docs: clarification on minimal example with full error handling

首先查看现有文档中涵盖 Client、Pool、connect、query、end 和 error events 的章节。将这些章节与 issue 中的两个最小示例进行比较；当文档清楚地解释相关的错误处理行为，以及针对单个查询使用 Client 或 Pool 的权衡时，文档工作就完成了。

The docs somewhat talk about it in pieces, but it wasn't 100% clear what a minimal code example with full error handling looks like, so I was hoping to get some clarity around that. I believe the following are both minimal examples, but I have a few questions surrounding them.

Client example:

import pg from 'pg';

const client = new pg.Client();

await client.connect();

try {
  await client.query('SELECT $1::text as message', ['Hello world!']);
} finally {
  await client.end();
}

Pool example:

import pg from 'pg';

const pool = new pg.Pool();

try {
  await pool.query('SELECT $1::text as message', ['Hello world!']);
} finally {
  await pool.end();
}

My couple questions:

1. Does either example need an actual event listener for error, or because I'm calling connect -> query -> end directly after each other, is it not possible for an error event to actually be emitted? I know in normal case errors that happen during those are rejected with the promise, but I wasn't sure if it was actually possible for an error event to be emitted between say the connect and query calls, etc.
2. Is there any advantage or disadvantage to using Pool even if it's only a single call and will then be torn down? The only advantage I can see is it requires one less line of code since you don't need to explicitly call connect, but I didn't know if there was other overhead or considered an antipattern*?

* I know instantiating a bunch of pools is an antipattern, but think of a usecase more akin to lambda or something, where it's not possible to keep a pool active, so I'm basically just spinning up some code, running a single query, then spinning everything down. Does using Pool have any negatives for that use case, or does defaulting to Pool over Client make sense?

Thanks for any info/help!