mirrors/ruby
0
0 Fork 0
mirror of https://github.com/ruby/ruby.git synced 2026-10-28 16:36:20 +08:00
Code Activity
Files
master
ruby/doc/extension.rdoc
T

2488 lines
86 KiB
Plaintext
Raw Permalink Blame History

# extension.rdoc - -*- RDoc -*- created at: Mon Aug 7 16:45:54 JST 1995
{日本語}[rdoc-ref:extension.ja.rdoc]
= Creating extension libraries for Ruby
This document explains how to make extension libraries for Ruby.
== Basic Knowledge
In C, variables have types and data do not have types. In contrast,
Ruby variables do not have a static type, and data themselves have
types, so data will need to be converted between the languages.
Objects in Ruby are represented by the C type +VALUE+. Each VALUE
data has its data type.
To retrieve C data from a VALUE, you need to:
1. Identify the VALUE's data type
2. Convert the VALUE into C data
Converting to the wrong data type may cause serious problems.
=== Ruby data types
The Ruby interpreter has the following data types:
+T_NIL+ :: +nil+
+T_OBJECT+ :: ordinary object
+T_CLASS+ :: class
+T_MODULE+ :: module
+T_FLOAT+ :: floating point number
+T_STRING+ :: string
+T_REGEXP+ :: regular expression
+T_ARRAY+ :: array
+T_HASH+ :: associative array
+T_STRUCT+ :: (Ruby) structure
+T_BIGNUM+ :: multi precision integer
+T_FIXNUM+ :: Fixnum(31bit or 63bit integer)
+T_COMPLEX+ :: complex number
+T_RATIONAL+ :: rational number
+T_FILE+ :: IO
+T_TRUE+ :: +true+
+T_FALSE+ :: +false+
+T_DATA+ :: data
+T_SYMBOL+ :: symbol
In addition, there are several other types used internally:
+T_ICLASS+ :: included module
+T_MATCH+ :: MatchData object
+T_UNDEF+ :: undefined
+T_NODE+ :: syntax tree node
+T_ZOMBIE+ :: object awaiting finalization
Most of the types are represented by C structures.
=== Check type of the VALUE data
The macro <tt>TYPE()</tt> defined in ruby.h shows the data type of the VALUE.
<tt>TYPE()</tt> returns the constant number <tt>T_</tt>_XXXX_ described above.
To handle data types, your code will look something like this:
switch (TYPE(obj)) {
case T_FIXNUM:
/* process Fixnum */
break;
case T_STRING:
/* process String */
break;
case T_ARRAY:
/* process Array */
break;
default:
/* raise exception */
rb_raise(rb_eTypeError, "not valid value");
break;
}
There is the data type check function
void Check_Type(VALUE value, int type)
which raises an exception if the VALUE does not have the type
specified.
There are also faster check macros for fixnums and nil.
FIXNUM_P(obj)
NIL_P(obj)
=== Convert VALUE into C data
The data for type +T_NIL+, +T_FALSE+, +T_TRUE+ are +nil+, +false+,
+true+ respectively. They are singletons for the data type. The
equivalent C constants are: +Qnil+, +Qfalse+, +Qtrue+.
<tt>RTEST()</tt> will return true if a VALUE is neither +Qfalse+ nor
+Qnil+. If you need to differentiate +Qfalse+ from +Qnil+,
specifically test against +Qfalse+.
The +T_FIXNUM+ data is a 31bit or 63bit length fixed integer. This
size depends on the size of long: if long is 32bit then +T_FIXNUM+ is
31bit, if long is 64bit then +T_FIXNUM+ is 63bit. +T_FIXNUM+ can be
converted to a C integer by using the <tt>FIX2INT()</tt> macro or
<tt>FIX2LONG()</tt>. Though you have to check that the data is really
FIXNUM before using them, they are faster. <tt>FIX2LONG()</tt> never
raises exceptions, but <tt>FIX2INT()</tt> raises RangeError if the
result is bigger or smaller than the size of int. There are also
<tt>NUM2INT()</tt> and <tt>NUM2LONG()</tt> which converts any Ruby
numbers into C integers. These macros include a type check, so an
exception will be raised if the conversion failed. <tt>NUM2DBL()</tt>
can be used to retrieve the double float value in the same way.
You can use the macros <tt>StringValue()</tt> and
<tt>StringValuePtr()</tt> to get a <tt>char*</tt> from a VALUE.
<tt>StringValue(var)</tt> replaces var's value with the result of
"var.to_str()". <tt>StringValuePtr(var)</tt> does the same
replacement and returns the char* representation of var. These macros
will skip the replacement if var is a String. Notice that the macros
take only the lvalue as their argument, to change the value of var in
place.
You can also use the macro named <tt>StringValueCStr()</tt>. This is
just like <tt>StringValuePtr()</tt>, but always adds a NUL character
at the end of the result. If the result contains a NUL character, this
macro causes the ArgumentError exception. <tt>StringValuePtr()</tt>
doesn't guarantee the existence of a NUL at the end of the result, and
the result may contain NUL.
Other data types have corresponding C structures, e.g. <tt>struct
RArray</tt> for +T_ARRAY+ etc. The VALUE of the type which has the
corresponding structure can be cast to retrieve the pointer to the
struct. The casting macro will be of the form RXXXX for each data
type; for instance, <tt>RARRAY(obj)</tt>. See "ruby.h". However, we
do not recommend to access RXXXX data directly because these data
structures are complex. Use corresponding rb_xxx() functions to
access the internal struct. For example, to access an entry of array,
use <tt>rb_ary_entry(ary, offset)</tt> and <tt>rb_ary_store(ary,
offset, obj)</tt>.
There are some accessing macros for structure members, for example
<code>RSTRING_LEN(str)</code> to get the size of the Ruby String object. The
allocated region can be accessed by <code>RSTRING_PTR(str)</code>.
Notice: Do not change the value of the structure directly, unless you
are responsible for the result. This ends up being the cause of
interesting bugs.
=== Convert C data into VALUE
To convert C data to Ruby values:
+FIXNUM+ ::
left shift 1 bit, and turn on its least significant bit (LSB).
Other pointer values ::
cast to VALUE.
You can determine whether a VALUE is a pointer or not by checking its LSB.
Notice: Ruby does not allow arbitrary pointer values to be a VALUE. They
should be pointers to the structures which Ruby knows about. The known
structures are defined in <ruby.h>.
To convert C numbers to Ruby values, use these macros:
<tt>INT2FIX()</tt> :: for integers within 31bits.
<tt>INT2NUM()</tt> :: for arbitrary sized integers.
<tt>INT2NUM()</tt> converts an integer into a Bignum if it is out
of the +FIXNUM+ range, but is a bit slower.
