CINXE.COM

<!DOCTYPE html> <html lang="en"> <head> <meta charset="utf-8"> <meta name="viewport" content="width=device-width, initial-scale=1.0"> <meta name="color-scheme" content="light dark"> <title>PEP 563 – Postponed Evaluation of Annotations | peps.python.org</title> <link rel="shortcut icon" href="../_static/py.png"> <link rel="canonical" href="https://peps.python.org/pep-0563/"> <link rel="stylesheet" href="../_static/style.css" type="text/css"> <link rel="stylesheet" href="../_static/mq.css" type="text/css"> <link rel="stylesheet" href="../_static/pygments.css" type="text/css" media="(prefers-color-scheme: light)" id="pyg-light"> <link rel="stylesheet" href="../_static/pygments_dark.css" type="text/css" media="(prefers-color-scheme: dark)" id="pyg-dark"> <link rel="alternate" type="application/rss+xml" title="Latest PEPs" href="https://peps.python.org/peps.rss"> <meta property="og:title" content='PEP 563 – Postponed Evaluation of Annotations | peps.python.org'> <meta property="og:description" content="PEP 3107 introduced syntax for function annotations, but the semantics were deliberately left undefined. PEP 484 introduced a standard meaning to annotations: type hints. PEP 526 defined variable annotations, explicitly tying them with the type hintin..."> <meta property="og:type" content="website"> <meta property="og:url" content="https://peps.python.org/pep-0563/"> <meta property="og:site_name" content="Python Enhancement Proposals (PEPs)"> <meta property="og:image" content="https://peps.python.org/_static/og-image.png"> <meta property="og:image:alt" content="Python PEPs"> <meta property="og:image:width" content="200"> <meta property="og:image:height" content="200"> <meta name="description" content="PEP 3107 introduced syntax for function annotations, but the semantics were deliberately left undefined. PEP 484 introduced a standard meaning to annotations: type hints. PEP 526 defined variable annotations, explicitly tying them with the type hintin..."> <meta name="theme-color" content="#3776ab"> </head> <body> <svg xmlns="http://www.w3.org/2000/svg" style="display: none;"> <symbol id="svg-sun-half" viewBox="0 0 24 24" pointer-events="all"> <title>Following system colour scheme</title> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"> <circle cx="12" cy="12" r="9"></circle> <path d="M12 3v18m0-12l4.65-4.65M12 14.3l7.37-7.37M12 19.6l8.85-8.85"></path> </svg> </symbol> <symbol id="svg-moon" viewBox="0 0 24 24" pointer-events="all"> <title>Selected dark colour scheme</title> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"> <path stroke="none" d="M0 0h24v24H0z" fill="none"></path> <path d="M12 3c.132 0 .263 0 .393 0a7.5 7.5 0 0 0 7.92 12.446a9 9 0 1 1 -8.313 -12.454z"></path> </svg> </symbol> <symbol id="svg-sun" viewBox="0 0 24 24" pointer-events="all"> <title>Selected light colour scheme</title> <svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"> <circle cx="12" cy="12" r="5"></circle> <line x1="12" y1="1" x2="12" y2="3"></line> <line x1="12" y1="21" x2="12" y2="23"></line> <line x1="4.22" y1="4.22" x2="5.64" y2="5.64"></line> <line x1="18.36" y1="18.36" x2="19.78" y2="19.78"></line> <line x1="1" y1="12" x2="3" y2="12"></line> <line x1="21" y1="12" x2="23" y2="12"></line> <line x1="4.22" y1="19.78" x2="5.64" y2="18.36"></line> <line x1="18.36" y1="5.64" x2="19.78" y2="4.22"></line> </svg> </symbol> </svg> <script> document.documentElement.dataset.colour_scheme = localStorage.getItem("colour_scheme") || "auto" </script> <section id="pep-page-section"> <header> <h1>Python Enhancement Proposals</h1> <ul class="breadcrumbs"> <li><a href="https://www.python.org/" title="The Python Programming Language">Python</a> » </li> <li><a href="../pep-0000/">PEP Index</a> » </li> <li>PEP 563</li> </ul> <button id="colour-scheme-cycler" onClick="setColourScheme(nextColourScheme())"> <svg aria-hidden="true" class="colour-scheme-icon-when-auto"><use href="#svg-sun-half"></use></svg> <svg aria-hidden="true" class="colour-scheme-icon-when-dark"><use href="#svg-moon"></use></svg> <svg aria-hidden="true" class="colour-scheme-icon-when-light"><use href="#svg-sun"></use></svg> Toggle light / dark / auto colour theme </button> </header> <article> <section id="pep-content"> <h1 class="page-title">PEP 563 – Postponed Evaluation of Annotations</h1> <dl class="rfc2822 field-list simple"> <dt class="field-odd">Author:</dt> <dd class="field-odd">Łukasz Langa <lukasz at python.org></dd> <dt class="field-even">Discussions-To:</dt> <dd class="field-even"><a class="reference external" href="https://mail.python.org/archives/list/python-dev@python.org/">Python-Dev list</a></dd> <dt class="field-odd">Status:</dt> <dd class="field-odd"><abbr title="Normative proposal accepted for implementation">Accepted</abbr></dd> <dt class="field-even">Type:</dt> <dd class="field-even"><abbr title="Normative PEP with a new feature for Python, implementation change for CPython or interoperability standard for the ecosystem">Standards Track</abbr></dd> <dt class="field-odd">Topic:</dt> <dd class="field-odd"><a class="reference external" href="../topic/typing/">Typing</a></dd> <dt class="field-even">Created:</dt> <dd class="field-even">08-Sep-2017</dd> <dt class="field-odd">Python-Version:</dt> <dd class="field-odd">3.7</dd> <dt class="field-even">Post-History:</dt> <dd class="field-even">01-Nov-2017, 21-Nov-2017</dd> <dt class="field-odd">Superseded-By:</dt> <dd class="field-odd"><a class="reference external" href="../pep-0649/">649</a></dd> <dt class="field-even">Resolution:</dt> <dd class="field-even"><a class="reference external" href="https://mail.python.org/pipermail/python-dev/2017-December/151042.html">Python-Dev message</a></dd> </dl> <hr class="docutils" /> <section id="contents"> <details><summary>Table of Contents</summary><ul class="simple"> <li><a class="reference internal" href="#abstract">Abstract</a></li> <li><a class="reference internal" href="#rationale-and-goals">Rationale and Goals</a><ul> <li><a class="reference internal" href="#non-goals">Non-goals</a></li> <li><a class="reference internal" href="#non-typing-usage-of-annotations">Non-typing usage of annotations</a></li> </ul> </li> <li><a