info > ATTRIBUTES(7)

🏷️ NAME

attributes - get/set subroutine or variable attributes

🚀 Quick Reference

🔧 Use Case	💻 Command	📝 Description
Retrieve attributes	my @attrs = attributes::get(\&foo);	Get list of attributes on a subroutine
Declare a method	sub foo : method {}	Suppress ambiguous‑call warnings
Declare an lvalue sub	sub foo : lvalue {}	Subroutine that can be assigned to; must return modifiable value
Constant anonymous sub	my $s = sub : const { 'value' };	Evaluated immediately and turned into a constant
Get reference type	my $type = attributes::reftype(\$var);	Returns 'SCALAR', 'ARRAY', etc.
Custom attribute validation	sub MODIFY_CODE_ATTRIBUTES { … }	Handle user‑defined attributes in a package

📜 SYNOPSIS

    sub foo : method ;
    my ($x,@y,%z) : Bent = 1;
    my $s = sub : method { ... };

    use attributes ();    # optional, to get subroutine declarations
    my @attrlist = attributes::get(\&foo);

    use attributes 'get'; # import the attributes::get subroutine
    my @attrlist = get \&foo;

📖 DESCRIPTION

Subroutine declarations and definitions may optionally have attribute lists associated with them. (Variable my declarations also may, but see the warning below.) Perl handles these declarations by passing information about the call site and the thing being declared along with the attribute list to this module. The first example above is equivalent to:

    use attributes __PACKAGE__, \&foo, 'method';

The second example does something equivalent to:

    use attributes ();
    my ($x,@y,%z);
    attributes::->import(__PACKAGE__, \$x, 'Bent');
    attributes::->import(__PACKAGE__, \@y, 'Bent');
    attributes::->import(__PACKAGE__, \%z, 'Bent');
    ($x,@y,%z) = 1;

Yes, that's a lot of expansion.

⚠️ WARNING: attribute declarations for variables are still evolving. The semantics and interfaces of such declarations could change in future versions. They are present for purposes of experimentation with what the semantics ought to be. Do not rely on the current implementation of this feature.

There are only a few attributes currently handled by Perl itself (or directly by this module). However, package‑specific attributes are allowed by an extension mechanism. (See “Package‑specific Attribute Handling” below.)

The setting of subroutine attributes happens at compile time. Variable attributes in our declarations are also applied at compile time. However, my variables get their attributes applied at run‑time. This means that you have to reach the run‑time component of the my before those attributes will get applied. For example:

    my $x : Bent = 42 if 0;

will neither assign 42 to $x nor apply the Bent attribute.

An attempt to set an unrecognized attribute is a fatal error. (The error is trappable, but it still stops compilation within that eval.) Setting an attribute whose name is all lowercase and not a built‑in (such as foo) will produce a warning with -w or use warnings 'reserved'.

❓ What import does

The description says

    sub foo : method;

is equivalent to

    use attributes __PACKAGE__, \&foo, 'method';

This calls the import function of attributes at compile time with these parameters: the module name, the caller’s package, the code reference and 'method'.

    attributes->import( __PACKAGE__, \&foo, 'method' );

So what does import actually do?

First, import gets the type of the third parameter ('CODE' in this case). attributes.pm checks if there is a subroutine called MODIFY_<reftype>_ATTRIBUTES in the caller’s namespace (here: 'main'). In this example a subroutine MODIFY_CODE_ATTRIBUTES is required. Then this method is called to check for “bad attributes”. The call would look like

    MODIFY_CODE_ATTRIBUTES( 'main', \&foo, 'method' );

MODIFY_<reftype>_ATTRIBUTES must return a list of all “bad attributes”. If any are present, import croaks.

(See “Package‑specific Attribute Handling” below.)

