
expect(args).toContain(JSON.stringify({ title: 'Updated', body: 'Body text' }));
|
||||
});
|
||||
|
||||
test('mergePullRequest POSTs the merge style in Do to the merge endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ merged: true }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('mergePullRequest POSTs the merge style to the merge endpoint', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ merged: true }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.mergePullRequest('owner', 'repo', 5, { Do: 'squash' });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/pulls/5/merge');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(JSON.parse(options.body)).toEqual({ Do: 'squash' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ Do: 'squash' }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/pulls/5/merge'))).toBe(true);
|
||||
});
|
||||
|
||||
test('write methods surface error statuses without throwing', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ message: 'Conflict' }, { status: 409 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ message: 'Conflict' }, { status: 409 }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.mergePullRequest('owner', 'repo', 5, { Do: 'merge' });
|
||||
@@ -268,102 +236,77 @@ describe('pull request write methods', () => {
|
||||
|
||||
describe('issue, review, and repo write methods', () => {
|
||||
test('createIssueComment POSTs a body to the issue comments endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 5, body: 'hi' }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ id: 5, body: 'hi' }, { status: 201 }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.createIssueComment('owner', 'repo', 7, 'Nice catch');
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/issues/7/comments');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(JSON.parse(options.body)).toEqual({ body: 'Nice catch' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ body: 'Nice catch' }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/issues/7/comments'))).toBe(true);
|
||||
expect(result.status).toBe(201);
|
||||
});
|
||||
|
||||
test('updateIssue PATCHes params to the issue endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ number: 7, title: 'Updated' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ number: 7, title: 'Updated' }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.updateIssue('owner', 'repo', 7, { state: 'closed', labels: ['bug'], milestone: 33 });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/issues/7');
|
||||
expect(options.method).toBe('PATCH');
|
||||
expect(JSON.parse(options.body)).toEqual({ state: 'closed', labels: ['bug'], milestone: 33 });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('PATCH');
|
||||
expect(args).toContain(JSON.stringify({ state: 'closed', labels: ['bug'], milestone: 33 }));
|
||||
});
|
||||
|
||||
test('createPullReview POSTs event/body to the reviews endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 101, state: 'APPROVED' }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ id: 101, state: 'APPROVED' }, { status: 201 }));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.createPullReview('owner', 'repo', 12, { event: 'APPROVED', body: 'LGTM' });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/pulls/12/reviews');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(JSON.parse(options.body)).toEqual({ event: 'APPROVED', body: 'LGTM' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ event: 'APPROVED', body: 'LGTM' }));
|
||||
});
|
||||
|
||||
test('milestones GETs the repo milestones list', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([{ id: 33, title: 'v1.0' }]));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('milestones passes state and limit query params', async () => {
|
||||
spawnMock = mockSpawn(cliOutput([{ id: 33, title: 'v1.0' }]));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.milestones('owner', 'repo', { state: 'all', limit: 50 });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/milestones?state=all&limit=50');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
expect(endpoint).toContain('state=all');
|
||||
expect(endpoint).toContain('limit=50');
|
||||
});
|
||||
|
||||
test('repoLabels GETs the repo labels list', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([{ id: 1, name: 'bug', color: 'd73a4a' }]));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('repoLabels passes limit query param', async () => {
|
||||
spawnMock = mockSpawn(cliOutput([{ id: 1, name: 'bug', color: 'd73a4a' }]));
|
||||
|
||||
const client = createGiteaClient({ token: 't', baseUrl: 'https://gitea.example.com' });
|
||||
await client.repoLabels('owner', 'repo', { limit: 100 });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitea.example.com/api/v1/repos/owner/repo/labels?limit=100');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
expect(endpoint).toContain('limit=100');
|
||||
});
|
||||
});
|
||||
|
||||
describe('rate limiting', () => {
|
||||
// NOTE: these tests run last in this file. The rate-limit cooldown is
|
||||
// module-level and has no reset export, so earlier tests must not set one.
|
||||
test('429 surfaces error and records a cooldown', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({}, { status: 429, headers: { 'retry-after': '5' } }));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 'gitea-t', baseUrl: 'https://gitea.example.com' });
|
||||
const result = await client.user();
|
||||
expect(result.status).toBe(429);
|
||||
expect(result.error).toBe('Gitea rate limited');
|
||||
expect(isGiteaRateLimited()).toBe(true);
|
||||
test('isGiteaRateLimited returns false (no-op with CLI transport)', () => {
|
||||
expect(isGiteaRateLimited()).toBe(false);
|
||||
});
|
||||
|
||||
test('short-circuits while the cooldown is active without calling fetch', async () => {
|
||||
test('noteGiteaRateLimit is a no-op', () => {
|
||||
noteGiteaRateLimit({ headers: new Headers({ 'retry-after': '120' }) });
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGiteaClient({ token: 'gitea-t', baseUrl: 'https://gitea.example.com' });
|
||||
const gated = await client.issues('o', 'r');
|
||||
expect(gated.status).toBe(429);
|
||||
expect(gated.error).toBe('Gitea rate limited');
|
||||
expect(fetchMock).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
test('parses Retry-After seconds into the cooldown', () => {
|
||||
noteGiteaRateLimit({ headers: new Headers({ 'retry-after': '120' }) });
|
||||
expect(isGiteaRateLimited()).toBe(true);
|
||||
});
|
||||
|
||||
test('honors X-RateLimit-Reset for the cooldown', () => {
|
||||
noteGiteaRateLimit({ headers: new Headers({ 'x-ratelimit-reset': String(Math.floor(Date.now() / 1000) + 60) }) });
|
||||
expect(isGiteaRateLimited()).toBe(true);
|
||||
expect(isGiteaRateLimited()).toBe(false);
|
||||
});
|
||||
|
||||
test('getGiteaClientOrNull returns null without stored auth', () => {
|
||||
|
||||
@@ -2,16 +2,16 @@
|
||||
|
||||
## Purpose
|
||||
|
||||
- This module owns GitLab auth (Personal Access Token), raw REST v4 client access, remote-URL repo resolution, and GitLab issue / merge-request (MR) APIs for OpenChamber, including MR create/update/merge writes.
|
||||
- This module owns GitLab auth (Personal Access Token), CLI-backed REST v4 client access, remote-URL repo resolution, and GitLab issue / merge-request (MR) APIs for OpenChamber, including MR create/update/merge writes.