class="reference internal" href="#implementation">Implementation</a><ul> <li><a class="reference internal" href="#enabling-the-future-behavior-in-python-3-7">Enabling the future behavior in Python 3.7</a></li> </ul> </li> <li><a class="reference internal" href="#resolving-type-hints-at-runtime">Resolving Type Hints at Runtime</a><ul> <li><a class="reference internal" href="#runtime-annotation-resolution-and-class-decorators">Runtime annotation resolution and class decorators</a></li> <li><a class="reference internal" href="#runtime-annotation-resolution-and-type-checking">Runtime annotation resolution and <code class="docutils literal notranslate">TYPE_CHECKING</code></a></li> </ul> </li> <li><a class="reference internal" href="#backwards-compatibility">Backwards Compatibility</a><ul> <li><a class="reference internal" href="#deprecation-policy">Deprecation policy</a></li> </ul> </li> <li><a class="reference internal" href="#forward-references">Forward References</a></li> <li><a class="reference internal" href="#rejected-ideas">Rejected Ideas</a><ul> <li><a class="reference internal" href="#keeping-the-ability-to-use-function-local-state-when-defining-annotations">Keeping the ability to use function local state when defining annotations</a></li> <li><a class="reference internal" href="#disallowing-local-state-usage-for-classes-too">Disallowing local state usage for classes, too</a></li> <li><a class="reference internal" href="#introducing-a-new-dictionary-for-the-string-literal-form-instead">Introducing a new dictionary for the string literal form instead</a></li> <li><a class="reference internal" href="#dropping-annotations-with-o">Dropping annotations with -O</a></li> <li><a class="reference internal" href="#passing-string-literals-in-annotations-verbatim-to-annotations">Passing string literals in annotations verbatim to <code class="docutils literal notranslate">__annotations__</code></a></li> <li><a class="reference internal" href="#making-the-name-of-the-future-import-more-verbose">Making the name of the future import more verbose</a></li> </ul> </li> <li><a class="reference internal" href="#prior-discussion">Prior discussion</a><ul> <li><a class="reference internal" href="#in-pep-484">In PEP 484</a></li> <li><a class="reference internal" href="#python-typing-400">python/typing#400</a></li> <li><a class="reference internal" href="#first-draft-discussion-on-python-ideas">First draft discussion on python-ideas</a></li> <li><a class="reference internal" href="#second-draft-discussion-on-python-dev">Second draft discussion on python-dev</a></li> </ul> </li> <li><a class="reference internal" href="#acknowledgements">Acknowledgements</a></li> <li><a class="reference internal" href="#copyright">Copyright</a></li> </ul> </details></section> <section id="abstract"> <h2><a class="toc-backref" href="#abstract" role="doc-backlink">Abstract</a></h2> <a class="pep reference internal" href="../pep-3107/" title="PEP 3107 – Function Annotations">PEP 3107</a> introduced syntax for function annotations, but the semantics were deliberately left undefined. <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> introduced a standard meaning to annotations: type hints. <a class="pep reference internal" href="../pep-0526/" title="PEP 526 – Syntax for Variable Annotations">PEP 526</a> defined variable annotations, explicitly tying them with the type hinting use case. This PEP proposes changing function annotations and variable annotations so that they are no longer evaluated at function definition time. Instead, they are preserved in <code class="docutils literal notranslate">__annotations__</code> in string form. This change is being introduced gradually, starting with a <code class="docutils literal notranslate">__future__</code> import in Python 3.7. </section> <section id="rationale-and-goals"> <h2><a class="toc-backref" href="#rationale-and-goals" role="doc-backlink">Rationale and Goals</a></h2> <a class="pep reference internal" href="../pep-3107/" title="PEP 3107 – Function Annotations">PEP 3107</a> added support for arbitrary annotations on parts of a function definition. Just like default values, annotations are evaluated at function definition time. This creates a number of issues for the type hinting use case: <ul class="simple"> <li>forward references: when a type hint contains names that have not been defined yet, that definition needs to be expressed as a string literal;</li> <li>type hints are executed at module import time, which is not computationally free.</li> </ul> Postponing the evaluation of annotations solves both problems. NOTE: <a class="pep reference internal" href="../pep-0649/" title="PEP 649 – Deferred Evaluation Of Annotations Using Descriptors">PEP 649</a> proposes an alternative solution to the above issues, putting this PEP in danger of being superseded. <section id="non-goals"> <h3><a class="toc-backref" href="#non-goals" role="doc-backlink">Non-goals</a></h3> Just like in <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> and <a class="pep reference internal" href="../pep-0526/" title="PEP 526 – Syntax for Variable Annotations">PEP 526</a>, it should be emphasized that Python will remain a dynamically typed language, and the authors have no desire to ever make type hints mandatory, even by convention. This PEP is meant to solve the problem of forward references in type annotations. There are still cases outside of annotations where forward references will require usage of string literals. Those are listed in a later section of this document. Annotations without forced evaluation enable opportunities to improve the syntax of type hints. This idea will require its own separate PEP and is not discussed further in this document. </section> <section id="non-typing-usage-of-annotations"> <h3><a class="toc-backref" href="#non-typing-usage-of-annotations" role="doc-backlink">Non-typing usage of annotations</a></h3> While annotations are still available for arbitrary use besides type checking, it is worth mentioning that the design of this PEP, as well as its precursors (<a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> and <a class="pep reference