🧱 Built-in Attributes

The following are the built‑in attributes for subroutines:

• lvalue – Indicates the referenced subroutine is a valid lvalue and can be assigned to. The subroutine must return a modifiable value such as a scalar variable (see perlsub).
  This module allows setting this attribute on a subroutine that is already defined. For Perl subroutines (XSUBs are fine), the result may or may not do what you want, depending on the code inside. A warning is emitted if warnings are enabled. You should only do this if you really know what you are doing.
• method – Indicates the referenced subroutine is a method. A subroutine so marked will not trigger the “Ambiguous call resolved as CORE::%s” warning.
• prototype(..) – Alternate means of specifying a prototype on a sub. The desired prototype is within the parens.
  The prototype from the attribute is assigned to the sub immediately after the prototype from the sub, which means that if both are declared at the same time, the traditionally defined prototype is ignored. In other words, sub foo($$) : prototype(@) {} is indistinguishable from sub foo(@){}.
  If illegalproto warnings are enabled, the prototype declared inside this attribute will be sanity checked at compile time.
• const – This experimental attribute, introduced in Perl 5.22, only applies to anonymous subroutines. It causes the subroutine to be called as soon as the sub expression is evaluated. The return value is captured and turned into a constant subroutine.

The following are the built‑in attributes for variables:

• shared – Indicates the referenced variable can be shared across different threads when used with the threads and threads::shared modules.

📦 Available Subroutines

The following subroutines are available for general use once this module has been loaded:

• get – Expects a single parameter – a reference to a subroutine or variable. Returns a list of attributes, which may be empty. If passed invalid arguments, it uses die() (via Carp::croak) to raise a fatal exception. If it can find an appropriate package name for a class method lookup, it will include the results from a FETCH_type_ATTRIBUTES call in its return list, as described in “Package‑specific Attribute Handling” below. Otherwise, only built‑in attributes will be returned.
• reftype – Expects a single parameter – a reference to a subroutine or variable. Returns the built‑in type of the referenced variable, ignoring any package into which it might have been blessed. Useful for determining the type value that forms part of the method names described in “Package‑specific Attribute Handling” below.

These routines are not exported by default.

🔧 Package‑specific Attribute Handling

⚠️ WARNING: the mechanisms described here are still experimental. Do not rely on the current implementation. In particular, there is no provision for applying package attributes to ‘cloned’ copies of subroutines used as closures. Package‑specific attribute handling may change incompatibly in a future release.

When an attribute list is present in a declaration, a check is made to see whether an attribute ‘modify’ handler is present in the appropriate package (or its @ISA inheritance tree). Similarly, when attributes::get is called on a valid reference, a check is made for an appropriate attribute ‘fetch’ handler. See “EXAMPLES” to see how the “appropriate package” determination works.

The handler names are based on the underlying type of the variable being declared or of the reference passed. This deliberately ignores any possibility of being blessed into some package. Thus, a subroutine declaration uses “CODE” as its type, and even a blessed hash reference uses “HASH” as its type.

The class methods invoked for modifying and fetching are:

• FETCH_type_ATTRIBUTES – Called with two arguments: the relevant package name, and a reference to a variable or subroutine for which package‑defined attributes are desired. The expected return value is a list of associated attributes (may be empty).
• MODIFY_type_ATTRIBUTES – Called with two fixed arguments, followed by the list of attributes from the relevant declaration. The two fixed arguments are the relevant package name and a reference to the declared subroutine or variable. The expected return value is a list of attributes that were not recognized by this handler. This allows a derived class to delegate a call to its base class and then only examine the attributes the base class didn’t already handle.

The call to MODIFY_type_ATTRIBUTES is made during the processing of the declaration. In particular, a subroutine reference will probably be for an undefined subroutine, even if this declaration is actually part of the definition.

Calling attributes::get() from within the scope of a null package declaration package ; for an unblessed variable reference will not provide any starting package name for the ‘fetch’ method lookup, and thus will not result in a method call for package‑defined attributes. A named subroutine knows to which symbol table entry it belongs and will use the corresponding package. An anonymous subroutine knows the package name into which it was compiled (unless compiled with a null package declaration), and so it will use that package name.

