From 401c9779714bc4ea308999c11189a7a461930b3a Mon Sep 17 00:00:00 2001 From: Guido van Rossum Date: Thu, 26 Feb 2015 12:17:47 -0800 Subject: [PATCH] Update to PEP 448 -- approved 2 out of 3, rejected *comprehension. --- pep-0448.txt | 115 +++++++++++++++++++++++++++++---------------------- 1 file changed, 65 insertions(+), 50 deletions(-) diff --git a/pep-0448.txt b/pep-0448.txt index ca7de5e1b..bae00befb 100644 --- a/pep-0448.txt +++ b/pep-0448.txt @@ -4,11 +4,11 @@ Version: $Revision$ Last-Modified: $Date$ Author: Joshua Landau Discussions-To: python-ideas@python.org -Status: Draft +Status: Accepted Type: Standards Track Content-Type: text/x-rst Created: 29-Jun-2013 -Python-Version: 3.4 +Python-Version: 3.5 Post-History: @@ -50,15 +50,9 @@ In dictionaries, later values will always override earlier ones:: >>> {**{'x': 2}, 'x': 1} {'x': 1} -Unpacking is proposed to be allowed inside list, set, -and dictionary comprehensions:: - - >>> ranges = [range(i) for i in range(5)] - >>> [*item for item in ranges] - [0, 0, 1, 0, 1, 2, 0, 1, 2, 3] - - >>> {*item for item in ranges} - {0, 1, 2, 3} +This PEP does not include unpacking operators inside list, set and +dictionary comprehensions although this has not been ruled out for +future proposals. Rationale @@ -121,14 +115,6 @@ list(my_range)`` which is now equivalent to just ``[*my_list, *my_tuple, *my_range]``. -The addition of unpacking to comprehensions is a logical extension. -It's usage will primarily be a neat replacement for ``[i for j in -list_of_lists for i in j]``, as the more readable -``[*l for l in list_of_lists]``. The iterable version, -``(*l for l in list_of_lists)``, replaces ``itertools.chain.from_iterable``. -Other uses are possible, but expected to occur rarely. - - Specification ============= @@ -145,7 +131,8 @@ Currently, if an argument is given multiple times — such as a positional argument given both positionally and by keyword — a ``TypeError`` is raised. This remains true for duplicate arguments provided through multiple ``**`` unpackings, -e.g. ``f(**{'x': 2}, **{'x': 3})``. +e.g. ``f(**{'x': 2}, **{'x': 3})``, except that the error will be +detected at runtime. A function looks like this:: @@ -159,40 +146,17 @@ Tuples, lists, sets and dictionaries will allow unpacking. This will act as if the elements from unpacked items were inserted in order at the site of unpacking, much as happens in unpacking in a function-call. Dictionaries require ``**`` unpacking; all the others require ``*`` unpacking. -A dictionary's key remain in a right-to-left priority order, so + +The keys in a dictionary remain in a right-to-left priority order, so ``{**{'a': 1}, 'a': 2, **{'a': 3}}`` evaluates to ``{'a': 3}``. There is no restriction on the number or position of unpackings. -Comprehensions, by simple extension, will support unpacking. As before, -dictionaries require ``**`` unpacking, all the others require ``*`` -unpacking and key priorities are unchanged. - -For example:: - - {*[1, 2, 3], 4, 5, *{6, 7, 8}} - - (*e for e in [[1], [3, 4, 5], [2]]) - - {**dictionary for dictionary in (globals(), locals())} - - {**locals(), "override": None} - -Unbracketed comprehensions in function calls, such as ``f(x for x in it)``, -are already valid. These could be extended to:: - - f(*x for x in it) == f((*x for x in it)) - f(**x for x in it) == f({**x for x in it}) - -However, this is likely to be confusing and is not included in this -PEP. These will throw ``SyntaxError`` and comprehensions with explicit -brackets should be used instead. - Disadvantages ============= The allowable orders for arguments in a function call are more -complicated than before. The simplest explanation for the rules +complicated than before. The simplest explanation for the rules may be "positional arguments precede keyword arguments and ``**`` unpacking; ``*`` unpacking precedes ``**`` unpacking". @@ -200,24 +164,75 @@ Whilst ``*elements, = iterable`` causes ``elements`` to be a list, ``elements = *iterable,`` causes ``elements`` to be a tuple. The reason for this may confuse people unfamiliar with the construct. +Concerns have been raised about the unexpected difference between +duplicate keys in dictionaries being allowed but duplicate keys +in function call syntax raising an error. Although this is already +the case with current syntax, this proposal might exacerbate the +issue. It remains to be seen how much of an issue this is in practice. + + +Variations +========== + +The PEP originally considered whether the ordering of argument types +in a function call (positional, keyword, ``*`` or ``**``) could become +less strict. This met little support so the idea was shelved. + +Earlier iterations of this PEP allowed unpacking operators inside +list, set, and dictionary comprehensions as a flattening operator +over iterables of containers:: + + >>> ranges = [range(i) for i in range(5)] + >>> [*item for item in ranges] + [0, 0, 1, 0, 1, 2, 0, 1, 2, 3] + + >>> {*item for item in ranges} + {0, 1, 2, 3} + +This was met with a mix of strong concerns about readability and mild +support. In order not to disadvantage the less controversial aspects +of the PEP, this was not accepted with the rest of the proposal. + +Unbracketed comprehensions in function calls, such as ``f(x for x in it)``, +are already valid. These could be extended to:: + + f(*x for x in it) == f((*x for x in it)) + f(**x for x in it) == f({**x for x in it}) + +However, it wasn't clear if this was the best behaviour or if it should +unpack into the arguments of the call to `f`. Since this is likely to be +confusing and is of only very marginal utility, it is not included in this +PEP. Instead, these will throw a ``SyntaxError`` and comprehensions with +explicit brackets should be used instead. + + +Approval +======== + +This PEP was accepted by Guido on February 25, 2015 [1]_. + Implementation ============== -An implementation for Python 3.5 is found at Issue 2292 on bug tracker [1]_. +An implementation for Python 3.5 is found at Issue 2292 on bug tracker [2]_. +This currently includes support for unpacking inside comprehensions, which +should be removed. References ========== -.. [1] Issue 2292, "Missing `*`-unpacking generalizations", Thomas Wouters +.. [1] PEP accepted, "PEP 448 review", Guido van Rossum + (https://mail.python.org/pipermail/python-dev/2015-February/138564.html) + +.. [2] Issue 2292, "Missing `*`-unpacking generalizations", Thomas Wouters (http://bugs.python.org/issue2292) -.. [2] Discussion on Python-ideas list, +.. [3] Discussion on Python-ideas list, "list / array comprehensions extension", Alexander Heger (http://mail.python.org/pipermail/python-ideas/2011-December/013097.html) - Copyright =========