internal" href="../pep-0526/" title="PEP 526 – Syntax for Variable Annotations">PEP 526</a>), is predominantly motivated by the type hinting use case. In Python 3.8 <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> will graduate from provisional status. Other enhancements to the Python programming language like <a class="pep reference internal" href="../pep-0544/" title="PEP 544 – Protocols: Structural subtyping (static duck typing)">PEP 544</a>, <a class="pep reference internal" href="../pep-0557/" title="PEP 557 – Data Classes">PEP 557</a>, or <a class="pep reference internal" href="../pep-0560/" title="PEP 560 – Core support for typing module and generic types">PEP 560</a>, are already being built on this basis as they depend on type annotations and the <code class="docutils literal notranslate">typing</code> module as defined by <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a>. In fact, the reason <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> is staying provisional in Python 3.7 is to enable rapid evolution for another release cycle that some of the aforementioned enhancements require. With this in mind, uses for annotations incompatible with the aforementioned PEPs should be considered deprecated. </section> </section> <section id="implementation"> <h2><a class="toc-backref" href="#implementation" role="doc-backlink">Implementation</a></h2> With this PEP, function and variable annotations will no longer be evaluated at definition time. Instead, a string form will be preserved in the respective <code class="docutils literal notranslate">__annotations__</code> dictionary. Static type checkers will see no difference in behavior, whereas tools using annotations at runtime will have to perform postponed evaluation. The string form is obtained from the AST during the compilation step, which means that the string form might not preserve the exact formatting of the source. Note: if an annotation was a string literal already, it will still be wrapped in a string. Annotations need to be syntactically valid Python expressions, also when passed as literal strings (i.e. <code class="docutils literal notranslate">compile(literal, '', 'eval')</code>). Annotations can only use names present in the module scope as postponed evaluation using local names is not reliable (with the sole exception of class-level names resolved by <code class="docutils literal notranslate">typing.get_type_hints()</code>). Note that as per <a class="pep reference internal" href="../pep-0526/" title="PEP 526 – Syntax for Variable Annotations">PEP 526</a>, local variable annotations are not evaluated at all since they are not accessible outside of the function’s closure. <section id="enabling-the-future-behavior-in-python-3-7"> <h3><a class="toc-backref" href="#enabling-the-future-behavior-in-python-3-7" role="doc-backlink">Enabling the future behavior in Python 3.7</a></h3> The functionality described above can be enabled starting from Python 3.7 using the following special import: <div class="highlight-default notranslate"><div class="highlight"><pre>from __future__ import annotations </pre></div> </div> A reference implementation of this functionality is available <a class="reference external" href="https://github.com/python/cpython/pull/4390">on GitHub</a>. </section> </section> <section id="resolving-type-hints-at-runtime"> <h2><a class="toc-backref" href="#resolving-type-hints-at-runtime" role="doc-backlink">Resolving Type Hints at Runtime</a></h2> To resolve an annotation at runtime from its string form to the result of the enclosed expression, user code needs to evaluate the string. For code that uses type hints, the <code class="docutils literal notranslate">typing.get_type_hints(obj, globalns=None, localns=None)</code> function correctly evaluates expressions back from its string form. Note that all valid code currently using <code class="docutils literal notranslate">__annotations__</code> should already be doing that since a type annotation can be expressed as a string literal. For code which uses annotations for other purposes, a regular <code class="docutils literal notranslate">eval(ann, globals, locals)</code> call is enough to resolve the annotation. In both cases it’s important to consider how globals and locals affect the postponed evaluation. An annotation is no longer evaluated at the time of definition and, more importantly, in the same scope where it was defined. Consequently, using local state in annotations is no longer possible in general. As for globals, the module where the annotation was defined is the correct context for postponed evaluation. The <code class="docutils literal notranslate">get_type_hints()</code> function automatically resolves the correct value of <code class="docutils literal notranslate">globalns</code> for functions and classes. It also automatically provides the correct <code class="docutils literal notranslate">localns</code> for classes. When running <code class="docutils literal notranslate">eval()</code>, the value of globals can be gathered in the following way: <ul> <li>function objects hold a reference to their respective globals in an attribute called <code class="docutils literal notranslate">__globals__</code>;</li> <li>classes hold the name of the module they were defined in, this can be used to retrieve the respective globals:<div class="highlight-default notranslate"><div class="highlight"><pre>cls_globals = vars(sys.modules[SomeClass.__module__]) </pre></div> </div> Note that this needs to be repeated for base classes to evaluate all <code class="docutils literal notranslate">__annotations__</code>. </li> <li>modules should use their own <code class="docutils literal notranslate">__dict__</code>.