🔡 Syntax of Attribute Lists

An attribute list is a sequence of attribute specifications, separated by whitespace or a colon (with optional whitespace). Each attribute specification is a simple name, optionally followed by a parenthesised parameter list. If a parameter list is present, it is scanned past as for the rules of the q() operator (see “Quote and Quote‑like Operators” in perlop). The parameter list is passed as it was found, however, and not as per q().

Some examples of syntactically valid attribute lists:

• switch(10,foo(7,3)) : expensive
• Ugly('\(") :Bad
• _5x5
• lvalue method

Some examples of syntactically invalid attribute lists (with annotation):

• switch(10,foo() – ()‑string not balanced
• Ugly('(') – ()‑string not balanced
• 5x5 – “5x5” not a valid identifier
• Y2::north – not a simple identifier
• foo + bar – “+” neither a colon nor whitespace

📤 EXPORTS

🛑 Default exports

None.

🔌 Available exports

The routines get and reftype are exportable.

🏷️ Export tags defined

The :ALL tag will get all of the above exports.

🔍 EXAMPLES

Samples of syntactically valid declarations, with annotation as to how they resolve internally into use attributes invocations by perl. These illustrate how the “appropriate package” is found for possible method lookups for package‑defined attributes.

1. Code:

       package Canine;
       package Dog;
       my Canine $spot : Watchful ;

   Effect:

       use attributes ();
       attributes::->import(Canine => \$spot, "Watchful");

2. Code:

       package Felis;
       my $cat : Nervous;

   Effect:

       use attributes ();
       attributes::->import(Felis => \$cat, "Nervous");

3. Code:

       package X;
       sub foo : lvalue ;

   Effect:

       use attributes X => \&foo, "lvalue";

4. Code:

       package X;
       sub Y::x : lvalue { 1 }

   Effect:

       use attributes Y => \&Y::x, "lvalue";

5. Code:

       package X;
       sub foo { 1 }
   
       package Y;
       BEGIN { *bar = \&X::foo; }
   
       package Z;
       sub Y::bar : lvalue ;

   Effect:

       use attributes X => \&X::foo, "lvalue";

This last example is purely for completeness. You should not be trying to mess with the attributes of something in a package that's not your own.

💡 MORE EXAMPLES

1.     sub MODIFY_CODE_ATTRIBUTES {
          my ($class,$code,@attrs) = @_;
   
          my $allowed = 'MyAttribute';
          my @bad = grep { $_ ne $allowed } @attrs;
   
          return @bad;
       }
   
       sub foo : MyAttribute {
          print "foo\n";
       }

   This example runs. At compile time MODIFY_CODE_ATTRIBUTES is called. We check if any attribute is disallowed and return a list of “bad attributes”. Since we return an empty list, everything is fine.

2.     sub MODIFY_CODE_ATTRIBUTES {
          my ($class,$code,@attrs) = @_;
   
          my $allowed = 'MyAttribute';
          my @bad = grep{ $_ ne $allowed }@attrs;
   
          return @bad;
       }
   
       sub foo : MyAttribute Test {
          print "foo\n";
       }

   This example is aborted at compile time because we use the attribute Test which isn’t allowed. MODIFY_CODE_ATTRIBUTES returns a list containing 'Test'.

🔗 SEE ALSO

“Private Variables via my()” in perlsub, “Subroutine Attributes” in perlsub for details on basic declarations; “use” in perlfunc for details on the normal invocation mechanism.