=== Manipulating Ruby object
As I already mentioned, it is not recommended to modify an object's
internal structure. To manipulate objects, use the functions supplied
by the Ruby interpreter. Some (not all) of the useful functions are
listed below:
==== String functions
<tt>rb_str_new(const char *ptr, rb_len_t len)</tt> ::
Creates a new Ruby string.
<tt>rb_str_new2(const char *ptr)</tt> ::
<tt>rb_str_new_cstr(const char *ptr)</tt> ::
Creates a new Ruby string from a C string. This is equivalent to
<tt>rb_str_new(ptr, strlen(ptr))</tt>.
<tt>rb_str_new_literal(const char *ptr)</tt> ::
Creates a new Ruby string from a C string literal.
<tt>rb_sprintf(const char *format, ...)</tt> ::
<tt>rb_vsprintf(const char *format, va_list ap)</tt> ::
Creates a new Ruby string with printf(3) format.
Note: In the format string, <tt>"%"PRIsVALUE</tt> can be used for Object#to_s
(or Object#inspect if <tt>+</tt> flag is set) output (and related argument
must be a VALUE). Since it conflicts with <tt>"%i"</tt>, for integers in
format strings, use <tt>"%d"</tt>.
<tt>rb_str_append(VALUE str1, VALUE str2)</tt> ::
Appends Ruby string str2 to Ruby string str1.
<tt>rb_str_cat(VALUE str, const char *ptr, rb_len_t len)</tt> ::
Appends len bytes of data from ptr to the Ruby string.
<tt>rb_str_cat2(VALUE str, const char* ptr)</tt> ::
<tt>rb_str_cat_cstr(VALUE str, const char* ptr)</tt> ::
Appends C string ptr to Ruby string str. This function is
equivalent <tt>to rb_str_cat(str, ptr, strlen(ptr))</tt>.
<tt>rb_str_catf(VALUE str, const char* format, ...)</tt> ::
<tt>rb_str_vcatf(VALUE str, const char* format, va_list ap)</tt> ::
Appends C string format and successive arguments to Ruby string
str according to a printf-like format. These functions are
equivalent to <tt>rb_str_append(str, rb_sprintf(format, ...))</tt> and
<tt>rb_str_append(str, rb_vsprintf(format, ap))</tt>, respectively.
<tt>rb_enc_str_new(const char *ptr, rb_len_t len, rb_encoding *enc)</tt> ::
<tt>rb_enc_str_new_cstr(const char *ptr, rb_encoding *enc)</tt> ::
Creates a new Ruby string with the specified encoding.
<tt>rb_enc_str_new_literal(const char *ptr, rb_encoding *enc)</tt> ::
Creates a new Ruby string from a C string literal with the specified
encoding.
<tt>rb_usascii_str_new(const char *ptr, rb_len_t len)</tt> ::
<tt>rb_usascii_str_new_cstr(const char *ptr)</tt> ::
Creates a new Ruby string with encoding US-ASCII.
<tt>rb_usascii_str_new_literal(const char *ptr)</tt> ::
Creates a new Ruby string from a C string literal with encoding
US-ASCII.
<tt>rb_utf8_str_new(const char *ptr, rb_len_t len)</tt> ::
<tt>rb_utf8_str_new_cstr(const char *ptr)</tt> ::
Creates a new Ruby string with encoding UTF-8.
<tt>rb_utf8_str_new_literal(const char *ptr)</tt> ::
Creates a new Ruby string from a C string literal with encoding
UTF-8.
<tt>rb_str_resize(VALUE str, rb_len_t len)</tt> ::
Resizes a Ruby string to len bytes. If str is not modifiable, this
function raises an exception. The length of str must be set in
advance. If len is less than the old length the content beyond
len bytes is discarded, else if len is greater than the old length
the content beyond the old length bytes will not be preserved but
will be garbage. Note that <tt>RSTRING_PTR(str)</tt> may change by calling
this function.
<tt>rb_str_set_len(VALUE str, rb_len_t len)</tt> ::
Sets the length of a Ruby string. If str is not modifiable, this
function raises an exception. This function preserves the content
up to len bytes, regardless <tt>RSTRING_LEN(str)</tt>. len must not exceed
the capacity of str.
<tt>rb_str_modify(VALUE str)</tt> ::
Prepares a Ruby string to modify. If str is not modifiable, this
function raises an exception, or if the buffer of str is shared,
this function allocates new buffer to make it unshared. Always
you MUST call this function before modifying the contents using
+RSTRING_PTR+ and/or +rb_str_set_len+.
==== Array functions
<tt>rb_ary_new()</tt> ::
Creates an array with no elements.
<tt>rb_ary_new2(rb_len_t len)</tt> ::
<tt>rb_ary_new_capa(rb_len_t len)</tt> ::
Creates an array with no elements, allocating internal buffer
for len elements.
<tt>rb_ary_new3(rb_len_t n, ...)</tt> ::
<tt>rb_ary_new_from_args(rb_len_t n, ...)</tt> ::
Creates an n-element array from the arguments.
<tt>rb_ary_new4(rb_len_t n, VALUE *elts)</tt> ::
<tt>rb_ary_new_from_values(rb_len_t n, VALUE *elts)</tt> ::
Creates an n-element array from a C array.
<tt>rb_ary_to_ary(VALUE obj)</tt> ::
Converts the object into an array.
Equivalent to Object#to_ary.
There are many functions to operate an array. They may dump core if other
types are given.
<tt>rb_ary_aref(int argc, const VALUE *argv, VALUE ary)</tt> ::
Equivalent to Array#[].
<tt>rb_ary_entry(VALUE ary, rb_len_t offset)</tt> ::
<tt>ary[offset]</tt>
<tt>rb_ary_store(VALUE ary, rb_len_t offset, VALUE obj)</tt> ::
<tt>ary[offset] = obj</tt>
<tt>rb_ary_subseq(VALUE ary, rb_len_t beg, rb_len_t len)</tt> ::
<tt>ary[beg, len]</tt>
<tt>rb_ary_push(VALUE ary, VALUE val)</tt> ::
<tt>rb_ary_pop(VALUE ary)</tt> ::
<tt>rb_ary_shift(VALUE ary)</tt> ::
<tt>rb_ary_unshift(VALUE ary, VALUE val)</tt> ::
<tt>ary.push(val)</tt>, <tt>ary.pop</tt>, <tt>ary.shift</tt>, <tt>ary.unshift(val)</tt>
<tt>rb_ary_cat(VALUE ary, const VALUE *ptr, rb_len_t len)</tt> ::
Appends _len_ elements of objects from _ptr_ to the array.