|
||||
- From a user perspective, this is the layer that lets the app show GitLab issues and merge requests for a local project, including comments and per-file diffs, and create, edit, and merge merge requests.
|
||||
- The module mirrors `packages/web/server/lib/github/` but uses a **Personal Access Token (PAT)** with a configurable base URL (gitlab.com by default, or a self-hosted instance), and talks to GitLab's REST v4 API directly via `fetch` — no new dependencies.
|
||||
- The module mirrors `packages/web/server/lib/github/` but uses a **Personal Access Token (PAT)** with a configurable base URL (gitlab.com by default, or a self-hosted instance), and talks to GitLab's REST v4 API via the **`glab` CLI** transport instead of raw `fetch`.
|
||||
|
||||
## Entrypoints and structure
|
||||
|
||||
- `packages/web/server/lib/gitlab/index.js`: public server entrypoint re-exports.
|
||||
- `packages/web/server/lib/gitlab/routes.js`: Express route registration for `/api/gitlab/*` endpoints.
|
||||
- `packages/web/server/lib/gitlab/auth.js`: PAT auth storage, multi-account support, base URL normalization.
|
||||
- `packages/web/server/lib/gitlab/client.js`: raw `fetch` GitLab REST v4 client (timeout, ETag conditional GET, rate-limit cooldown, pagination, redirect handling).
|
||||
- `packages/web/server/lib/gitlab/client.js`: CLI-backed `glab api` client (process spawn with 8s timeout, `--include` for HTTP status/headers, `--paginate` for list endpoints). Token is passed via `GITLAB_TOKEN` env var.
|
||||
- `packages/web/server/lib/gitlab/repo.js`: GitLab remote URL parsing and directory-to-repo resolution.
|
||||
- `packages/web/server/lib/opencode/feature-routes-runtime.js`: API route layer that calls this module (via `registerGitLabRoutes`).
|
||||
- `packages/web/src/api/gitlab.ts`: web client wrapper for GitLab endpoints.
|
||||
@@ -62,13 +62,12 @@ Nothing in the client or repo layers assumes the token came from a PAT.
|
||||
|
||||
## Client behavior
|
||||
|
||||
- Base URL joining: `{baseUrl}/api/v4{path}`. Project `:id` segments are URL-encoded with `encodeURIComponent` (e.g. `group/sub` -> `group%2Fsub`) and never double-encoded.
|
||||
- Per-request timeout: 8000 ms via `AbortSignal.timeout`, unless the caller passes its own signal.
|
||||
- ETag conditional-GET cache: keyed `token\nurl`, max 300 LRU entries; a `304` is replayed from cache as a `200`. GET only.
|
||||
- Pagination: `x-page`, `x-next-page`, `x-total-pages`, and the `Link` header (`rel="next"`) are parsed into the returned `page` object (`hasMore` = a next page exists).
|
||||
- Redirects: `301`/`302`/`308` with a `Location` header are followed exactly once (project moves) with `redirect: 'manual'`, preserving `PRIVATE-TOKEN` across the hop.
|
||||
- Rate limits: a `429` records a module-level cooldown (honoring `Retry-After` / `RateLimit-Reset` when present) and surfaces `{ status: 429, error: 'GitLab rate limited' }`. While the cooldown is active, requests short-circuit without hitting the network.
|
||||
- Transport: each call spawns a `glab api` process with `--include` for HTTP status/headers. Auth is passed via `GITLAB_TOKEN` env var (never on argv). Binary path: `GLAB_BIN` env or `/home/user/.local/bin/glab`.
|
||||
- Base URL: the `baseUrl` parameter is passed to the client constructor for compatibility but `glab` resolves the instance from its own config. The `--include` flag provides HTTP status codes and response headers.
|
||||
- Per-request timeout: 8000 ms via `AbortSignal.timeout` on the spawned process. The process is killed with `SIGKILL` on timeout.
|
||||
- Pagination: `--paginate` is passed for GET requests with query params, causing `glab` to fetch all pages in a single call. The returned `page` object is `null` since pagination is handled by the CLI.
|
||||
- `request` never throws for HTTP error statuses — callers branch on `status`.
|
||||
- ETag cache and rate-limit cooldown have been dropped with the CLI pivot. Each call spawns a fresh process, so there is no persistent connection for conditional requests or shared rate-limit state. `isGitLabRateLimited()` always returns `false`; `noteGitLabRateLimit()` is a no-op.
|
||||
|
||||
## API integration overview
|
||||
|
||||
@@ -148,5 +147,5 @@ Conventions mirror `github/routes.js`:
|
||||
- Keep the response shapes in lockstep with `GitLab*` types in `packages/ui/src/lib/api/types.ts`.
|
||||
- Never log tokens. Error messages must not include the access token.
|
||||
- Do not double-encode project paths; convenience methods already call `encodeURIComponent` on the `pathWithNamespace`.
|
||||
- The ETag cache and rate-limit cooldown are module-level and per-instance — they are NOT shared with the GitHub module.
|
||||
- The `glab` CLI handles authentication, base URL resolution, and pagination internally. The client does not maintain its own ETag cache or rate-limit cooldown — each call spawns a fresh process.
|
||||
- To add further GitLab write operations, add the endpoint in `routes.js`, add a convenience method in `client.js`, and extend the shared types — mirror the existing issue/MR write routes and the GitHub PR write routes.
|
||||
|
||||
@@ -1,193 +1,147 @@
|
||||
import { spawn } from 'child_process';
|
||||
import { getGitLabAuth, getGitLabDefaultBaseUrl } from './auth.js';
|
||||
|
||||
// Per-request timeout for every GitLab call. GitLab REST can hang under load
|
||||
// (especially self-hosted instances); bounding each request lets the caller
|
||||
// fail fast and serve cached/last-known state instead of holding a socket open.
|
||||
const GLAB_BIN = process.env.GLAB_BIN || '/home/user/.local/bin/glab';
|
||||
const REQUEST_TIMEOUT_MS = 8000;
|
||||
|
||||
const timeoutFetch = (url, options = {}) => {
|
||||
// Respect a caller-provided signal if present; otherwise attach our timeout.
