Lectures

We break down the streaming I/O model in Go: why files and streams are read in “chunks”, how the contracts of io.Reader and io.Writer work, how to correctly interpret the pair n/ err (including io.EOF), what a short write is, and how to write reliable read/write loops in practice (counting lines, dumping a file to stdout).

We break down why buffering is needed in Go and how it is implemented in the bufio package: how bufio.Reader “pre-reads” data on top of io.Reader, how to read lines via ReadString/ ReadBytes, and how to handle io.EOF correctly. Separately, we cover how bufio.Writer works, why data can “get stuck” until Flush(), how Flush() and Close() differ, and which beginner mistakes are most common.

We break down how bufio.Scanner works: what a “token” is, why the scanner has a MaxScanTokenSize limit (by default about 64 KiB), how the bufio.ErrTooLong error shows up, how to correctly increase the maximum via Scanner.Buffer, and when it’s better to use bufio.Reader with ReadString/ ReadBytes for full control over reading.

We break down why in Go you should use io.Copy and io.CopyN instead of a manual Read/ Write loop: the argument order dst/ src, correct handling of io.EOF, “partial progress” in written, adding error context via fmt.Errorf, as well as the buffering nuance with bufio.Writer and the need for Flush(). Practice: copying files via os.Open/ os.Create, backups, and previewing the first N bytes.

We break down the “I/O as a layer” approach in Go: why business logic should not know about files, how to design functions around io.Reader/ io.Writer, who is responsible for calling Close(), why bufio.Writer requires Flush(), and how this lets you test code without a disk using strings.NewReader and bytes.Buffer.