Implement a Python Built-in Function in C

Topic Overview

Built-in Functions in C

Builtins like len, abs, and sum are C functions registered on the builtins module. They parse PyObject* arguments, do native work, and return new Python objects without ever entering the bytecode interpreter loop.

PyCFunction Pattern

A builtin uses the same C API as extensions but lives in Python/bltinmodule.c or similar:

cCopy
Loading…

Every builtin you study follows the same checklist:

• Parse: PyArg_ParseTuple with a format string like "i" or "O"
• Validate: PyLong_Check, PySequence_Check, or similar type guards
• Return: a new reference (PyLong_FromLong) or NULL with PyErr_SetString

Type checks, error handling, and correct refcounting are mandatory. Fast paths often bypass generic PyNumber dispatch when the argument type is known.

For example, abs(-42) calls C code that reads the PyLong digits directly instead of re-entering the interpreter loop.

Design Notes

Fast paths should validate types once, then read internal struct fields directly. Falling back to abstract PyNumber APIs preserves semantics but costs the speed that motivated writing C in the first place.

For This Exercise

This exercise asks you to implement a Python builtin in C following CPython conventions. You will write argument parsing, type validation, and return value construction for a function callable from the REPL.

You will use the same mental model here when reading production interpreter source later in the track. Sketch one concrete input on paper, predict the outcome, then confirm with code. That discipline catches logic errors early and makes debugging far faster when you extend the implementation in follow-on tasks.

Your Task

Implement a new built-in, reverse(s), the way Python/bltinmodule.c implements builtins, together with the few builtins needed to inspect strings: len, ord, chr, getsizeof and kind. The real work is Unicode. A CPython str is a sequence of code points, not bytes. PEP 393 stores it with 1, 2 or 4 bytes per code point, chosen by its largest character. So reverse must decode the UTF-8 source, reverse the code points, and the result keeps the same storage kind.

Input

One expression per line: a string literal, an integer, or a call name(arg, ...) whose arguments are expressions (calls nest).

• String literals use '...' or "...". They contain UTF-8 text and the escapes \\n \\t \\\\ \\' \\" \\xNN \\uNNNN \\UNNNNNNNN.
• Integers are decimal, with an optional -.

Builtins

The empty string is ascii.

Errors (printed instead of a result)

• TypeError: NAME() takes exactly one argument (N given)
• TypeError: reverse() argument must be str, not int (the same for getsizeof and kind)
• TypeError: object of type 'int' has no len()
• TypeError: ord() expected a character, but string of length N found
• TypeError: ord() expected string of length 1, but int found
• TypeError: an integer is required (got type str) (from chr)
• ValueError: chr() arg not in range(0x110000)
• NameError: name 'X' is not defined
• Syntax errors:
  • SyntaxError: EOL while scanning string literal
  • SyntaxError: invalid UTF-8 in source (malformed, overlong or surrogate encodings)
  • SyntaxError: illegal Unicode character (\\U above 0x10FFFF)
  • SyntaxError: truncated \\xXX escape (likewise \\uXXXX, \\UXXXXXXXX) when the escape lacks hex digits
  • SyntaxError: unsupported escape
  • SyntaxError: invalid syntax

Output

Print each result's repr:

• Integers: in decimal.
• Quotes: strings are in single quotes, or double quotes if they contain ' and no ". The chosen quote and \\ are escaped with a backslash.
• Escapes: \\n \\t \\r stay escapes. Other characters below 0x20, and 0x7F..0x9F, become \\xNN, and surrogates (from chr) become \\udNNN, all in lowercase hex.
• Everything else is written as UTF-8.

Example

Input:

cCopy
Loading…

Output:

cCopy
Loading…

Requirements

1. Decode UTF-8 into code points before doing anything else. Reversing bytes would corrupt every multibyte character.
2. Compute the storage kind from the maximum code point, as PyUnicode_New(size, maxchar) does.

Hidden tests cover 4-byte characters (emoji), a combining accent that moves when reversed, escapes and quote choice in repr, a lone surrogate from chr, and the type, value, name and syntax errors.