🏷️ NAME

attributes - POSIX safety concepts

🚀 Quick Reference

🛡️ Use Case	📋 Concept	📝 Description
Check thread safety	MT‑Safe / MT‑Unsafe markers	Look in function’s ATTRIBUTES section of man page
Initialise safely	init feature	Call function once in single‑thread context before starting threads
Avoid data races	race feature	Use mutexes around functions that modify shared state
Handle constant objects	const feature	Use read‑write lock; readers are safe, writers need write lock
Signal interference	sig feature	Block or ensure exclusive use of temporary signal handler
Terminal safety	term feature	Use a mutex to protect terminal attribute changes
Locale / environment reads	locale, env	Safe as long as modifying functions are not called concurrently

📖 DESCRIPTION

Note: the text of this man page is based on the material taken from the “POSIX Safety Concepts” section of the GNU C Library manual. Further details on the topics described here can be found in that manual.

Various function manual pages include a section ATTRIBUTES that describes the safety of calling the function in various contexts. This section annotates functions with the following safety markings:

• MT‑Safe – MT‑Safe or Thread‑Safe functions are safe to call in the presence of other threads. MT stands for Multi Thread.
  Being MT‑Safe does not imply atomicity, nor that the function uses memory synchronisation primitives. Calling two MT‑Safe functions in sequence is not guaranteed to be MT‑Safe. Whole‑program optimisations that inline functions across library interfaces may expose unsafe reordering, so inlining across the GNU C Library interface is not recommended. However, functions defined in user‑visible headers are designed to be safe for inlining.
• MT‑Unsafe – MT‑Unsafe functions are not safe to call in a multithreaded program.

Other keywords that appear in safety notes are defined in subsequent sections.

⚠️ Conditionally safe features

For some features that make functions unsafe to call in certain contexts, there are known ways to avoid the safety problem other than refraining from calling the function altogether. The keywords that follow refer to such features, and each of their definitions indicates how the whole program needs to be constrained in order to remove the safety problem indicated by the keyword. Only when all the reasons that make a function unsafe are observed and addressed, by applying the documented constraints, does the function become safe to call in a context.

• 🏁 init – Functions marked with init as an MT‑Unsafe feature perform MT‑Unsafe initialisation when they are first called.
  Calling such a function at least once in single‑threaded mode removes this specific cause for the function to be regarded as MT‑Unsafe. If no other cause remains, the function can then be safely called after other threads are started.
• 🏎️ race – Functions annotated with race operate on objects in ways that may cause data races or similar forms of destructive interference out of concurrent execution. In some cases, the objects are passed to the functions by users; in others, they are used by the functions to return values to users; in others, they are not even exposed to users.
• 🔒 const – Functions marked with const non‑atomically modify internal objects that are better regarded as constant, because a substantial portion of the GNU C Library accesses them without synchronisation. Unlike race, which causes both readers and writers of internal objects to be regarded as MT‑Unsafe, this mark is applied to writers only. Writers remain MT‑Unsafe to call, but the then‑mandatory constness of objects they modify enables readers to be regarded as MT‑Safe (as long as no other reasons for them to be unsafe remain).
  The identifier that follows the const mark will appear by itself as a safety note in readers. Programs that wish to work around this safety issue, so as to call writers, may use a non‑recursive read‑write lock associated with the identifier, and guard all calls to functions marked with const followed by the identifier with a write lock, and all calls to functions marked with the identifier by itself with a read lock.
• 📶 sig – Functions marked with sig may temporarily install a signal handler for internal purposes, which may interfere with other uses of the signal, identified after a colon.
  This safety problem can be worked around by ensuring that no other uses of the signal will take place for the duration of the call. Holding a non‑recursive mutex while calling all functions that use the same temporary signal; blocking that signal before the call and resetting its handler afterwards is recommended.
• 💻 term – Functions marked with term may change the terminal settings in the recommended way, namely: call tcgetattr(3), modify some flags, and then call tcsetattr(3), this creates a window in which changes made by other threads are lost. Thus, functions marked with term are MT‑Unsafe.
  It is advisable for applications using the terminal to avoid concurrent and reentrant interactions with it, by not using it in signal handlers or blocking signals that might use it, and holding a lock while calling these functions and interacting with the terminal. This lock should also be used for mutual exclusion with functions marked with race:tcattr(fd), where fd is a file descriptor for the controlling terminal.