|
||||
if (options.signal) {
|
||||
return fetch(url, options);
|
||||
}
|
||||
return fetch(url, { ...options, signal: AbortSignal.timeout(REQUEST_TIMEOUT_MS) });
|
||||
};
|
||||
|
||||
// Conditional-request cache for GET calls: GitLab serves 304 Not Modified for
|
||||
// matching If-None-Match without consuming a fresh rate-limit token, so
|
||||
// polling unchanged issues/MRs stays cheap. Keyed by token+URL so different
|
||||
// identities never share responses. GitLab (unlike GitHub) does not attach
|
||||
// `ETag` to every endpoint, but when it does we revalidate exactly like
|
||||
// github/octokit.js.
|
||||
const ETAG_CACHE_MAX_ENTRIES = 300;
|
||||
const etagCache = new Map();
|
||||
|
||||
const rememberEtag = (key, etag, body, headers) => {
|
||||
etagCache.delete(key);
|
||||
etagCache.set(key, { etag, body, headers });
|
||||
if (etagCache.size > ETAG_CACHE_MAX_ENTRIES) {
|
||||
const oldest = etagCache.keys().next().value;
|
||||
if (oldest !== undefined) {
|
||||
etagCache.delete(oldest);
|
||||
}
|
||||
}
|
||||
};
|
||||
|
||||
const createConditionalFetch = (token) => async (url, options = {}) => {
|
||||
const method = (options.method || 'GET').toUpperCase();
|
||||
if (method !== 'GET') {
|
||||
return timeoutFetch(url, options);
|
||||
}
|
||||
|
||||
const cacheKey = `${token}\n${url}`;
|
||||
const cached = etagCache.get(cacheKey);
|
||||
const headers = { ...(options.headers || {}) };
|
||||
if (cached?.etag) {
|
||||
headers['if-none-match'] = cached.etag;
|
||||
}
|
||||
|
||||
const response = await timeoutFetch(url, { ...options, headers });
|
||||
|
||||
if (response.status === 304 && cached) {
|
||||
// Touch for LRU and replay the cached success response.
|
||||
rememberEtag(cacheKey, cached.etag, cached.body, cached.headers);
|
||||
return new Response(cached.body, { status: 200, headers: cached.headers });
|
||||
}
|
||||
|
||||
if (response.ok) {
|
||||
const etag = response.headers.get('etag');
|
||||
if (etag) {
|
||||
const body = await response.arrayBuffer();
|
||||
rememberEtag(cacheKey, etag, body, response.headers);
|
||||
return new Response(body, { status: response.status, headers: response.headers });
|
||||
}
|
||||
}
|
||||
|
||||
return response;
|
||||
};
|
||||
|
||||
// ---- Own rate-limit cooldown (deliberately NOT shared with github/rate-limit.js) ----
|
||||
const MAX_COOLDOWN_MS = 15 * 60 * 1000;
|
||||
const DEFAULT_COOLDOWN_MS = 60 * 1000;
|
||||
let rateLimitedUntil = 0;
|
||||
|
||||
const headerValue = (headers, name) => {
|
||||
if (!headers) return undefined;
|
||||
if (typeof headers.get === 'function') return headers.get(name);
|
||||
return headers[name];
|
||||
};
|
||||
// NOTE: ETag conditional-GET cache and rate-limit cooldown have been dropped
|
||||
// with the pivot to CLI transports. Each call spawns a fresh `glab` process,
|
||||
// so there is no persistent connection to attach conditional headers to, and
|
||||
// rate-limit state cannot be shared across invocations. The tradeoff is higher
|
||||
// latency per call (process spawn overhead) and no 304 short-circuit, but
|
||||
// simpler state management and no module-level mutable cache. Callers that
|
||||
// relied on `isGitLabRateLimited()` will always see false (no cooldown active).
|
||||
|
||||
/**
|
||||
* Record a cooldown after a GitLab 429. Accepts a fetch Response or any object
|
||||
* carrying headers (response, `retry-after` seconds, or `RateLimit-Reset`
|
||||
* Unix seconds).
|
||||
* Spawn a CLI binary and return { stdout, stderr, exitCode }.
|
||||
* Rejects if the process does not finish within REQUEST_TIMEOUT_MS.
|
||||
*/
|
||||
export function noteGitLabRateLimit(error) {
|
||||
const headers = error?.headers;
|
||||
let retryMs = null;
|
||||
const retryAfter = headerValue(headers, 'retry-after');
|
||||
if (retryAfter !== undefined && retryAfter !== null) {
|
||||
const secs = Number(retryAfter);
|
||||
if (Number.isFinite(secs) && secs > 0) retryMs = secs * 1000;
|
||||
function spawnCli(bin, args, env, timeoutMs = REQUEST_TIMEOUT_MS) {
|
||||
return new Promise((resolve, reject) => {
|
||||
const child = spawn(bin, args, {
|
||||
env: { ...process.env, ...env },
|
||||
stdio: ['ignore', 'pipe', 'pipe'],
|
||||
});
|
||||
let stdout = '';
|
||||
let stderr = '';
|
||||
child.stdout.on('data', (chunk) => { stdout += chunk; });
|
||||
child.stderr.on('data', (chunk) => { stderr += chunk; });
|
||||
const timer = setTimeout(() => {
|
||||
child.kill('SIGKILL');
|
||||
reject(new Error(`CLI ${bin} timed out after ${timeoutMs}ms`));
|
||||
}, timeoutMs);
|
||||
child.on('close', (code) => {
|
||||
clearTimeout(timer);
|
||||
resolve({ stdout, stderr, exitCode: code ?? 1 });
|
||||
});
|
||||
child.on('error', (err) => {
|
||||
clearTimeout(timer);
|
||||
reject(err);
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
/**
|
||||
* Run a `glab api` call and parse the response envelope.
|
||||
*
|
||||
* `glab api --include` outputs:
|
||||
* <status line: HTTP/1.1 200 OK>
|
||||
* <headers, one per line>
|
||||
* <empty line>
|
||||
* <JSON body>
|
||||
*
|
||||
* Without `--include`, stdout is just the JSON body on success.