</li> </ul> The value of <code class="docutils literal notranslate">localns</code> cannot be reliably retrieved for functions because in all likelihood the stack frame at the time of the call no longer exists. For classes, <code class="docutils literal notranslate">localns</code> can be composed by chaining vars of the given class and its base classes (in the method resolution order). Since slots can only be filled after the class was defined, we don’t need to consult them for this purpose. <section id="runtime-annotation-resolution-and-class-decorators"> <h3><a class="toc-backref" href="#runtime-annotation-resolution-and-class-decorators" role="doc-backlink">Runtime annotation resolution and class decorators</a></h3> Metaclasses and class decorators that need to resolve annotations for the current class will fail for annotations that use the name of the current class. Example: <div class="highlight-default notranslate"><div class="highlight"><pre>def class_decorator(cls): annotations = get_type_hints(cls) # raises NameError on 'C' print(f'Annotations for {cls}: {annotations}') return cls @class_decorator class C: singleton: 'C' = None </pre></div> </div> This was already true before this PEP. The class decorator acts on the class before it’s assigned a name in the current definition scope. </section> <section id="runtime-annotation-resolution-and-type-checking"> <h3><a class="toc-backref" href="#runtime-annotation-resolution-and-type-checking" role="doc-backlink">Runtime annotation resolution and <code class="docutils literal notranslate">TYPE_CHECKING</code></a></h3> Sometimes there’s code that must be seen by a type checker but should not be executed. For such situations the <code class="docutils literal notranslate">typing</code> module defines a constant, <code class="docutils literal notranslate">TYPE_CHECKING</code>, that is considered <code class="docutils literal notranslate">True</code> during type checking but <code class="docutils literal notranslate">False</code> at runtime. Example: <div class="highlight-default notranslate"><div class="highlight"><pre>import typing if typing.TYPE_CHECKING: import expensive_mod def a_func(arg: expensive_mod.SomeClass) -> None: a_var: expensive_mod.SomeClass = arg ... </pre></div> </div> This approach is also useful when handling import cycles. Trying to resolve annotations of <code class="docutils literal notranslate">a_func</code> at runtime using <code class="docutils literal notranslate">typing.get_type_hints()</code> will fail since the name <code class="docutils literal notranslate">expensive_mod</code> is not defined (<code class="docutils literal notranslate">TYPE_CHECKING</code> variable being <code class="docutils literal notranslate">False</code> at runtime). This was already true before this PEP. </section> </section> <section id="backwards-compatibility"> <h2><a class="toc-backref" href="#backwards-compatibility" role="doc-backlink">Backwards Compatibility</a></h2> This is a backwards incompatible change. Applications depending on arbitrary objects to be directly present in annotations will break if they are not using <code class="docutils literal notranslate">typing.get_type_hints()</code> or <code class="docutils literal notranslate">eval()</code>. Annotations that depend on locals at the time of the function definition will not be resolvable later. Example: <div class="highlight-default notranslate"><div class="highlight"><pre>def generate(): A = Optional[int] class C: field: A = 1 def method(self, arg: A) -> None: ... return C X = generate() </pre></div> </div> Trying to resolve annotations of <code class="docutils literal notranslate">X</code> later by using <code class="docutils literal notranslate">get_type_hints(X)</code> will fail because <code class="docutils literal notranslate">A</code> and its enclosing scope no longer exists. Python will make no attempt to disallow such annotations since they can often still be successfully statically analyzed, which is the predominant use case for annotations. Annotations using nested classes and their respective state are still valid. They can use local names or the fully qualified name. Example: <div class="highlight-default notranslate"><div class="highlight"><pre>class C: field = 'c_field' def method(self) -> C.field: # this is OK ... def method(self) -> field: # this is OK ... def method(self) -> C.D: # this is OK ... def method(self) -> D: # this is OK ... class D: field2 = 'd_field' def method(self) -> C.D.field2: # this is OK ... def method(self) -> D.field2: # this FAILS, class D is local to C ... # and is therefore only available # as C.D. This was already true # before the PEP. def method(self) -> field2: # this is OK ... def method(self) -> field: # this FAILS, field is local to C and # is therefore not visible to D unless # accessed as C.field. This was already # true before the PEP. </pre></div> </div> In the presence of an annotation that isn’t a syntactically valid expression, SyntaxError is raised at compile time. However, since names aren’t resolved at that time, no attempt is made to validate whether used names are correct or not. <section id="deprecation-policy"> <h3><a class="toc-backref" href="#deprecation-policy" role="doc-backlink">Deprecation policy</a></h3> Starting with Python 3.7, a <code class="docutils literal notranslate">__future__</code> import is required to use the described functionality. No warnings are raised. NOTE: Whether this will eventually become the default behavior is currently unclear pending decision on <a class="pep reference internal" href="../pep-0649/" title="PEP 649 – Deferred Evaluation Of Annotations Using Descriptors">PEP 649</a>. In any case, use of annotations that depend upon their eager evaluation is incompatible with both proposals and is no longer supported. </section> </section> <section id="forward-references"> <h2><a class="toc-backref" href="#forward-references" role="doc-backlink">Forward References</a></h2> Deliberately using a name before it was defined in the module is called a forward reference. For the purpose of this section, we’ll call any name imported or defined within a <code class="docutils literal notranslate">if TYPE_CHECKING:</code> block a forward reference, too. This PEP addresses the issue of forward references in type annotations. The use of string literals will no longer be required in this case. However, there are APIs in the <code class="docutils literal notranslate">typing</code> module that use other syntactic constructs of the language, and those will still require working around forward references