Message-ID: <gh-pgjdbc-pgjdbc-3750-c3162565793@github.com>
From: "vlsi (@vlsi)" <noreply+vlsi@github.com>
To: "pgjdbc/pgjdbc" <noreply+pgjdbc-pgjdbc@github.com>
Date: Thu, 07 Aug 2025 05:38:56 +0000
Subject: Re: [pgjdbc/pgjdbc] PR #3750: Use `BufferedInputStream` with `FileInputStream`
In-Reply-To: <gh-pgjdbc-pgjdbc-3750@github.com>
References: <gh-pgjdbc-pgjdbc-3750@github.com>
List-Id: <gh-pgjdbc-pgjdbc.github.com>
X-GitHub-Author-Login: vlsi
X-GitHub-Comment-Id: 3162565793
X-GitHub-Comment-Type: issue_comment
X-GitHub-Issue: 3750
X-GitHub-Repo: pgjdbc/pgjdbc
X-GitHub-Type: comment
X-GitHub-Url: https://github.com/pgjdbc/pgjdbc/pull/3750#issuecomment-3162565793
Content-Type: text/plain; charset=utf-8

`FileUtils` is indeed a better class name.

> The reason being: if you are always going to expect file streams to be buffered, then there is no real point to put it in the name. If you want a file stream, it will be buffered, whether you like it or not

Well, `Files.newBufferedReader` does have `Buffered` in the name. I just thought having `.newBufferedInputStream` would would convey the intent.

>The exclusions list references an imaginary `Unsafe.class` file. Should I replace that with the new file?

Let's replace it indeed.

> Should test code be excluded or updated to use the new util?

AFAIK there are two instances, and they are worth updating:

> `new StrangeInputStream(seed, new FileInputStream("target/buffer.txt"));`

It is fine to convert to a new util method. It looks like a true case when the necessary buffering was missing.

> `new BufferedReader(new InputStreamReader(new FileInputStream(path)));`

It could rather be `Files.newBufferedReader(...)` instead