ℹ️ Other safety remarks

Additional keywords may be attached to functions, indicating features that do not make a function unsafe to call, but that may need to be taken into account in certain classes of programs:

• 🌐 locale – Functions annotated with locale read from the locale object without any form of synchronisation. Functions annotated with locale called concurrently with locale changes may behave in ways that do not correspond to any of the locales active during their execution, but an unpredictable mix thereof.
  We do not mark these functions as MT‑Unsafe, however, because functions that modify the locale object are marked with const:locale and regarded as unsafe. Being unsafe, the latter are not to be called when multiple threads are running or asynchronous signals are enabled, and so the locale can be considered effectively constant in these contexts, which makes the former safe.
• 🧬 env – Functions marked with env access the environment with getenv(3) or similar, without any guards to ensure safety in the presence of concurrent modifications.
  We do not mark these functions as MT‑Unsafe, however, because functions that modify the environment are all marked with const:env and regarded as unsafe. Being unsafe, the latter are not to be called when multiple threads are running or asynchronous signals are enabled, and so the environment can be considered effectively constant in these contexts, which makes the former safe.
• 🆔 hostid – The function marked with hostid reads from the system‑wide data structures that hold the “host ID” of the machine. These data structures cannot generally be modified atomically. Since it is expected that the “host ID” will not normally change, the function that reads from it (gethostid(3)) is regarded as safe, whereas the function that modifies it (sethostid(3)) is marked with const:hostid, indicating it may require special care if it is to be called. In this specific case, the special care amounts to system‑wide (not merely intra‑process) coordination.
• 🚨 sigintr – Functions marked with sigintr access the GNU C Library _sigintr internal data structure without any guards to ensure safety in the presence of concurrent modifications.
  We do not mark these functions as MT‑Unsafe, however, because functions that modify this data structure are all marked with const:sigintr and regarded as unsafe. Being unsafe, the latter are not to be called when multiple threads are running or asynchronous signals are enabled, and so the data structure can be considered effectively constant in these contexts, which makes the former safe.
• 📁 cwd – Functions marked with cwd may temporarily change the current working directory during their execution, which may cause relative pathnames to be resolved in unexpected ways in other threads or within asynchronous signal or cancellation handlers.
  This is not enough of a reason to mark so‑marked functions as MT‑Unsafe, but when this behavior is optional (e.g., nftw(3) with FTW_CHDIR), avoiding the option may be a good alternative to using full pathnames or file descriptor‑relative (e.g., openat(2)) system calls.
• :identifier – Annotations may sometimes be followed by identifiers, intended to group several functions that, for example, access the data structures in an unsafe way, as in race and const, or to provide more specific information, such as naming a signal in a function marked with sig. It is envisioned that it may be applied to lock and corrupt as well in the future.
  In most cases, the identifier will name a set of functions, but it may name global objects or function arguments, or identifiable properties or logical components associated with them, with a notation such as, for example, :buf(arg) to denote a buffer associated with the argument arg, or :tcattr(fd) to denote the terminal attributes of a file descriptor fd.
  The most common use for identifiers is to provide logical groups of functions and arguments that need to be protected by the same synchronisation primitive in order to ensure safe operation in a given context.
• /condition – Some safety annotations may be conditional, in that they only apply if a boolean expression involving arguments, global variables or even the underlying kernel evaluates to true. For example, /!ps and /one_per_line indicate the preceding marker only applies when argument ps is NULL, or global variable one_per_line is nonzero.
  When all marks that render a function unsafe are adorned with such conditions, and none of the named conditions hold, then the function can be regarded as safe.

🔗 SEE ALSO

pthreads(7), signal-safety(7)

📚 COLOPHON

This page is part of release 5.10 of the Linux man-pages project. A description of the project, information about reporting bugs, and the latest version of this page, can be found at https://www.kernel.org/doc/man-pages/.