Add --max-bytes flag to limit bytes searched per file

[manascb1344]

This PR adds a new --max-bytes NUM flag that tells ripgrep to only read and search the first NUM bytes of each file or stream. This is useful when matches are expected only in a small header region (for example HTTP headers in large cache files), and can greatly reduce I/O.

• New flag: --max-bytes NUM (per file/stream)
• Applied consistently to mmap, regular files, stdin, and multiline searches
• Multiline preallocation is capped by min(NUM, file_size) when --max-bytes is set
• Plays well with --max-count 1 for fast “find first hit in header” use cases
• Includes tests for normal files, stdin, and binary-looking files

Implements the feature requested in #3035.

[manascb1344]

@BurntSushi
All tests and CI checks are now passing for this PR. When you have a moment, could you please review and consider merging the --max-bytes flag change?