== Extending Ruby with C
=== Adding new features to Ruby
You can add new features (classes, methods, etc.) to the Ruby
interpreter. Ruby provides APIs for defining the following things:
- Classes, Modules
- Methods, singleton methods
- Constants
==== Class and Module Definition
To define a class or module, use the functions below:
VALUE rb_define_class(const char *name, VALUE super)
VALUE rb_define_module(const char *name)
These functions return the newly created class or module. You may
want to save this reference into a variable to use later.
To define nested classes or modules, use the functions below:
VALUE rb_define_class_under(VALUE outer, const char *name, VALUE super)
VALUE rb_define_module_under(VALUE outer, const char *name)
==== Method and singleton method definition
To define methods or singleton methods, use these functions:
void rb_define_method(VALUE klass, const char *name,
VALUE (*func)(ANYARGS), int argc)
void rb_define_singleton_method(VALUE object, const char *name,
VALUE (*func)(ANYARGS), int argc)
The _argc_ represents the number of the arguments to the C function,
which must be less than 17. But I doubt you'll need that many.
If _argc_ is negative, it specifies the calling sequence, not number of
the arguments.
If _argc_ is -1, the function will be called as:
VALUE func(int argc, VALUE *argv, VALUE obj)
where _argc_ is the actual number of arguments, _argv_ is the C array of
the arguments, and _obj_ is the receiver.
If _argc_ is -2, the arguments are passed in a Ruby array. The function
will be called like:
VALUE func(VALUE obj, VALUE args)
where _obj_ is the receiver, and _args_ is the Ruby array containing
actual arguments.
There are some more functions to define methods. One takes an ID
as the name of method to be defined. See also ID or Symbol below.
void rb_define_method_id(VALUE klass, ID name,
VALUE (*func)(ANYARGS), int argc)
There are two functions to define private/protected methods:
void rb_define_private_method(VALUE klass, const char *name,
VALUE (*func)(ANYARGS), int argc)
void rb_define_protected_method(VALUE klass, const char *name,
VALUE (*func)(ANYARGS), int argc)
At last, +rb_define_module_function+ defines a module function,
which are private AND singleton methods of the module.
For example, +sqrt+ is a module function defined in the Math module.
It can be called in the following way:
Math.sqrt(4)
or
include Math
sqrt(4)
To define module functions, use:
void rb_define_module_function(VALUE module, const char *name,
VALUE (*func)(ANYARGS), int argc)
In addition, function-like methods, which are private methods defined
in the Kernel module, can be defined using:
void rb_define_global_function(const char *name, VALUE (*func)(ANYARGS), int argc)
To define an alias for the method,
void rb_define_alias(VALUE module, const char* new, const char* old);
To define a reader/writer for an attribute,
void rb_define_attr(VALUE klass, const char *name, int read, int write)
To define and undefine the +allocate+ class method,
void rb_define_alloc_func(VALUE klass, VALUE (*func)(VALUE klass));
void rb_undef_alloc_func(VALUE klass);
_func_ has to take the _klass_ as the argument and return a newly
allocated instance. This instance should be as empty as possible,
without any expensive (including external) resources.
If you are overriding an existing method of any ancestor of your class,
you may rely on:
VALUE rb_call_super(int argc, const VALUE *argv)
To specify whether keyword arguments are passed when calling super:
VALUE rb_call_super_kw(int argc, const VALUE *argv, int kw_splat)
_kw_splat_ can have these possible values (used by all methods that accept
_kw_splat_ argument):
+RB_NO_KEYWORDS+ :: Do not pass keywords
+RB_PASS_KEYWORDS+ :: Pass keywords, final argument should be a hash of keywords
+RB_PASS_CALLED_KEYWORDS+ :: Pass keywords if current method was called with
keywords, useful for argument delegation
To achieve the receiver of the current scope (if no other way is
available), you can use:
VALUE rb_current_receiver(void)
==== Constant definition
We have 2 functions to define constants:
void rb_define_const(VALUE klass, const char *name, VALUE val)
void rb_define_global_const(const char *name, VALUE val)
The former is to define a constant under specified class/module. The
latter is to define a global constant.
=== Use Ruby features from C
There are several ways to invoke Ruby's features from C code.
==== Evaluate Ruby programs in a string
The easiest way to use Ruby's functionality from a C program is to
evaluate the string as Ruby program. This function will do the job:
VALUE rb_eval_string(const char *str)
Evaluation is done under the current context, thus current local variables
of the innermost method (which is defined by Ruby) can be accessed.
Note that the evaluation can raise an exception. There is a safer
function:
VALUE rb_eval_string_protect(const char *str, int *state)
It returns +nil+ when an error occurred. Moreover, <code>*state</code> is
zero if _str_ was successfully evaluated, or nonzero otherwise.
==== ID or Symbol
You can invoke methods directly, without parsing the string. First I
need to explain about ID. ID is the integer number to represent
Ruby's identifiers such as variable names. The Ruby data type
corresponding to ID is Symbol. It can be accessed from Ruby in the
form:
:Identifier
or
:"any kind of string"
You can get the ID value from a string within C code by using
rb_intern(const char *name)
rb_intern_str(VALUE name)
You can retrieve ID from Ruby object (Symbol or String) given as an
argument by using
rb_to_id(VALUE symbol)
rb_check_id(volatile VALUE *name)
rb_check_id_cstr(const char *name, rb_len_t len, rb_encoding *enc)
These functions try to convert the argument to a String if it was not a
Symbol nor a String. The second function stores the converted result into
<code>*name</code>, and returns 0 if the string is not a known symbol.
After this function returned a non-zero value, <code>*name</code> is
always a Symbol or a String, otherwise it is a String if the result is
0. The third function takes C string (pointer to +char+ array) and its
length, not Ruby VALUE.
You can retrieve Symbol from Ruby object (Symbol or String) given as
an argument by using
rb_to_symbol(VALUE name)
rb_check_symbol(volatile VALUE *namep)
rb_check_symbol_cstr(const char *ptr, rb_len_t len, rb_encoding *enc)
These functions are similar to above functions except that these
return a Symbol instead of an ID.
You can convert C ID to Ruby Symbol by using
VALUE ID2SYM(ID id)
and to convert Ruby Symbol object to ID, use
ID SYM2ID(VALUE symbol)
==== Invoke Ruby method from C
To invoke methods directly, you can use the function below
VALUE rb_funcall(VALUE recv, ID mid, int argc, ...)
This function invokes a method on the _recv_, with the method name
specified by the symbol _mid_.
==== Accessing the variables and constants
You can access class variables and instance variables using access
functions. Also, global variables can be shared between both
environments. There's no way to access Ruby's local variables.
The functions to access/modify instance variables are below:
VALUE rb_ivar_get(VALUE obj, ID id)
VALUE rb_ivar_set(VALUE obj, ID id, VALUE val)
_id_ must be the symbol, which can be retrieved by <tt>rb_intern()</tt>.
To access the constants of the class/module:
VALUE rb_const_get(VALUE obj, ID id)
See also Constant Definition above.