with string literals. The list includes: <ul> <li>type definitions:<div class="highlight-default notranslate"><div class="highlight"><pre>T = TypeVar('T', bound='<type>') UserId = NewType('UserId', '<type>') Employee = NamedTuple('Employee', [('name', '<type>'), ('id', '<type>')]) </pre></div> </div> </li> <li>aliases:<div class="highlight-default notranslate"><div class="highlight"><pre>Alias = Optional['<type>'] AnotherAlias = Union['<type>', '<type>'] YetAnotherAlias = '<type>' </pre></div> </div> </li> <li>casting:<div class="highlight-default notranslate"><div class="highlight"><pre>cast('<type>', value) </pre></div> </div> </li> <li>base classes:<div class="highlight-default notranslate"><div class="highlight"><pre>class C(Tuple['<type>', '<type>']): ... </pre></div> </div> </li> </ul> Depending on the specific case, some of the cases listed above might be worked around by placing the usage in a <code class="docutils literal notranslate">if TYPE_CHECKING:</code> block. This will not work for any code that needs to be available at runtime, notably for base classes and casting. For named tuples, using the new class definition syntax introduced in Python 3.6 solves the issue. In general, fixing the issue for all forward references requires changing how module instantiation is performed in Python, from the current single-pass top-down model. This would be a major change in the language and is out of scope for this PEP. </section> <section id="rejected-ideas"> <h2><a class="toc-backref" href="#rejected-ideas" role="doc-backlink">Rejected Ideas</a></h2> <section id="keeping-the-ability-to-use-function-local-state-when-defining-annotations"> <h3><a class="toc-backref" href="#keeping-the-ability-to-use-function-local-state-when-defining-annotations" role="doc-backlink">Keeping the ability to use function local state when defining annotations</a></h3> With postponed evaluation, this would require keeping a reference to the frame in which an annotation got created. This could be achieved for example by storing all annotations as lambdas instead of strings. This would be prohibitively expensive for highly annotated code as the frames would keep all their objects alive. That includes predominantly objects that won’t ever be accessed again. To be able to address class-level scope, the lambda approach would require a new kind of cell in the interpreter. This would proliferate the number of types that can appear in <code class="docutils literal notranslate">__annotations__</code>, as well as wouldn’t be as introspectable as strings. Note that in the case of nested classes, the functionality to get the effective “globals” and “locals” at definition time is provided by <code class="docutils literal notranslate">typing.get_type_hints()</code>. If a function generates a class or a function with annotations that have to use local variables, it can populate the given generated object’s <code class="docutils literal notranslate">__annotations__</code> dictionary directly, without relying on the compiler. </section> <section id="disallowing-local-state-usage-for-classes-too"> <h3><a class="toc-backref" href="#disallowing-local-state-usage-for-classes-too" role="doc-backlink">Disallowing local state usage for classes, too</a></h3> This PEP originally proposed limiting names within annotations to only allow names from the model-level scope, including for classes. The author argued this makes name resolution unambiguous, including in cases of conflicts between local names and module-level names. This idea was ultimately rejected in case of classes. Instead, <code class="docutils literal notranslate">typing.get_type_hints()</code> got modified to populate the local namespace correctly if class-level annotations are needed. The reasons for rejecting the idea were that it goes against the intuition of how scoping works in Python, and would break enough existing type annotations to make the transition cumbersome. Finally, local scope access is required for class decorators to be able to evaluate type annotations. This is because class decorators are applied before the class receives its name in the outer scope. </section> <section id="introducing-a-new-dictionary-for-the-string-literal-form-instead"> <h3><a class="toc-backref" href="#introducing-a-new-dictionary-for-the-string-literal-form-instead" role="doc-backlink">Introducing a new dictionary for the string literal form instead</a></h3> Yury Selivanov shared the following idea: <ol class="arabic simple"> <li>Add a new special attribute to functions: <code class="docutils literal notranslate">__annotations_text__</code>.</li> <li>Make <code class="docutils literal notranslate">__annotations__</code> a lazy dynamic mapping, evaluating expressions from the corresponding key in <code class="docutils literal notranslate">__annotations_text__</code> just-in-time.</li> </ol> This idea is supposed to solve the backwards compatibility issue, removing the need for a new <code class="docutils literal notranslate">__future__</code> import. Sadly, this is not enough. Postponed evaluation changes which state the annotation has access to. While postponed evaluation fixes the forward reference problem, it also makes it impossible to access function-level locals anymore. This alone is a source of backwards incompatibility which justifies a deprecation period. A <code class="docutils literal notranslate">__future__</code> import is an obvious and explicit indicator of opting in for the new functionality. It also makes it trivial for external tools to recognize the difference between a Python files using the old or the new approach. In the former case, that tool would recognize that local state access is allowed, whereas in the latter case it would recognize that forward references are allowed. Finally, just-in-time evaluation in <code class="docutils literal notranslate">__annotations__</code> is an unnecessary step if <code class="docutils literal notranslate">get_type_hints()</code> is used later. </section> <section id="dropping-annotations-with-o"> <h3><a class="toc-backref" href="#dropping-annotations-with-o" role="doc-backlink">Dropping annotations