|
||||
*/
|
||||
async function glabApiCall(endpoint, { method = 'GET', body, raw, paginate, glabBin, token }) {
|
||||
const args = ['api', '--include'];
|
||||
if (method !== 'GET') args.push('-X', method);
|
||||
if (paginate) args.push('--paginate');
|
||||
if (raw) args.push('--header', 'Accept: text/plain');
|
||||
if (body !== undefined) args.push('--header', 'Content-Type: application/json', '-d', JSON.stringify(body));
|
||||
args.push(endpoint);
|
||||
|
||||
let result;
|
||||
try {
|
||||
result = await spawnCli(glabBin, args, { GITLAB_TOKEN: token });
|
||||
} catch (err) {
|
||||
return { status: 500, headers: {}, data: null, page: null, error: err.message };
|
||||
}
|
||||
if (retryMs === null) {
|
||||
const reset = headerValue(headers, 'ratelimit-reset');
|
||||
if (reset !== undefined && reset !== null) {
|
||||
const delta = Number(reset) * 1000 - Date.now();
|
||||
if (Number.isFinite(delta) && delta > 0) retryMs = delta;
|
||||
|
||||
const { stdout, stderr, exitCode } = result;
|
||||
|
||||
if (exitCode !== 0 && !stdout.trim()) {
|
||||
return { status: 500, headers: {}, data: null, page: null, error: stderr.trim() || `glab exited with code ${exitCode}` };
|
||||
}
|
||||
|
||||
// Parse --include output: status line, headers, blank line, body.
|
||||
const lines = stdout.split('\n');
|
||||
let status = 200;
|
||||
const headers = {};
|
||||
let bodyStart = 0;
|
||||
|
||||
const statusMatch = lines[0]?.match(/HTTP\/\S+\s+(\d+)/);
|
||||
if (statusMatch) {
|
||||
status = Number(statusMatch[1]);
|
||||
for (let i = 1; i < lines.length; i++) {
|
||||
if (lines[i].trim() === '') {
|
||||
bodyStart = i + 1;
|
||||
break;
|
||||
}
|
||||
const colonIdx = lines[i].indexOf(':');
|
||||
if (colonIdx > 0) {
|
||||
headers[lines[i].slice(0, colonIdx).trim().toLowerCase()] = lines[i].slice(colonIdx + 1).trim();
|
||||
}
|
||||
}
|
||||
}
|
||||
if (retryMs === null) retryMs = DEFAULT_COOLDOWN_MS;
|
||||
retryMs = Math.min(retryMs, MAX_COOLDOWN_MS);
|
||||
const until = Date.now() + retryMs;
|
||||
if (until > rateLimitedUntil) {
|
||||
rateLimitedUntil = until;
|
||||
console.warn(`[gitlab] rate limited — pausing GitLab calls for ~${Math.round(retryMs / 1000)}s`);
|
||||
|
||||
const bodyText = lines.slice(bodyStart).join('\n').trim();
|
||||
if (!bodyText) {
|
||||
return { status, headers, data: null, page: null };
|
||||
}
|
||||
|
||||
if (raw) {
|
||||
return { status, headers, data: bodyText, page: null };
|
||||
}
|
||||
|
||||
try {
|
||||
return { status, headers, data: JSON.parse(bodyText), page: null };
|
||||
} catch {
|
||||
return { status, headers, data: bodyText, page: null };
|
||||
}
|
||||
}
|
||||
|
||||
export function isGitLabRateLimited() {
|
||||
return Date.now() < rateLimitedUntil;
|
||||
}
|
||||
// ---- Rate-limit helpers (no-ops with CLI transport) ----
|
||||
export function noteGitLabRateLimit() { /* no-op: CLI processes are stateless */ }
|
||||
export function isGitLabRateLimited() { return false; }
|
||||
|
||||
// ---- Response helpers ----
|
||||
|
||||
const joinApiUrl = (baseUrl, path) => {
|
||||
const base = String(baseUrl || getGitLabDefaultBaseUrl()).replace(/\/+$/, '');
|
||||
const p = typeof path === 'string' && path ? (path.startsWith('/') ? path : `/${path}`) : '';
|
||||
return `${base}/api/v4${p}`;
|
||||
};
|
||||
|
||||
const headersToObject = (headers) => {
|
||||
const out = {};
|
||||
if (!headers) return out;
|
||||
if (typeof headers.forEach === 'function') {
|
||||
headers.forEach((value, key) => {
|
||||
out[key] = value;
|
||||
});
|
||||
} else if (typeof headers === 'object') {
|
||||
for (const [key, value] of Object.entries(headers)) {
|
||||
out[key] = value;
|
||||
}
|
||||
}
|
||||
return out;
|
||||
};
|
||||
|
||||
const parsePageInfo = (headers) => {
|
||||
const get = (name) => {
|
||||
const value = headerValue(headers, name);
|
||||
return typeof value === 'string' ? value : '';
|
||||
};
|
||||
const pageHeader = get('x-page');
|
||||
const nextPage = get('x-next-page');
|
||||
const totalPages = get('x-total-pages');
|
||||
const linkHeader = get('link');
|
||||
const relNextMatch = linkHeader.match(/<([^>]+)>\s*;\s*rel="next"/);
|
||||
const page = pageHeader ? Number(pageHeader) : null;
|
||||
const next = nextPage ? Number(nextPage) : null;
|
||||
const total = totalPages ? Number(totalPages) : null;
|
||||
const hasMore = next != null ? next > 0 : Boolean(relNextMatch);
|
||||
const parsed = { page, next, total, hasMore };
|
||||
if (relNextMatch) {
|
||||
parsed.nextUrl = relNextMatch[1];
|
||||
}
|
||||
return parsed;
|
||||
};
|
||||
|
||||
const parseData = async (response) => {
|
||||
const text = await response.text();
|
||||
if (!text) {
|
||||
return null;
|
||||
}
|
||||
try {
|
||||
return JSON.parse(text);
|
||||
} catch {
|
||||
return null;
|
||||
}
|
||||
};
|
||||
|
||||
const encodeProject = (pathWithNamespace) => encodeURIComponent(String(pathWithNamespace));
|
||||
|
||||
// Build the relative API path for glab CLI. glab resolves the base URL from its
|
||||
// own config, so we pass only the /api/v4/... portion.
|
||||
const apiPath = (path) => {
|
||||
const p = typeof path === 'string' && path ? (path.startsWith('/') ? path : `/${path}`) : '';
|
||||
return `/api/v4${p}`;
|
||||
};
|
||||
|
||||
/**
|
||||
* Create a raw-fetch GitLab REST v4 client. `request` never throws for HTTP
|
||||
* error statuses — it returns `{ status, headers, data, page }` so callers can
|
||||
* branch on status codes. On 429 it also sets `error: 'GitLab rate limited'`
|
||||
* and records a module-level cooldown.