== Information sharing between Ruby and C
=== Ruby constants that can be accessed from C
As stated in section 1.3,
the following Ruby constants can be referred from C.
<tt>Qtrue</tt> ::
<tt>Qfalse</tt> ::
Boolean values. +Qfalse+ is +false+ in C also (i.e. 0).
<tt>Qnil</tt> ::
Ruby +nil+ in C scope.
=== Global variables shared between C and Ruby
Information can be shared between the two environments using shared global
variables. To define them, you can use functions listed below:
void rb_define_variable(const char *name, VALUE *var)
This function defines the variable which is shared by both environments.
The value of the global variable pointed to by _var_ can be accessed
through Ruby's global variable named _name_.
You can define read-only (from Ruby, of course) variables using the
function below.
void rb_define_readonly_variable(const char *name, VALUE *var)
You can define hooked variables. The accessor functions (_getter_ and
_setter_) are called on access to the hooked variables.
void rb_define_hooked_variable(const char *name, VALUE *var,
VALUE (*getter)(), void (*setter)())
If you need to supply either setter or getter, just supply 0 for the
hook you don't need. If both hooks are 0, <tt>rb_define_hooked_variable()</tt>
works just like <tt>rb_define_variable()</tt>.
The prototypes of the _getter_ and _setter_ functions are as follows:
VALUE (*getter)(ID id, VALUE *var);
void (*setter)(VALUE val, ID id, VALUE *var);
Also you can define a Ruby global variable without a corresponding C
variable. The value of the variable will be set/get only by hooks.
void rb_define_virtual_variable(const char *name,
VALUE (*getter)(), void (*setter)())
The prototypes of the _getter_ and _setter_ functions are as follows:
VALUE (*getter)(ID id);
void (*setter)(VALUE val, ID id);
=== Encapsulate C data into a Ruby object
Sometimes you need to expose your struct in the C world as a Ruby
object.
In a situation like this, making use of the TypedData_XXX macro
family, the pointer to the struct and the Ruby object can be mutually
converted.
--
The old (non-Typed) Data_XXX macro family has been deprecated.
In the future version of Ruby, it is possible old macros will not
work.
++
==== C struct to Ruby object
You can convert _sval_, a pointer to your struct, into a Ruby object with the
next macro.
TypedData_Wrap_Struct(klass, data_type, sval)
<tt>TypedData_Wrap_Struct()</tt> returns a created Ruby object as a
VALUE.
This macro cannot be used with data types that have the +RUBY_TYPED_EMBEDDABLE+
flag; use +TypedData_Make_Struct+ or +rb_data_typed_object_zalloc+ for such
types.
The _klass_ argument is the class for the object. The _klass_ should derive
from _rb_cObject_, and the allocator must be set by calling
+rb_define_alloc_func+ or +rb_undef_alloc_func+.
_data_type_ is a pointer to a const +rb_data_type_t+ which describes how Ruby
should manage the struct.
+rb_data_type_t+ is roughly defined like this. Fields reserved for future use
are omitted.
typedef struct rb_data_type_struct rb_data_type_t;
struct rb_data_type_struct {
const char *wrap_struct_name;
struct {
void (*dmark)(void*);
void (*dfree)(void*);
size_t (*dsize)(const void *);
void (*dcompact)(void*);
/* reserved fields */
} function;
const rb_data_type_t *parent;
void *data;
VALUE flags;
};
+wrap_struct_name+ is an identifier of this instance of the struct. It is
basically used for collecting and emitting statistics. So the identifier should
be unique in the process, but doesn't need to be valid as a C or Ruby
identifier.
These +dmark+ / +dfree+ functions are invoked during GC execution. No object
allocations are allowed during it, so do not allocate ruby objects inside them.
+dmark+ is a function to mark Ruby objects referred from your struct. It must
mark all references from your struct with +rb_gc_mark+ or its family if your
struct keeps such references.
--
Note that it is recommended to avoid such a reference.
++
+dfree+ is a function to free the pointer allocation. If this is
+RUBY_DEFAULT_FREE+, the pointer will be just freed.
+dsize+ calculates memory consumption in bytes by the struct. Its parameter is
a pointer to your struct. You can pass 0 as dsize if it is hard to implement
such a function. But it is still recommended to avoid 0.
+dcompact+ is invoked when memory compaction took place. Referred Ruby objects
that were marked by <tt>rb_gc_mark_movable()</tt> can here be updated per
<tt>rb_gc_location()</tt>.
+parent+ can point to another C type definition that the Ruby object is
inherited from. Then <tt>TypedData_Get_Struct()</tt> does also accept
derived objects.
You can fill +data+ with an arbitrary value for your use. Ruby does nothing
with the member.
+flags+ is a bitwise-OR of the following flag values. Since they require deep
understanding of garbage collector in Ruby, you can just set 0 to +flags+ if you
are not sure.
+RUBY_TYPED_FREE_IMMEDIATELY+ ::
This flag allows the garbage collector to immediately invoke <tt>dfree()</tt>
during GC when it needs to free your struct.
You can specify this flag if the <tt>dfree</tt> never unlocks Ruby's
internal lock (GVL).
If this flag is not set, Ruby defers invocation of <tt>dfree()</tt>
and invokes <tt>dfree()</tt> at the same time as finalizers.
+RUBY_TYPED_THREAD_SAFE_FREE+ ::
This flag declares that the <tt>dfree()</tt> function is thread-safe.
With this flag set the garbage collector MAY invoke <tt>dfree()</tt>
* On any thread, whether a Ruby Thread or a native thread
* In parallel with calls to the same or other <tt>dfree</tt> functions
* Concurrently with other Ruby code
This should NOT be set on <tt>dfree()</tt> functions which access shared state
in a thread-unsafe way. A <tt>dfree()</tt> function which uses only
<tt>xfree()</tt> on the struct and memory it exclusively owns is thread-safe.
This flag implies +RUBY_TYPED_FREE_IMMEDIATELY+, since a thread-safe
<tt>dfree()</tt> is also safe to invoke immediately.
+RUBY_TYPED_WB_PROTECTED+ ::
It shows that implementation of the object supports write barriers.
If this flag is set, Ruby is better able to do garbage collection
of the object.
When it is set, however, you are responsible for putting write
barriers in all implementations of methods of that object as
appropriate. Otherwise Ruby might crash while running.
More about write barriers can be found in {Generational
GC}[rdoc-ref:@Appendix+D.+Generational+GC].
+RUBY_TYPED_FROZEN_SHAREABLE+ ::
This flag indicates that the object is shareable object if the object
is frozen. See {Ractor support}[rdoc-ref:@Appendix+F.+Ractor+support]
more details.
If this flag is not set, the object can not become a shareable
object by <tt>Ractor.make_shareable()</tt> method.