with -O</a></h3> There are two reasons this is not satisfying for the purpose of this PEP. First, this only addresses runtime cost, not forward references, those still cannot be safely used in source code. A library maintainer would never be able to use forward references since that would force the library users to use this new hypothetical -O switch. Second, this throws the baby out with the bath water. Now no runtime annotation use can be performed. <a class="pep reference internal" href="../pep-0557/" title="PEP 557 – Data Classes">PEP 557</a> is one example of a recent development where evaluating type annotations at runtime is useful. All that being said, a granular -O option to drop annotations is a possibility in the future, as it’s conceptually compatible with existing -O behavior (dropping docstrings and assert statements). This PEP does not invalidate the idea. </section> <section id="passing-string-literals-in-annotations-verbatim-to-annotations"> <h3><a class="toc-backref" href="#passing-string-literals-in-annotations-verbatim-to-annotations" role="doc-backlink">Passing string literals in annotations verbatim to <code class="docutils literal notranslate">__annotations__</code></a></h3> This PEP originally suggested directly storing the contents of a string literal under its respective key in <code class="docutils literal notranslate">__annotations__</code>. This was meant to simplify support for runtime type checkers. Mark Shannon pointed out this idea was flawed since it wasn’t handling situations where strings are only part of a type annotation. The inconsistency of it was always apparent but given that it doesn’t fully prevent cases of double-wrapping strings anyway, it is not worth it. </section> <section id="making-the-name-of-the-future-import-more-verbose"> <h3><a class="toc-backref" href="#making-the-name-of-the-future-import-more-verbose" role="doc-backlink">Making the name of the future import more verbose</a></h3> Instead of requiring the following import: <div class="highlight-default notranslate"><div class="highlight"><pre>from __future__ import annotations </pre></div> </div> the PEP could call the feature more explicitly, for example <code class="docutils literal notranslate">string_annotations</code>, <code class="docutils literal notranslate">stringify_annotations</code>, <code class="docutils literal notranslate">annotation_strings</code>, <code class="docutils literal notranslate">annotations_as_strings</code>, <code class="docutils literal notranslate">lazy_annotations</code>, <code class="docutils literal notranslate">static_annotations</code>, etc. The problem with those names is that they are very verbose. Each of them besides <code class="docutils literal notranslate">lazy_annotations</code> would constitute the longest future feature name in Python. They are long to type and harder to remember than the single-word form. There is precedence of a future import name that sounds overly generic but in practice was obvious to users as to what it does: <div class="highlight-default notranslate"><div class="highlight"><pre>from __future__ import division </pre></div> </div> </section> </section> <section id="prior-discussion"> <h2><a class="toc-backref" href="#prior-discussion" role="doc-backlink">Prior discussion</a></h2> <section id="in-pep-484"> <h3><a class="toc-backref" href="#in-pep-484" role="doc-backlink">In PEP 484</a></h3> The forward reference problem was discussed when <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> was originally drafted, leading to the following statement in the document: <blockquote> <div>A compromise is possible where a <code class="docutils literal notranslate">__future__</code> import could enable turning all annotations in a given module into string literals, as follows:<div class="highlight-default notranslate"><div class="highlight"><pre>from __future__ import annotations class ImSet: def add(self, a: ImSet) -> List[ImSet]: ... assert ImSet.add.__annotations__ == { 'a': 'ImSet', 'return': 'List[ImSet]' } </pre></div> </div> Such a <code class="docutils literal notranslate">__future__</code> import statement may be proposed in a separate PEP. </div></blockquote> </section> <section id="python-typing-400"> <h3><a class="toc-backref" href="#python-typing-400" role="doc-backlink">python/typing#400</a></h3> The problem was discussed at length on the typing module’s GitHub project, under <a class="reference external" href="https://github.com/python/typing/issues/400">Issue 400</a>. The problem statement there includes critique of generic types requiring imports from <code class="docutils literal notranslate">typing</code>. This tends to be confusing to beginners: <blockquote> <div>Why this:<div class="highlight-default notranslate"><div class="highlight"><pre>from typing import List, Set def dir(o: object = ...) -> List[str]: ... def add_friends(friends: Set[Friend]) -> None: ... </pre></div> </div> But not this: <div class="highlight-default notranslate"><div class="highlight"><pre>def dir(o: object = ...) -> list[str]: ... def add_friends(friends: set[Friend]) -> None ... </pre></div> </div> Why this: <div class="highlight-default notranslate"><div class="highlight"><pre>up_to_ten = list(range(10)) friends = set() </pre></div> </div> But not this: <div class="highlight-default notranslate"><div class="highlight"><pre>from typing import List, Set up_to_ten = List[int](range(10)) friends = Set[Friend]() </pre></div> </div> </div></blockquote> While typing usability is an interesting problem, it is out of scope of this PEP. Specifically, any extensions of the typing syntax standardized in <a class="pep reference internal" href="../pep-0484/" title="PEP 484 – Type Hints">PEP 484</a> will require their own respective PEPs and approval. Issue 400 ultimately suggests postponing evaluation of annotations and keeping them as strings in <code class="docutils literal notranslate">__annotations__</code>, just like this PEP specifies. This idea was received well. Ivan Levkivskyi supported using the <code class="docutils literal notranslate">__future__</code> import and suggested unparsing the AST in <code