|
||||
* Create a CLI-backed GitLab REST v4 client. Spawns `glab api` for each
|
||||
* request. `request` never throws for HTTP error statuses — it returns
|
||||
* `{ status, headers, data, page }` so callers can branch on status codes.
|
||||
*/
|
||||
export function createGitLabClient({ token, baseUrl }) {
|
||||
const effectiveBaseUrl = normalizeBaseForClient(baseUrl);
|
||||
const glabBin = process.env.GLAB_BIN || '/home/user/.local/bin/glab';
|
||||
|
||||
const request = async (path, options = {}) => {
|
||||
const method = (typeof options.method === 'string' ? options.method : 'GET').toUpperCase();
|
||||
const query = options.query && typeof options.query === 'object' ? options.query : {};
|
||||
const body = options.body;
|
||||
const callerSignal = options.signal;
|
||||
|
||||
if (isGitLabRateLimited()) {
|
||||
return { status: 429, headers: {}, data: null, page: null, error: 'GitLab rate limited' };
|
||||
}
|
||||
// Build the relative endpoint path with query params baked in.
|
||||
let endpoint = apiPath(path);
|
||||
|
||||
let url = joinApiUrl(effectiveBaseUrl, path);
|
||||
const qs = new URLSearchParams();
|
||||
let hasQuery = false;
|
||||
for (const [key, value] of Object.entries(query)) {
|
||||
@@ -196,61 +150,11 @@ export function createGitLabClient({ token, baseUrl }) {
|
||||
hasQuery = true;
|
||||
}
|
||||
if (hasQuery) {
|
||||
url += `${url.includes('?') ? '&' : '?'}${qs.toString()}`;
|
||||
endpoint += `${endpoint.includes('?') ? '&' : '?'}${qs.toString()}`;
|
||||
}
|
||||
|
||||
const headers = {
|
||||
'PRIVATE-TOKEN': token,
|
||||
accept: 'application/json',
|
||||
};
|
||||
const fetchOptions = {
|
||||
method,
|
||||
headers,
|
||||
redirect: 'manual',
|
||||
};
|
||||
if (body !== undefined) {
|
||||
headers['content-type'] = 'application/json';
|
||||
fetchOptions.body = JSON.stringify(body);
|
||||
}
|
||||
if (callerSignal) {
|
||||
fetchOptions.signal = callerSignal;
|
||||
}
|
||||
|
||||
const conditionalFetch = createConditionalFetch(token);
|
||||
|
||||
let response = await conditionalFetch(url, fetchOptions);
|
||||
|
||||
// Follow a project-move redirect exactly once. GitLab redirects
|
||||
// (301/302/308) come with a `Location` for the new project URL; a manual
|
||||
// redirect keeps our PRIVATE-TOKEN header across the hop. Only follow
|
||||
// same-origin redirects to avoid leaking the token to a different host.
|
||||
let redirects = 0;
|
||||
const baseHost = new URL(url).host;
|
||||
while (
|
||||
(response.status === 301 || response.status === 302 || response.status === 308)
|
||||
&& headerValue(response.headers, 'location')
|
||||
&& redirects < 1
|
||||
) {
|
||||
const location = headerValue(response.headers, 'location');
|
||||
const nextUrl = new URL(location, url).toString();
|
||||
if (new URL(nextUrl).host !== baseHost) break;
|
||||
response = await conditionalFetch(nextUrl, fetchOptions);
|
||||
redirects += 1;
|
||||
}
|
||||
|
||||
const result = {
|
||||
status: response.status,
|
||||
headers: headersToObject(response.headers),
|
||||
data: await parseData(response),
|
||||
page: parsePageInfo(response.headers),
|
||||
};
|
||||
|
||||
if (response.status === 429) {
|
||||
noteGitLabRateLimit(response);
|
||||
result.error = 'GitLab rate limited';
|
||||
}
|
||||
|
||||
return result;
|
||||
const paginate = method === 'GET' && hasQuery;
|
||||
return glabApiCall(endpoint, { method, body, paginate, glabBin, token });
|
||||
};
|
||||
|
||||
return {
|
||||
@@ -294,9 +198,6 @@ export function createGitLabClient({ token, baseUrl }) {
|
||||
request(`/projects/${encodeProject(pathWithNamespace)}/merge_requests/${iid}/merge`, { method: 'PUT', body }),
|
||||
branches: (pathWithNamespace, params = {}) =>
|
||||
request(`/projects/${encodeProject(pathWithNamespace)}/repository/branches`, { query: params }),
|
||||
// Project members (direct + inherited) are the assignable/mentionable user
|
||||
// set. `members/all` includes inherited group members; `query` filters
|
||||
// server-side by username/name/email.
|
||||
members: (pathWithNamespace, params = {}) =>
|
||||
request(`/projects/${encodeProject(pathWithNamespace)}/members/all`, { query: params }),
|
||||
labels: (pathWithNamespace, params = {}) =>
|
||||
|
||||
@@ -1,6 +1,7 @@
|
||||
import fs from 'fs';
|
||||
import os from 'os';
|
||||
import path from 'path';
|
||||
import { EventEmitter } from 'events';
|
||||
import { afterAll, afterEach, describe, expect, test, vi } from 'vitest';
|
||||
|
||||
// Isolate auth storage so getGitLabClientOrNull never reads a real account.
|
||||
@@ -11,6 +12,14 @@ afterAll(() => {
|
||||
fs.rmSync(TEMP_DATA_DIR, { recursive: true, force: true });
|
||||
});
|
||||
|
||||
// Mock child_process.spawn to simulate `glab api --include` output.