+RUBY_TYPED_EMBEDDABLE+ ::
This flag indicates that Ruby may store the C struct inside the object
slot, rather than allocate it separately with +malloc+.
However, it is not a guarantee. Ruby may decide not to embed the object.
For instance if it's too large to fit into one of the available slot sizes.
Embedding the C struct inside the object slot reduces pointer chasing,
+malloc+ overhead, and improves sweep performance. In some cases, it can also
reduce the memory footprint of the object.
To be embeddable, types must abide by some restrictions:
* Pointers to the C struct, or into the C struct, MUST NOT be stored,
as they become invalid when GC compaction occurs.
It is however valid to pass and use such pointers for as long as the Ruby
object remains on the stack.
In a sense, this is similar to the restrictions of a stack allocated struct.
The +RB_GC_GUARD+ macro must be used to ensure the object is not moved by
compaction and not freed, unless the object is passed directly as an
argument from Ruby to C, i.e. as a parameter of a function used with
+rb_define_method+ and similar.
* The +DATA_PTR+, +RTYPEDDATA_DATA+ and +TypedData_Wrap_Struct+ macro can't be
used.
Only +RTYPEDDATA_GET_DATA+ or +TypedData_Get_Struct+ macros can be used with
embeddable objects. Accessing <tt>RTYPEDDATA(obj)->data</tt> is invalid too.
* The +dfree+ function MUST NOT free the C struct itself.
Setting +dfree+ to +RUBY_DEFAULT_FREE+ is fine.
To support older Ruby versions without this feature, you can
conditionally free the C struct if +RUBY_TYPED_EMBEDDABLE+ isn't defined.
* The type must have the +RUBY_TYPED_FREE_IMMEDIATELY+ flag set.
If the embedded C struct is of variable size, +rb_data_typed_object_zalloc+
can be used instead of +TypedData_Make_Struct+.
See {Embedded TypedData}[rdoc-ref:@Appendix+G.+Embedded+TypedData] for a
commented example of how to use +RUBY_TYPED_EMBEDDABLE+.
Fields other than those listed above are reserved for future use, and
initializers are typically omitted. In the rare case that you need to
dynamically allocate a +rb_data_type_t+, these fields must be filled with
zeros.
Note that this macro can raise an exception. If _sval_ already holds a
resource that must be released, such as allocated memory or a handle from an
external library, protect the call with +rb_protect+ and release the resource
when an exception is raised. Alternatively, create the object first with a
+NULL+ data pointer, then allocate the resource and store it in the object.
You can allocate and wrap the structure in one step, in more
preferable manner.
obj = TypedData_Make_Struct(klass, type, data_type, sval)
This macro returns an allocated +T_DATA+ object, wrapping the pointer to the
structure, which is also allocated. This macro works like:
obj = TypedData_Wrap_Struct(klass, data_type, NULL);
RTYPEDDATA_DATA(obj) = sval = ZALLOC(type);
However, you should use this macro instead of "allocation then wrap" like the
above code if it is simply allocated, because the "wrap" can raise a
+NoMemoryError+ and _sval_ will be memory leaked in that case.
Arguments _klass_ and _data_type_ work like their counterparts in
<tt>TypedData_Wrap_Struct()</tt>. A pointer to the allocated structure will be
assigned to _sval_, which should be a pointer of the type specified.
Same for other than simple memory allocation:
obj = TypedData_Wrap_Struct(klass, data_type, NULL);
sval = allocate_some_resource();
if (success(sval)) RTYPEDDATA_DATA(obj) = sval;
You may have to cast _sval_ to <tt>void*</tt>, if it is not a pointer type.
==== Declaratively marking/compacting struct references
In the case where your struct refers to Ruby objects that are simple values,
not wrapped in conditional logic or complex data structures an alternative
approach to marking and reference updating is provided, by declaring offset
references to the VALUEs (or edge) in your struct.
Doing this allows the Ruby GC to support marking these references and GC
compaction without the need to define the +dmark+ and +dcompact+ callbacks.
You must define a static list of VALUE pointers to the offsets within your
struct where the references are located, and set the +data+ member to point
to this reference list. The reference list must end with +RUBY_END_REFS+.
Some macros have been provided to make edge referencing easier:
<tt>RUBY_TYPED_DECL_MARKING</tt> ::
A flag that can be set on the +ruby_data_type_t+ to indicate that
references are being declared as edges.
<tt>RUBY_REFERENCES(ref_list_name)</tt> ::
Define _ref_list_name_ as a list of references
<tt>RUBY_REF_END</tt> :: The end mark of the references list.
<tt>RUBY_REF_EDGE(struct, member)</tt> ::
Declare _member_ as a VALUE edge from _struct_. Use this inside
the block of +RUBY_REFERENCES+.
+RUBY_REFS_LIST_PTR+ ::
Coerce the reference list into a format that can be accepted by the
existing +dmark+ interface.
The example below is from Dir (defined in +dir.c+)
// The struct being wrapped. Notice this contains 3 members of which the second
// is a VALUE reference to another ruby object.
struct dir_data {
DIR *dir;
const VALUE path;
rb_encoding *enc;
}
// Define a reference list `dir_refs` containing a single entry to `path`.
// Needs terminating with RUBY_REF_END
RUBY_REFERENCES(dir_refs) = {
RUBY_REF_EDGE(dir_data, path),
RUBY_REF_END
};
// Override the "dmark" field with the defined reference list now that we
// no longer need a marking callback and add RUBY_TYPED_DECL_MARKING to the
// flags field
static const rb_data_type_t dir_data_type = {
"dir",
{RUBY_REFS_LIST_PTR(dir_refs), dir_free, dir_memsize,},
0, NULL, RUBY_TYPED_WB_PROTECTED | RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_DECL_MARKING
};
Declaring simple references declaratively in this manner allows the GC to both
mark, and move the underlying object, and automatically update the reference to
it during compaction.
==== Ruby object to C struct
To retrieve the C pointer from the +T_DATA+ object, use the macro
<tt>TypedData_Get_Struct()</tt>.
TypedData_Get_Struct(obj, type, &data_type, sval)
A pointer to the structure will be assigned to the variable _sval_.
See the example below for details.
== Example - Creating the dbm Extension
OK, here's the example of making an extension library. This is the
extension to access DBMs. The full source is included in the ext/
directory in the Ruby's source tree.
=== Make the directory
% mkdir ext/dbm
Make a directory for the extension library under ext directory if you
want to link it to statically Ruby. You can use an arbitrary name.
=== Design the Library
You need to design the library features, before making it.
=== Write the C Code
You need to write C code for your extension library. If your library
has only one source file, choosing ``LIBRARY.c'' as a file name is
preferred.
