| `Scanner.useDelimiter("\\A")` is not an efficient way to read an entire |
| `InputStream`. |
| |
| ```java |
| Scanner scanner = new Scanner(inputStream, UTF_8).useDelimiter("\\A"); |
| String s = scanner.hasNext() ? scanner.next() : ""; |
| ``` |
| |
| `Scanner` separates its input into "tokens" based on a delimiter that is a |
| regular expression. The regular expression `\A` matches the beginning of the |
| input, only, so there is no later delimiter and the single token consists of |
| every character read from the `InputStream`. |
| |
| This works, but has multiple drawbacks: |
| |
| * You need a special case for an empty `InputStream`. In that case there is no |
| token after `\A`. That's why the extract above checks `hasNext()`. If you |
| forget to do that, you get `NoSuchElementException` in the empty case. |
| * It swallows `IOException`. Quoting the `Scanner` specification: |
| |
| > A scanner can read text from any object which implements the `Readable` |
| > interface. If an invocation of the underlying readable's `read()` method |
| > throws an `IOException` then the scanner assumes that the end of the input |
| > has been reached. The most recent `IOException` thrown by the underlying |
| > readable can be retrieved via the `ioException()` method. |
| |
| * It is much slower than calling `inputStream.readAllBytes()`. |
| |
| Instead, prefer one of the following alternatives: |
| |
| Since Java 9, it has been possible to write this: |
| |
| ```java |
| String s = new String(inputStream.readAllBytes(), UTF_8); |
| ``` |
| |
| On Android, that does require API level 33, though. Guava's |
| [`ByteStreams.toByteArray(inputStream)`](https://guava.dev/releases/snapshot-jre/api/docs/com/google/common/io/ByteStreams.html#toByteArray\(java.io.InputStream\)) |
| is equivalent to `inputStream.readAllBytes()`. |