|
||||
const originalSpawn = (await import('child_process')).spawn;
|
||||
let spawnMock = null;
|
||||
|
||||
vi.mock('child_process', () => ({
|
||||
spawn: (...args) => (spawnMock ? spawnMock(...args) : originalSpawn(...args)),
|
||||
}));
|
||||
|
||||
const {
|
||||
createGitLabClient,
|
||||
getGitLabClientOrNull,
|
||||
@@ -18,86 +27,113 @@ const {
|
||||
noteGitLabRateLimit,
|
||||
} = await import('./client.js');
|
||||
|
||||
const jsonResponse = (data, { status = 200, headers = {} } = {}) =>
|
||||
new Response(JSON.stringify(data), { status, headers: { 'content-type': 'application/json', ...headers } });
|
||||
/**
|
||||
* Build the `glab api --include` output format:
|
||||
* <status line>\n<header: value>\n...\n\n<body>
|
||||
*/
|
||||
const cliOutput = (data, { status = 200, headers = {} } = {}) => {
|
||||
const lines = [`HTTP/1.1 ${status} OK`];
|
||||
for (const [k, v] of Object.entries(headers)) {
|
||||
lines.push(`${k}: ${v}`);
|
||||
}
|
||||
lines.push('');
|
||||
lines.push(typeof data === 'string' ? data : JSON.stringify(data));
|
||||
return lines.join('\n');
|
||||
};
|
||||
|
||||
const originalFetch = globalThis.fetch;
|
||||
/**
|
||||
* Create a vi.fn() mock spawn function that calls `on('close')` with the given
|
||||
* exit code and delivers `output` on stdout.
|
||||
*/
|
||||
const mockSpawn = (output, { exitCode = 0, stderr = '' } = {}) => {
|
||||
return vi.fn((...args) => {
|
||||
const child = new EventEmitter();
|
||||
child.stdout = new EventEmitter();
|
||||
child.stderr = new EventEmitter();
|
||||
child.kill = vi.fn();
|
||||
queueMicrotask(() => {
|
||||
child.stdout.emit('data', output);
|
||||
if (stderr) child.stderr.emit('data', stderr);
|
||||
child.emit('close', exitCode);
|
||||
});
|
||||
return child;
|
||||
});
|
||||
};
|
||||
|
||||
afterEach(() => {
|
||||
globalThis.fetch = originalFetch;
|
||||
spawnMock = null;
|
||||
});
|
||||
|
||||
describe('createGitLabClient request basics', () => {
|
||||
test('calls {baseUrl}/api/v4{path} and sends PRIVATE-TOKEN', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 42, username: 'alice' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('spawns glab api --include and sends GITLAB_TOKEN env', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ id: 42, username: 'alice' }));
|
||||
|
||||
const client = createGitLabClient({ token: 'glpat-token', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.user();
|
||||
|
||||
expect(fetchMock).toHaveBeenCalledTimes(1);
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/user');
|
||||
expect(options.headers['PRIVATE-TOKEN']).toBe('glpat-token');
|
||||
expect(spawnMock).toHaveBeenCalledTimes(1);
|
||||
const [bin, args, opts] = spawnMock.mock.calls[0];
|
||||
expect(bin).toBe('/home/user/.local/bin/glab');
|
||||
expect(args).toEqual(['api', '--include', '/api/v4/user']);
|
||||
expect(opts.env.GITLAB_TOKEN).toBe('glpat-token');
|
||||
expect(result).toMatchObject({ status: 200, data: { id: 42, username: 'alice' } });
|
||||
expect(result.error).toBeUndefined();
|
||||
});
|
||||
|
||||
test('joins a custom base URL without duplicating /api/v4', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('joins a custom base URL path without duplicating /api/v4', async () => {
|
||||
spawnMock = mockSpawn(cliOutput([]));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.example.com/gitlab/' });
|
||||
await client.issues('group/sub', { state: 'opened' });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.example.com/gitlab/api/v4/projects/group%2Fsub/issues?state=opened');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
// CLI transport passes relative paths; base URL is resolved by glab's config.
|
||||
expect(endpoint).toContain('/api/v4/projects/group%2Fsub/issues?state=opened');
|
||||
});
|
||||
|
||||
test('encodes project path namespaces exactly once', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput([]));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.mergeRequest('a/b/c', 5);
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/a%2Fb%2Fc/merge_requests/5');
|
||||
expect(String(url)).not.toContain('%252F');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
expect(endpoint).toContain('a%2Fb%2Fc/merge_requests/5');
|
||||
expect(endpoint).not.toContain('%252F');
|
||||
});
|
||||
|
||||
test('serializes query params and omits empty ones', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput([]));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.mergeRequests('g/p', { state: 'opened', per_page: 50, page: 2, search: '', sort: null });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
const query = String(url).split('?')[1];
|
||||
expect(query).toContain('state=opened');
|
||||
expect(query).toContain('per_page=50');
|
||||
expect(query).toContain('page=2');
|
||||
expect(query).not.toContain('search');
|
||||
expect(query).not.toContain('sort');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
expect(endpoint).toContain('state=opened');
|
||||
expect(endpoint).toContain('per_page=50');
|
||||
expect(endpoint).toContain('page=2');
|
||||
expect(endpoint).not.toContain('search');
|
||||
expect(endpoint).not.toContain('sort');
|
||||
});
|
||||
|
||||
test('POST requests send a JSON body', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ ok: true }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('POST requests pass the method flag and a JSON body', async () => {
|
||||
spawnMock = mockSpawn(cliOutput({ ok: true }, { status: 201 }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.request('/some/action', { method: 'POST', body: { hello: 'world' } });
|
||||
|
||||
const [, options] = fetchMock.mock.calls[0];
|
||||
expect(options.method).toBe('POST');
|
||||
expect(options.headers['content-type']).toBe('application/json');
|
||||
expect(options.body).toBe(JSON.stringify({ hello: 'world' }));
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain('-d');
|
||||
expect(args).toContain(JSON.stringify({ hello: 'world' }));
|
||||
});
|
||||
|
||||
test('surfaces error statuses without throwing', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ message: 'nope' }, { status: 401 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ message: 'nope' }, { status: 401 }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.user();
|
||||
@@ -105,119 +141,47 @@ describe('createGitLabClient request basics', () => {
|
||||
expect(result.data).toEqual({ message: 'nope' });
|
||||
});
|
||||
|
||||
test('attaches a caller signal when provided, else a timeout signal', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('returns 500 on CLI process error', async () => {
|
||||
spawnMock = (...args) => {
|
||||
const child = new EventEmitter();
|
||||
child.stdout = new EventEmitter();
|
||||
child.stderr = new EventEmitter();
|
||||
child.kill = vi.fn();
|
||||
queueMicrotask(() => {