Note that some functions in mkmf library described below generate
a file ``conftest.c'' for checking with compilation. You shouldn't
choose ``conftest.c'' as a name of a source file.
Ruby will execute the initializing function named +Init_LIBRARY+ in
the library. For example, <code>Init_dbm()</code> will be executed when
loading the library.
Here's the example of an initializing function.
#include <ruby.h>
void
Init_dbm(void)
{
/* define DBM class */
VALUE cDBM = rb_define_class("DBM", rb_cObject);
/* Redefine DBM.allocate
rb_define_alloc_func(cDBM, fdbm_alloc);
/* DBM includes Enumerable module */
rb_include_module(cDBM, rb_mEnumerable);
/* DBM has class method open(): arguments are received as C array */
rb_define_singleton_method(cDBM, "open", fdbm_s_open, -1);
/* DBM instance method close(): no args */
rb_define_method(cDBM, "close", fdbm_close, 0);
/* DBM instance method []: 1 argument */
rb_define_method(cDBM, "[]", fdbm_aref, 1);
/* ... */
/* ID for a instance variable to store DBM data */
id_dbm = rb_intern("dbm");
}
The dbm extension wraps the dbm struct in the C environment using
+TypedData_Make_Struct+.
struct dbmdata {
int di_size;
DBM *di_dbm;
};
static const rb_data_type_t dbm_type = {
"dbm",
{0, free_dbm, memsize_dbm,},
0, 0,
RUBY_TYPED_FREE_IMMEDIATELY,
};
static VALUE
fdbm_alloc(VALUE klass)
{
struct dbmdata *dbmp;
/* Allocate T_DATA object and C struct and fill struct with zero bytes */
return TypedData_Make_Struct(klass, struct dbmdata, &dbm_type, dbmp);
}
This code wraps the +dbmdata+ structure into a Ruby object. We avoid
wrapping <tt>DBM*</tt> directly, because we want to cache size information.
Since Class.allocate allocates an ordinary +T_OBJECT+ type (instead
of +T_DATA+), it's important to either use <tt>rb_define_alloc_func()</tt> to
overwrite it or <tt>rb_undef_alloc_func()</tt> to delete it.
To retrieve the +dbmdata+ structure from a Ruby object, we define the
following macro:
#define GetDBM(obj, dbmp) do {\
TypedData_Get_Struct((obj), struct dbmdata, &dbm_type, (dbmp));\
if ((dbmp) == 0) closed_dbm();\
if ((dbmp)->di_dbm == 0) closed_dbm();\
} while (0)
This sort of complicated macro does the retrieving and close checking
for the DBM.
There are three kinds of way to receive method arguments. First,
methods with a fixed number of arguments receive arguments like this:
static VALUE
fdbm_aref(VALUE obj, VALUE keystr)
{
struct dbmdata *dbmp;
GetDBM(obj, dbmp);
/* Use dbmp to access the key */
dbm_fetch(dbmp->di_dbm, StringValueCStr(keystr));
/* ... */
}
The first argument of the C function is the +self+, the rest are the
arguments to the method.
Second, methods with an arbitrary number of arguments receive
arguments like this:
static VALUE
fdbm_s_open(int argc, VALUE *argv, VALUE klass)
{
/* ... */
if (rb_scan_args(argc, argv, "11", &file, &vmode) == 1) {
mode = 0666; /* default value */
}
/* ... */
}
The first argument is the number of method arguments, the second
argument is the C array of the method arguments, and the third
argument is the receiver of the method.
You can use the function <tt>rb_scan_args()</tt> to check and retrieve
the arguments. The third argument is a string that specifies how to
capture method arguments and assign them to the following VALUE
references.
You can just check the argument number with <tt>rb_check_arity()</tt>,
this is handy in the case you want to treat the arguments as a list.
The following is an example of a method that takes arguments by Ruby's
array:
static VALUE
thread_initialize(VALUE thread, VALUE args)
{
/* ... */
}
The first argument is the receiver, the second one is the Ruby array
which contains the arguments to the method.
<b>Notice</b>: GC should know about global variables which refer to Ruby's objects,
but are not exported to the Ruby world. You need to protect them by
void rb_global_variable(VALUE *var)
or the objects themselves by
void rb_gc_register_mark_object(VALUE object)
Registration is scoped to the calling Ractor. Any Ractor may register an
address. The stored value must be a special constant, a shareable object,
or an unshareable object owned by the registering Ractor. If the address
holds another Ractor's unshareable object, the owning Ractor's GC cannot
see the registration and can free the stored object while the registered
address still refers to it. This can cause a use-after-free if the
registering Ractor attempts to use the Object stored.
If the address has process lifetime (a +static VALUE+), either register it
from the main Ractor or keep the stored values shareable.
Calling +Ractor#value+ on a Ractor moves its registrations to the calling
Ractor; otherwise they move to the main Ractor once the dead Ractor is
collected. If you rely on a Ractor inheriting another Ractor's registered
globals, call +Ractor#value+.
=== Prepare extconf.rb
If the file named extconf.rb exists, it will be executed to generate
Makefile.
extconf.rb is the file for checking compilation conditions etc. You
need to put
require 'mkmf'
at the top of the file. You can use the functions below to check
various conditions.
append_cppflags(array-of-flags[, opt]): append each flag to $CPPFLAGS if usable
append_cflags(array-of-flags[, opt]): append each flag to $CFLAGS if usable
append_ldflags(array-of-flags[, opt]): append each flag to $LDFLAGS if usable
have_macro(macro[, headers[, opt]]): check whether macro is defined
have_library(lib[, func[, headers[, opt]]]): check whether library containing function exists
find_library(lib[, func, *paths]): find library from paths
have_func(func[, headers[, opt]): check whether function exists
have_var(var[, headers[, opt]]): check whether variable exists
have_header(header[, preheaders[, opt]]): check whether header file exists
find_header(header, *paths): find header from paths
have_framework(fw): check whether framework exists (for MacOS X)
have_struct_member(type, member[, headers[, opt]]): check whether struct has member
have_type(type[, headers[, opt]]): check whether type exists
find_type(type, opt, *headers): check whether type exists in headers
have_const(const[, headers[, opt]]): check whether constant is defined
check_sizeof(type[, headers[, opts]]): check size of type
check_signedness(type[, headers[, opts]]): check signedness of type
convertible_int(type[, headers[, opts]]): find convertible integer type
find_executable(bin[, path]): find executable file path
create_header(header): generate configured header
create_makefile(target[, target_prefix]): generate Makefile
See MakeMakefile for full documentation of these functions.
The value of the variables below will affect the Makefile.
$CFLAGS: included in CFLAGS make variable (such as -O)
$CPPFLAGS: included in CPPFLAGS make variable (such as -I, -D)
$LDFLAGS: included in LDFLAGS make variable (such as -L)
$objs: list of object file names
Compiler/linker flags are not portable usually, you should use
+append_cppflags+, +append_cflags+ and +append_ldflags+ respectively
instead of appending the above variables directly.
