Bug: TablesDB listRows total:false uses a GET encoding that does not suppress totals in measured Cloud requests

オープン
#168 コメント 3 件 リアクション 0 件 担当者 0 名 GitHub で見る

まだ誰も着手していません。

評価

難易度
4/5
見積もり時間
3〜5日
初心者へのやさしさ
48/100
issue の種類
バグ
明瞭さ
おおむね明確
活発さ
活発
技術スタック
nodejs, typescript

調査の方向性

Start with the pinned src/client.ts source and the corresponding SDK Generator platform templates, then verify GET boolean parsing against a nonempty list; the report notes that an empty-table fixture is insufficient. Confirm the intended server contract before choosing whether the serializer, server parser, or both should change. Done means regression coverage for true, false, omission, query arrays, and POST-body behavior, while preserving JSON booleans and booleans inside serialized Query JSON.

索引モデルが issue の本文から書いたものです。

説明

Reproduction steps

With an isolated, frozen install of [email protected], run the following .mjs file with Node 26. It sends one HTTP GET to a local server and requires no Appwrite credentials or Cloud writes.

import { Client, TablesDB } from "node-appwrite";
import assert from "node:assert/strict";
import { createServer } from "node:http";

let observed;
const server = createServer((request, response) => {
  observed = new URL(request.url, "http://fixture.invalid").searchParams.get("total");
  response.writeHead(200, { "content-type": "application/json" });
  response.end('{"total":0,"rows":[]}');
});
await new Promise(resolve => server.listen(0, "127.0.0.1", resolve));
try {
  const client = new Client().setEndpoint(
    `http://127.0.0.1:${server.address().port}/v1`,
  ).setProject("fixture-project");
  await new TablesDB(client).listRows({
    databaseId: "db",
    tableId: "cost",
    total: false,
  });
  console.log({ observed }); // { observed: 'false' }
  assert.equal(observed, "false"); // Characterization, not the desired fix.
} finally {
  server.closeAllConnections();
  await new Promise(resolve => server.close(resolve));
}
Expected behavior

Passing total: false should suppress total counting on a nonempty list. The SDK and the server should agree on the GET boolean representation. In the measured Cloud boundary below, total=0 produced the intended response while total=false did not. Please confirm the intended server contract before deciding whether to fix the serializer, server parser, or both.

Actual behavior

The published SDK emits the string false in the GET URL. A separate, read-only Appwrite Cloud SGP probe queried an existing nonempty table with identical limit/order queries:

GET flag Response total Rows returned
total=false 8 1
total=0 0 1
total=true 8 1
total=1 8 1

The returned row hash matched between the false/zero variants. The behavior held with and without X-Appwrite-Response-Format: 2.3.0. No row was inserted or changed. An empty-table fixture alone cannot establish that counting was suppressed. This is a response/counting observation, not a measured billing or query-performance claim.

Environment and source pin
  • Published dependency: [email protected], exact frozen lock.
  • SDK tag 29.0.0 resolves to commit f333acaff4fcc23956b6086a537c51c87123f40d: client source.
  • Fresh local replay: Windows, Node 26.10.0, pnpm 12.9.0. The real-SDK compatibility suite passed two tests, including raw string-false characterization and an explicit GET-total-only zero/one shim. The same pinned suite also passed Linux CI.
  • Cloud measurement: SGP, captured October 2026. The Cloud engine version was not independently identified; 2.3.0 above is the requested response-format header, not a claim about the deployed engine.
  • No claim that other SDK/server versions behave identically. Project IDs, credentials, private receipts and row content are excluded.
Suggested investigation and regression coverage

Please verify GET boolean parsing against a nonempty list and the corresponding SDK Generator platform templates. If zero/one serialization is required, preserve JSON request-body booleans and booleans inside serialized Query JSON. Cover true, false, omission, query arrays and POST-body behavior; make the desired-fix regression fail against the previous implementation first.

Our prototype uses an explicit local shim for the GET total flag. We have not patched the vendor package or migrated production Functions to the Node SDK. A general upstream change should be reviewed at the generator/server contract boundary rather than copying that narrow workaround.

Duplicate check

Searched open and closed sdk-for-node issues/PRs for total and false, and sdk-generator issues/PRs for total with false; no matching report was found at posting time. Only this report is being opened; no parallel issue is being filed in the generator repository.

主要言語
TypeScript
スター
256
フォーク
45
平均マージ
8時間 59分
マージ済み PR(30日)
4

環境構築

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

appwrite/sdk-for-node のほかの issue

appwrite/sdk-for-node の issue をすべて見る

似ている issue

TypeScript の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。