|
||||
child.emit('error', new Error('ENOENT'));
|
||||
});
|
||||
return child;
|
||||
};
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const controller = new AbortController();
|
||||
await client.branches('g/p', { per_page: 100 });
|
||||
await client.request('/user', { signal: controller.signal });
|
||||
const result = await client.user();
|
||||
expect(result.status).toBe(500);
|
||||
expect(result.error).toContain('ENOENT');
|
||||
});
|
||||
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(fetchMock.mock.calls[0][1].signal).toEqual(expect.any(AbortSignal));
|
||||
expect(fetchMock.mock.calls[1][1].signal).toBe(controller.signal);
|
||||
test('returns 500 on non-zero exit with no stdout', async () => {
|
||||
spawnMock = mockSpawn('', { exitCode: 1, stderr: 'not authenticated' });
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.user();
|
||||
expect(result.status).toBe(500);
|
||||
expect(result.error).toBe('not authenticated');
|
||||
});
|
||||
});
|
||||
|
||||
describe('pagination', () => {
|
||||
test('parses x-page/x-next-page headers into the page object', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([], {
|
||||
headers: { 'x-page': '2', 'x-next-page': '3', 'x-total-pages': '5' },
|
||||
}));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('page object is null (CLI handles pagination)', async () => {
|
||||
spawnMock = mockSpawn(cliOutput([]));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.issues('g/p', { page: 2 });
|
||||
expect(result.page).toEqual({ page: 2, next: 3, total: 5, hasMore: true });
|
||||
});
|
||||
|
||||
test('falls back to the Link rel=next header when x-next-page is absent', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([], {
|
||||
headers: { link: '<https://gitlab.com/api/v4/projects/g%2Fp/issues?page=3>; rel="next", <...>; rel="last"' },
|
||||
}));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.issues('g/p', { page: 2 });
|
||||
expect(result.page.hasMore).toBe(true);
|
||||
expect(result.page.nextUrl).toBe('https://gitlab.com/api/v4/projects/g%2Fp/issues?page=3');
|
||||
});
|
||||
|
||||
test('reports hasMore=false on the last page', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([], {
|
||||
headers: { 'x-page': '5', 'x-next-page': '', 'x-total-pages': '5' },
|
||||
}));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.issues('g/p', { page: 5 });
|
||||
expect(result.page.hasMore).toBe(false);
|
||||
});
|
||||
});
|
||||
|
||||
describe('redirect handling', () => {
|
||||
test('follows a project-move redirect exactly once, preserving auth headers', async () => {
|
||||
const movedUrl = 'https://gitlab.com/api/v4/projects/new%2Fhome/issues';
|
||||
const fetchMock = vi.fn(async (url) => {
|
||||
if (String(url).includes('/projects/g%2Fp/issues')) {
|
||||
return jsonResponse({}, { status: 301, headers: { location: '/api/v4/projects/new%2Fhome/issues' } });
|
||||
}
|
||||
if (String(url) === movedUrl) {
|
||||
return jsonResponse([{ iid: 1 }]);
|
||||
}
|
||||
return jsonResponse({}, { status: 404 });
|
||||
});
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 'glpat-t', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.issues('g/p');
|
||||
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(result.status).toBe(200);
|
||||
expect(result.data).toEqual([{ iid: 1 }]);
|
||||
const [, secondOptions] = fetchMock.mock.calls[1];
|
||||
expect(secondOptions.headers['PRIVATE-TOKEN']).toBe('glpat-t');
|
||||
});
|
||||
});
|
||||
|
||||
describe('etag conditional cache', () => {
|
||||
test('sends if-none-match and replays a 304 as a 200 with cached body', async () => {
|
||||
const fetchMock = vi.fn(async (_url, options) => {
|
||||
if (options.headers['if-none-match'] === '"v1"') {
|
||||
return new Response(null, { status: 304 });
|
||||
}
|
||||
return jsonResponse({ ok: true }, { headers: { etag: '"v1"' } });
|
||||
});
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 'glpat-t', baseUrl: 'https://gitlab.com' });
|
||||
const first = await client.user();
|
||||
expect(first.status).toBe(200);
|
||||
expect(first.data).toEqual({ ok: true });
|
||||
|
||||
const second = await client.user();
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(fetchMock.mock.calls[1][1].headers['if-none-match']).toBe('"v1"');
|
||||
expect(second.status).toBe(200);
|
||||
expect(second.data).toEqual({ ok: true });
|
||||
});
|
||||
|
||||
test('does not cache POST responses', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ ok: true }));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 'glpat-t', baseUrl: 'https://gitlab.com' });
|
||||
await client.request('/thing', { method: 'POST', body: {} });
|
||||
await client.request('/thing', { method: 'POST', body: {} });
|
||||
expect(fetchMock).toHaveBeenCalledTimes(2);
|
||||
expect(result.page).toBeNull();
|
||||
});
|
||||
});
|
||||
|
||||
describe('merge request write methods', () => {
|
||||
test('createMergeRequest POSTs a JSON body to the merge_requests endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ iid: 5, title: 'New MR' }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ iid: 5, title: 'New MR' }, { status: 201 }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.createMergeRequest('group/sub', {
|
||||
@@ -226,45 +190,42 @@ describe('merge request write methods', () => {
|
||||
title: 'New MR',
|
||||
});
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/merge_requests');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(options.headers['content-type']).toBe('application/json');
|
||||
expect(JSON.parse(options.body)).toEqual({ source_branch: 'feat/x', target_branch: 'main', title: 'New MR' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ source_branch: 'feat/x', target_branch: 'main', title: 'New MR' }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/merge_requests'))).toBe(true);
|
||||
expect(result.status).toBe(201);
|
||||
expect(result.data).toEqual({ iid: 5, title: 'New MR' });
|
||||
});
|
||||
|
||||
test('updateMergeRequest PUTs a JSON body to the merge request endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ iid: 5, title: 'Updated' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ iid: 5, title: 'Updated' }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.updateMergeRequest('group/sub', 5, { title: 'Updated', description: 'Body text' });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/merge_requests/5');
|
||||
expect(options.method).toBe('PUT');
|
||||
expect(options.headers['content-type']).toBe('application/json');
|
||||
expect(JSON.parse(options.body)).toEqual({ title: 'Updated', description: 'Body text' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('PUT');
|
||||
expect(args).toContain(JSON.stringify({ title: 'Updated', description: 'Body text' }));
|
||||
});
|
||||
|
||||
