Abstract

    This PEP proposes the overloading of the '@' operator for use of
    the string formatting system introduced in Python 3000.

Contents

    This repository holds a subversion checkout of the latest Python
    3000 trunk which is updated using the 'subversion' git branch.
    Updates are then merged back to the 'master' git branch, to keep
    official repository code changes from changes made for this PEP.

    When/if appropriate, a simple 'svn diff' will create the necessary
    patch needed for integration.

Installation

    $ cd py3k && ./configure && make
    $ ./python.exe ../tests/simple.py

Rationale

    The new string formatting system [1] introduced in Python 3000 and
    later backported to the 2.x series provides a far more versatile
    mechanism than the original '%' operator, the latter of which is
    deprecated and slated for removal.

    Part of the design specifically addressed the binary operator
    dilemma, which forced a choice between non-named items (string or
    tuples, for example) versus named items (dictionaries).

    This PEP attempts to reintroduce a binary operator as an optional
    syntactical alternative to a direct function call, and to propose
    a convention which may address the issues of its predecessor.

Examples

    >>> "Simple test with {} {} via tuples" @ ("unnamed", "args")
    'Simple test with unnamed args via tuples'

    >>> "Simple test with {foo} args via dicts" @ {"foo": "named"}
    'Simple test with named args via dicts'

    >>> "Mix of {foo} and {} args" @ ("unnamed", dict("foo"="named"))
    'Mix of named and unnamed args'

    >>> "No, really a {}" @ ({"literal":"dictionary"}, {})
    "No, really a {'literal': 'dictionary'}"