Normally, the object files list is automatically generated by searching
source files, but you must define them explicitly if any sources will
be generated while building.
If a compilation condition is not fulfilled, you should not call
{create_makefile}[rdoc-ref:MakeMakefile#create_makefile]. The Makefile will not be generated,
compilation will not be done.
=== Prepare depend (Optional)
If the file named depend exists, Makefile will include that file to
check dependencies. You can make this file by invoking
% gcc -MM *.c > depend
It's harmless. Prepare it.
=== Generate Makefile
Try generating the Makefile by:
ruby extconf.rb
If the library should be installed under vendor_ruby directory
instead of site_ruby directory, use --vendor option as follows.
ruby extconf.rb --vendor
You don't need this step if you put the extension library under the ext
directory of the ruby source tree. In that case, compilation of the
interpreter will do this step for you.
=== Run make
Type
make
to compile your extension. You don't need this step either if you have
put the extension library under the ext directory of the ruby source tree.
=== Debug
You may need to rb_debug the extension. Extensions can be linked
statically by adding the directory name in the ext/Setup file so that
you can inspect the extension with the debugger.
=== Done! Now you have the extension library
You can do anything you want with your library. The author of Ruby
will not claim any restrictions on your code depending on the Ruby API.
Feel free to use, modify, distribute or sell your program.
== Appendix A. Ruby header and source files overview
=== Ruby header files
Everything under <tt>$repo_root/include/ruby</tt> is installed with
<tt>make install</tt>.
It should be included per <tt>#include <ruby.h></tt> from C extensions.
All symbols are public API with the exception of symbols prefixed with
+rbimpl_+ or +RBIMPL_+. They are implementation details and shouldn't
be used by C extensions.
Only <tt>$repo_root/include/ruby/*.h</tt> whose corresponding macros
are defined in the <tt>$repo_root/include/ruby.h</tt> header are
allowed to be <tt>#include</tt>-d by C extensions.
Header files under <tt>$repo_root/internal/</tt> or directly under the
root <tt>$repo_root/*.h</tt> are not make-installed.
They are internal headers with only internal APIs.
=== Ruby language core
class.c :: classes and modules
error.c :: exception classes and exception mechanism
gc.c :: memory management
load.c :: library loading
object.c :: objects
variable.c :: variables and constants
=== Ruby syntax parser
parse.y :: grammar definition
parse.c :: automatically generated from parse.y
defs/keywords :: reserved keywords
lex.c :: automatically generated from keywords
id.c :: pre-defined IDs
=== Ruby evaluator (a.k.a. YARV)
eval.c ::
eval_error.c ::
eval_jump.c ::
Evaluator
compile.c :: AST to VM instruction compiler
iseq.c :: implementation of VM::ISeq
thread.c ::
thread_win32.c ::
thread_pthread.c ::
thread management and context switching
vm.c ::
vm_dump.c ::
vm_eval.c ::
vm_exec.c ::
vm_insnhelper.c ::
vm_method.c ::
VM implementation
insns.def :: definition of VM instructions
defs/opt_insns_unif.def :: instruction unification
defs/opt_operand.def :: definitions for optimization
-> insn*.inc ::
-> opt*.inc ::
-> vm.inc ::
automatically generated
=== Regular expression engine (Onigumo)
regcomp.c :: compiler
regenc.c :: encodings
regerror.c :: error handling
regexec.c :: executer
regparse.c :: parser
regsyntax.c :: pre-defined syntax options
=== Utility functions
debug.c :: debug symbols for C debugger
dln.c :: dynamic loading
st.c :: general purpose hash table
strftime.c :: formatting times
util.c :: misc utilities
=== Ruby interpreter implementation
dmyext.c ::
dmydln.c ::
dmyencoding.c ::
for miniruby
id.c ::
inits.c ::
main.c ::
ruby.c ::
version.c ::
Ruby command
builtijns.c ::
gem_prelude.rb ::
prelude.rb ::
for Ruby code loaded at start up
=== Class library
array.c :: Array
bignum.c :: Integer (Bignum)
compar.c :: Comparable
complex.c :: Complex
cont.c :: Fiber, Continuation
dir.c :: Dir
enum.c :: Enumerable
enumerator.c :: Enumerator
file.c :: File
hash.c :: Hash
io.c :: IO
marshal.c :: Marshal
math.c :: Math
numeric.c :: Numeric, Integer (Fixnum), Float
pack.c :: Array#pack, String#unpack
proc.c :: Binding, Proc
process.c :: Process
random.c :: Random number generator
range.c :: Range
rational.c :: Rational
re.c :: Regexp, MatchData
signal.c :: Signal
sprintf.c :: String#sprintf
string.c :: String
struct.c :: Struct
symbol.c :: Symbol
time.c :: Time
defs/known_errors.def :: exception classes under Errno
-> known_errors.inc :: automatically generated
=== Multilingualization
encoding.c :: Encoding
transcode.c :: Encoding::Converter
enc/*.c :: encoding classes
enc/trans/* :: codepoint mapping tables
=== goruby interpreter implementation
goruby.c ::
golf_prelude.rb :: goruby specific libraries.
-> golf_prelude.c :: automatically generated
== Appendix B. Ruby extension API reference
=== Types
+VALUE+ ::
The type for the Ruby object. Actual structures are defined in ruby.h,
such as <tt>struct RString</tt>, etc. To refer the values in structures, use
casting macros like <tt>RSTRING(obj)</tt>.
=== Variables and constants
+Qnil+ ::
+nil+ object
+Qtrue+ ::
+true+ object (default true value)
+Qfalse+ ::
+false+ object
=== Checking VALUE types
<tt>RB_TYPE_P(value, type)</tt> ::
Is _value_ an internal _type_ (+T_NIL+, +T_FIXNUM+, etc.)?
<tt>TYPE(value)</tt> ::
Internal type (+T_NIL+, +T_FIXNUM+, etc.)
<tt>FIXNUM_P(value)</tt> ::
Is _value_ a Fixnum?
<tt>NIL_P(value)</tt> ::
Is _value_ +nil+?
<tt>RB_INTEGER_TYPE_P(value)</tt> ::
Is _value_ an Integer?
<tt>RB_FLOAT_TYPE_P(value)</tt> ::
Is _value_ a Float?