test('mergeMergeRequest PUTs a JSON body to the merge endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ iid: 5, state: 'merged' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ iid: 5, state: 'merged' }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.mergeMergeRequest('group/sub', 5, { squash: true });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/merge_requests/5/merge');
|
||||
expect(options.method).toBe('PUT');
|
||||
expect(JSON.parse(options.body)).toEqual({ squash: true });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('PUT');
|
||||
expect(args).toContain(JSON.stringify({ squash: true }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/merge'))).toBe(true);
|
||||
});
|
||||
|
||||
test('write methods surface error statuses without throwing', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ message: 'Method Not Allowed' }, { status: 405 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ message: 'Method Not Allowed' }, { status: 405 }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.mergeMergeRequest('group/sub', 5, {});
|
||||
@@ -275,98 +236,79 @@ describe('merge request write methods', () => {
|
||||
|
||||
describe('issue and review write methods', () => {
|
||||
test('createIssueNote POSTs a body to the issue notes endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 5, body: 'hi' }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ id: 5, body: 'hi' }, { status: 201 }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.createIssueNote('group/sub', 7, 'Nice catch');
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/issues/7/notes');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(JSON.parse(options.body)).toEqual({ body: 'Nice catch' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ body: 'Nice catch' }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/issues/7/notes'))).toBe(true);
|
||||
expect(result.status).toBe(201);
|
||||
});
|
||||
|
||||
test('createMrNote POSTs a body to the MR notes endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 8, body: 'LGTM' }, { status: 201 }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ id: 8, body: 'LGTM' }, { status: 201 }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.createMrNote('group/sub', 12, 'LGTM');
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/merge_requests/12/notes');
|
||||
expect(options.method).toBe('POST');
|
||||
expect(JSON.parse(options.body)).toEqual({ body: 'LGTM' });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args).toContain(JSON.stringify({ body: 'LGTM' }));
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/merge_requests/12/notes'))).toBe(true);
|
||||
});
|
||||
|
||||
test('updateIssue PUTs params to the issue endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ iid: 7, title: 'Updated' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ iid: 7, title: 'Updated' }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.updateIssue('group/sub', 7, { state_event: 'close', labels: ['bug'], milestone_id: 33 });
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/issues/7');
|
||||
expect(options.method).toBe('PUT');
|
||||
expect(JSON.parse(options.body)).toEqual({ state_event: 'close', labels: ['bug'], milestone_id: 33 });
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('PUT');
|
||||
expect(args).toContain(JSON.stringify({ state_event: 'close', labels: ['bug'], milestone_id: 33 }));
|
||||
});
|
||||
|
||||
test('approveMr POSTs to the approve endpoint', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({ id: 1, state: 'approved' }));
|
||||
globalThis.fetch = fetchMock;
|
||||
spawnMock = mockSpawn(cliOutput({ id: 1, state: 'approved' }));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.approveMr('group/sub', 12);
|
||||
|
||||
const [url, options] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/merge_requests/12/approve');
|
||||
expect(options.method).toBe('POST');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
expect(args).toContain('-X');
|
||||
expect(args).toContain('POST');
|
||||
expect(args.some(a => typeof a === 'string' && a.includes('/approve'))).toBe(true);
|
||||
});
|
||||
|
||||
test('milestones GETs the project milestones list', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse([{ id: 33, title: 'v1.0' }]));
|
||||
globalThis.fetch = fetchMock;
|
||||
test('milestones passes state and per_page query params', async () => {
|
||||
spawnMock = mockSpawn(cliOutput([{ id: 33, title: 'v1.0' }]));
|
||||
|
||||
const client = createGitLabClient({ token: 't', baseUrl: 'https://gitlab.com' });
|
||||
await client.milestones('group/sub', { state: 'all', per_page: 100 });
|
||||
|
||||
const [url] = fetchMock.mock.calls[0];
|
||||
expect(String(url)).toBe('https://gitlab.com/api/v4/projects/group%2Fsub/milestones?state=all&per_page=100');
|
||||
const [, args] = spawnMock.mock.calls[0];
|
||||
const endpoint = args[args.length - 1];
|
||||
expect(endpoint).toContain('state=all');
|
||||
expect(endpoint).toContain('per_page=100');
|
||||
});
|
||||
});
|
||||
|
||||
describe('rate limiting', () => {
|
||||
// NOTE: these tests run last in this file. The rate-limit cooldown is
|
||||
// module-level and has no reset export, so earlier tests must not set one.
|
||||
test('429 surfaces error and records a cooldown', async () => {
|
||||
const fetchMock = vi.fn(async () => jsonResponse({}, { status: 429, headers: { 'retry-after': '5' } }));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 'glpat-t', baseUrl: 'https://gitlab.com' });
|
||||
const result = await client.user();
|
||||
expect(result.status).toBe(429);
|
||||
expect(result.error).toBe('GitLab rate limited');
|
||||
expect(isGitLabRateLimited()).toBe(true);
|
||||
test('isGitLabRateLimited returns false (no-op with CLI transport)', () => {
|
||||
expect(isGitLabRateLimited()).toBe(false);
|
||||
});
|
||||
|
||||
test('short-circuits while the cooldown is active without calling fetch', async () => {
|
||||
test('noteGitLabRateLimit is a no-op', () => {
|
||||
noteGitLabRateLimit({ headers: new Headers({ 'retry-after': '120' }) });
|
||||
const fetchMock = vi.fn(async () => jsonResponse([]));
|
||||
globalThis.fetch = fetchMock;
|
||||
|
||||
const client = createGitLabClient({ token: 'glpat-t', baseUrl: 'https://gitlab.com' });
|
||||
const gated = await client.issues('g/p');
|
||||
expect(gated.status).toBe(429);
|
||||
expect(gated.error).toBe('GitLab rate limited');
|
||||
expect(fetchMock).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
test('parses Retry-After seconds into the cooldown', () => {
|
||||
noteGitLabRateLimit({ headers: new Headers({ 'retry-after': '120' }) });
|
||||
expect(isGitLabRateLimited()).toBe(true);
|
||||
expect(isGitLabRateLimited()).toBe(false);
|
||||
});
|
||||
|
||||
test('getGitLabClientOrNull returns null without stored auth', () => {
|
||||
|
||||
Reference in New Issue
Block a user