class="docutils literal notranslate">compile.c</code>. Jukka Lehtosalo pointed out that there are some cases of forward references where types are used outside of annotations and postponed evaluation will not help those. For those cases using the string literal notation would still be required. Those cases are discussed briefly in the “Forward References” section of this PEP. The biggest controversy on the issue was Guido van Rossum’s concern that untokenizing annotation expressions back to their string form has no precedent in the Python programming language and feels like a hacky workaround. He said: <blockquote> <div>One thing that comes to mind is that it’s a very random change to the language. It might be useful to have a more compact way to indicate deferred execution of expressions (using less syntax than <code class="docutils literal notranslate">lambda:</code>). But why would the use case of type annotations be so all-important to change the language to do it there first (rather than proposing a more general solution), given that there’s already a solution for this particular use case that requires very minimal syntax?</div></blockquote> Eventually, Ethan Smith and schollii voiced that feedback gathered during PyCon US suggests that the state of forward references needs fixing. Guido van Rossum suggested coming back to the <code class="docutils literal notranslate">__future__</code> idea, pointing out that to prevent abuse, it’s important for the annotations to be kept both syntactically valid and evaluating correctly at runtime. </section> <section id="first-draft-discussion-on-python-ideas"> <h3><a class="toc-backref" href="#first-draft-discussion-on-python-ideas" role="doc-backlink">First draft discussion on python-ideas</a></h3> Discussion happened largely in two threads, <a class="reference external" href="https://mail.python.org/pipermail/python-ideas/2017-September/thread.html#47031">the original announcement</a> and a follow-up called <a class="reference external" href="https://mail.python.org/pipermail/python-ideas/2017-September/thread.html#47108">PEP 563 and expensive backwards compatibility</a>. The PEP received rather warm feedback (4 strongly in favor, 2 in favor with concerns, 2 against). The biggest voice of concern on the former thread being Steven D’Aprano’s review stating that the problem definition of the PEP doesn’t justify breaking backwards compatibility. In this response Steven seemed mostly concerned about Python no longer supporting evaluation of annotations that depended on local function/class state. A few people voiced concerns that there are libraries using annotations for non-typing purposes. However, none of the named libraries would be invalidated by this PEP. They do require adapting to the new requirement to call <code class="docutils literal notranslate">eval()</code> on the annotation with the correct <code class="docutils literal notranslate">globals</code> and <code class="docutils literal notranslate">locals</code> set. This detail about <code class="docutils literal notranslate">globals</code> and <code class="docutils literal notranslate">locals</code> having to be correct was picked up by a number of commenters. Alyssa (Nick) Coghlan benchmarked turning annotations into lambdas instead of strings, sadly this proved to be much slower at runtime than the current situation. The latter thread was started by Jim J. Jewett who stressed that the ability to properly evaluate annotations is an important requirement and backwards compatibility in that regard is valuable. After some discussion he admitted that side effects in annotations are a code smell and modal support to either perform or not perform evaluation is a messy solution. His biggest concern remained loss of functionality stemming from the evaluation restrictions on global and local scope. Alyssa Coghlan pointed out that some of those evaluation restrictions from the PEP could be lifted by a clever implementation of an evaluation helper, which could solve self-referencing classes even in the form of a class decorator. She suggested the PEP should provide this helper function in the standard library. </section> <section id="second-draft-discussion-on-python-dev"> <h3><a class="toc-backref" href="#second-draft-discussion-on-python-dev" role="doc-backlink">Second draft discussion on python-dev</a></h3> Discussion happened mainly in the <a class="reference external" href="https://mail.python.org/pipermail/python-dev/2017-November/150062.html">announcement thread</a>, followed by a brief discussion under <a class="reference external" href="https://mail.python.org/pipermail/python-dev/2017-November/150637.html">Mark Shannon’s post</a>. Steven D’Aprano was concerned whether it’s acceptable for typos to be allowed in annotations after the change proposed by the PEP. Brett Cannon responded that type checkers and other static analyzers (like linters or programming text editors) will catch this type of error. Jukka Lehtosalo added that this situation is analogous to how names in function bodies are not resolved until the function is called. A major topic of discussion was Alyssa Coghlan’s suggestion to store annotations in “thunk form”, in other words as a specialized lambda which would be able to access class-level scope (and allow for scope customization at call time). He presented a possible design for it (<a class="reference external" href="https://mail.python.org/pipermail/python-dev/2017-November/150141.html">indirect attribute cells</a>). This was later seen as equivalent to “special forms” in Lisp. Guido van Rossum expressed worry that this sort of feature cannot be safely implemented in twelve weeks (i.e. in time before the Python 3.7 beta freeze). After a while it became clear that the point of division between supporters of the string form vs. supporters of the thunk form is actually about whether annotations should be perceived as a general syntactic element vs. something tied to the type checking use case. Finally, Guido van Rossum declared he’s rejecting the thunk idea based on the fact that it would require a new building block in the interpreter. This block would be exposed in annotations, multiplying possible types of values stored in <code class="docutils literal notranslate">__annotations__</code> (arbitrary objects, strings, and now thunks). Moreover, thunks aren’t as introspectable as strings. Most importantly, Guido van Rossum explicitly stated interest in gradually restricting the use of annotations to static typing (with an optional runtime component). Alyssa Coghlan got convinced to <a class="pep reference internal" href="../pep-0563/" title="PEP 563 – Postponed Evaluation of Annotations">PEP 563</a>, too, promptly beginning the mandatory bike shedding session on the name of the <code class="docutils literal notranslate">__future__</code> import. Many debaters agreed that <code class="docutils literal notranslate">annotations</code> seems like an overly broad name for the feature name. Guido van Rossum briefly decided to call it <code class="docutils literal notranslate">string_annotations</code> but then changed his mind, arguing that <code class="docutils literal notranslate">division</code> is a precedent of a broad name with a clear meaning. The final improvement to the PEP suggested in the discussion by Mark Shannon was the rejection of the temptation to pass string literals through to <code class="docutils literal notranslate">__annotations__</code> verbatim. A side-thread of discussion started around the runtime penalty of static typing, with topic like the import time of the <code class="docutils literal notranslate">typing</code> module (which is comparable to <code class="docutils literal notranslate">re</code> without dependencies, and three times as heavy as <code class="docutils literal notranslate">re</code> when counting dependencies). </section> </section> <section id="acknowledgements"> <h2><a class="toc-backref" href="#acknowledgements" role="doc-backlink">Acknowledgements</a></h2> This document could not be completed without valuable input, encouragement and advice from Guido van Rossum, Jukka Lehtosalo, and Ivan Levkivskyi. The implementation was thoroughly reviewed by Serhiy Storchaka who found all sorts of issues, including bugs, bad readability, and performance problems. </section> <section id="copyright"> <h2><a class="toc-backref" href="#copyright" role="doc-backlink">Copyright</a></h2> This document has been placed in the public domain. </section> </section> <hr class="docutils" /> Source: <a class="reference external" href="https://github.com/python/peps/blob/main/peps/pep-0563.rst">https://github.com/python/peps/blob/main/peps/pep-0563.rst</a> Last modified: <a class="reference external" href="https://github.com/python/peps/commits/main/peps/pep-0563.rst">2024-03-24 01:43:58 GMT</a> </article> <nav id="pep-sidebar"> <h2>Contents</h2> <ul> <li><a class="reference internal" href="#abstract">Abstract</a></li> <li><a class="reference internal" href="#rationale-and-goals">Rationale and Goals</a><ul> <li><a class="reference internal" href="#non-goals">Non-goals</a></li> <li><a class="reference internal" href="#non-typing-usage-of-annotations">Non-typing usage of annotations</a></li> </ul> </li> <li><a class="reference internal" href="#implementation">Implementation</a><ul> <li><a class="reference internal" href="#enabling-the-future-behavior-in-python-3-7">Enabling the future behavior in Python 3.7</a></li> </ul> </li> <li><a class="reference internal" href="#resolving-type-hints-at-runtime">Resolving Type Hints at Runtime</a><ul> <li><a class="reference internal" href="#runtime-annotation-resolution-and-class-decorators">Runtime annotation resolution and class decorators</a></li> <li><a class="reference internal" href="#runtime-annotation-resolution-and-type-checking">Runtime annotation resolution and <code class="docutils literal notranslate">TYPE_CHECKING</code></a></li> </ul> </li> <li><a class="reference internal" href="#backwards-compatibility">Backwards Compatibility</a><ul> <li><a class="reference internal" href="#deprecation-policy">Deprecation policy</a></li> </ul> </li> <li><a class="reference internal" href="#forward-references">Forward References</a></li> <li><a class="reference internal" href="#rejected-ideas">Rejected Ideas</a><ul> <li><a class="reference internal" href="#keeping-the-ability-to-use-function-local-state-when-defining-annotations">Keeping the ability to use function local state when defining annotations</a></li> <li><a class="reference internal" href="#disallowing-local-state-usage-for-classes-too">Disallowing local state usage for classes, too</a></li> <li><a class="reference internal" href="#introducing-a-new-dictionary-for-the-string-literal-form-instead">Introducing a new dictionary for the string literal form instead</a></li> <li><a class="reference internal" href="#dropping-annotations-with-o">Dropping annotations with -O</a></li> <li><a class="reference internal" href="#passing-string-literals-in-annotations-verbatim-to-annotations">Passing string literals in annotations verbatim to <code class="docutils literal notranslate">__annotations__</code></a></li> <li><a class="reference internal" href="#making-the-name-of-the-future-import-more-verbose">Making the name of the future import more verbose</a></li> </ul> </li> <li><a class="reference internal" href="#prior-discussion">Prior discussion</a><ul> <li><a class="reference internal" href="#in-pep-484">In PEP 484</a></li> <li><a class="reference internal" href="#python-typing-400">python/typing#400</a></li> <li><a class="reference internal" href="#first-draft-discussion-on-python-ideas">First draft discussion on python-ideas</a></li> <li><a class="reference internal" href="#second-draft-discussion-on-python-dev">Second draft discussion on python-dev</a></li> </ul> </li> <li><a class="reference internal" href="#acknowledgements">Acknowledgements</a></li> <li><a class="reference internal" href="#copyright">Copyright</a></li> </ul> <a id="source" href="https://github.com/python/peps/blob/main/peps/pep-0563.rst">Page Source (GitHub)</a> </nav> </section> <script src="../_static/colour_scheme.js"></script> <script src="../_static/wrap_tables.js"></script> <script src="../_static/sticky_banner.js"></script> </body> </html>