<tt>void Check_Type(VALUE value, int type)</tt> ::
Ensures _value_ is of the given internal _type_ or raises a TypeError
=== VALUE type conversion
<tt>FIX2INT(value)</tt>, <tt>INT2FIX(i)</tt> ::
Fixnum <-> integer
<tt>FIX2LONG(value)</tt>, <tt>LONG2FIX(l)</tt> ::
Fixnum <-> long
<tt>NUM2INT(value)</tt>, <tt>INT2NUM(i)</tt> ::
Numeric <-> integer
<tt>NUM2UINT(value)</tt>, <tt>UINT2NUM(ui)</tt> ::
Numeric <-> unsigned integer
<tt>NUM2LONG(value)</tt>, <tt>LONG2NUM(l)</tt> ::
Numeric <-> long
<tt>NUM2ULONG(value)</tt>, <tt>ULONG2NUM(ul)</tt> ::
Numeric <-> unsigned long
<tt>NUM2LL(value)</tt>, <tt>LL2NUM(ll)</tt> ::
Numeric <-> long long
<tt>NUM2ULL(value)</tt>, <tt>ULL2NUM(ull)</tt> ::
Numeric <-> unsigned long long
<tt>NUM2OFFT(value)</tt>, <tt>OFFT2NUM(off)</tt> ::
Numeric <-> off_t
<tt>NUM2SIZET(value)</tt>, <tt>SIZET2NUM(size)</tt> ::
Numeric <-> size_t
<tt>NUM2SSIZET(value)</tt>, <tt>SSIZET2NUM(ssize)</tt> ::
Numeric <-> ssize_t
<tt>rb_integer_pack(value, words, numwords, wordsize, nails, flags)</tt>, <tt>rb_integer_unpack(words, numwords, wordsize, nails, flags)</tt> ::
Numeric <-> Arbitrary size integer buffer
<tt>NUM2DBL(value)</tt> ::
Numeric -> double
<tt>rb_float_new(f)</tt> ::
double -> Float
<tt>RSTRING_LEN(str)</tt> ::
String -> length of String data in bytes
<tt>RSTRING_PTR(str)</tt> ::
String -> pointer to String data
Note that the result pointer may not be NUL-terminated
<tt>StringValue(value)</tt> ::
Object with <tt>#to_str</tt> -> String
<tt>StringValuePtr(value)</tt> ::
Object with <tt>#to_str</tt> -> pointer to String data
<tt>StringValueCStr(value)</tt> ::
Object with <tt>#to_str</tt> -> pointer to String data without NUL bytes
It is guaranteed that the result data is NUL-terminated
<tt>rb_str_new2(s)</tt> ::
char * -> String
=== Defining classes and modules
<tt>VALUE rb_define_class(const char *name, VALUE super)</tt> ::
Defines a Ruby class as a subclass of _super_ with named specified by
_name_ at the top-level.
<tt>VALUE rb_define_class_under(VALUE module, const char *name, VALUE super)</tt> ::
Defines a Ruby class as a subclass of _super_ with named specified by
_name_ under the <i>module</i>'s namespace.
<tt>VALUE rb_define_module(const char *name)</tt> ::
Defines a Ruby module with named specified by _name_ at the top-level.
<tt>VALUE rb_define_module_under(VALUE module, const char *name)</tt> ::
Defines a Ruby module with named specified by _name_ under the
<i>module</i>'s namespace.
<tt>void rb_include_module(VALUE klass, VALUE module)</tt> ::
Includes _module_ into _class_. If _class_ already includes it, just
ignored.
<tt>void rb_extend_object(VALUE object, VALUE module)</tt> ::
Extend the _object_ with the <i>module</i>'s attributes.
=== Defining global variables
<tt>void rb_define_variable(const char *name, VALUE *var)</tt> ::
Defines a global variable _name_ which is shared between C and Ruby. If
_name_ does not start with +$+, it is prefixed automatically. If that
variable name contains a character which is not allowed a global
variable, that variable can't be seen from Ruby programs.
<tt>void rb_define_readonly_variable(const char *name, VALUE *var)</tt> ::
Defines a read-only global variable. Works just like
<tt>rb_define_variable()</tt>, except the defined variable is read-only.
<tt>void rb_define_virtual_variable(const char *name, VALUE (*getter)(), void (*setter)())</tt> ::
Defines a virtual variable, whose behavior is defined by a pair of C
functions. The _getter_ function is called when the variable is
referenced. The _setter_ function is called when the variable is set to a
value.
These functions are the same as +rb_define_hooked_variable+, but the
_var_ argument is meaningless.
<tt>void rb_define_hooked_variable(const char *name, VALUE *var, VALUE (*getter)(), void (*setter)())</tt> ::
Defines hooked variable. It's a virtual variable with a C variable.
The _getter_ is called as
VALUE getter(ID id, VALUE *var)
returning a new value. The _setter_ is called as
void setter(VALUE val, ID id, VALUE *var)
_id_ is the ID interned from _name_, and _var_ is the argument itself.
_val_ is the value to be set.
<tt>void rb_global_variable(VALUE *var)</tt> ::
Tells GC to protect C global variable, which holds Ruby value to be marked.
The registration belongs to the calling Ractor: the stored value must be
a special constant, a shareable object, or an unshareable object owned by
the registering Ractor.
<tt>void rb_gc_register_mark_object(VALUE object)</tt> ::
Tells GC to protect the _object_, which may not be referenced anywhere.
=== Constant definition
<tt>void rb_define_const(VALUE klass, const char *name, VALUE val)</tt> ::
Defines a new constant under the class/module.
<tt>void rb_define_global_const(const char *name, VALUE val)</tt> ::
Defines a global constant. This is just the same as
rb_define_const(rb_cObject, name, val)
=== \Method definition
<tt>rb_define_method(VALUE klass, const char *name, VALUE (*func)(ANYARGS), int argc)</tt> ::
Defines a method for the class. func is the function pointer. argc
is the number of arguments. if _argc_ is -1, the function will receive
3 arguments: _argc_, _argv_, and _self_. if _argc_ is -2, the function will
receive 2 arguments, _self_ and _args_, where _args_ is a Ruby array of
the method arguments.
<tt>rb_define_private_method(VALUE klass, const char *name, VALUE (*func)(ANYARGS), int argc)</tt> ::
Defines a private method for the class. Arguments are same as
<tt>rb_define_method()</tt>.
<tt>rb_define_singleton_method(VALUE klass, const char *name, VALUE (*func)(ANYARGS), int argc)</tt> ::
Defines a singleton method. Arguments are same as <tt>rb_define_method()</tt>.
<tt>rb_check_arity(int argc, int min, int max)</tt> ::
Check the number of arguments, _argc_ is in the range of +min..max+. If
_max_ is +UNLIMITED_ARGUMENTS+, upper bound is not checked. If _argc_ is
out of bounds, an ArgumentError will be raised.
<tt>rb_scan_args(int argc, VALUE *argv, const char *fmt, ...)</tt> ::
Retrieve argument from _argc_ and _argv_ to given VALUE references
according to the format string _fmt_. The format can be described in ABNF
as follows:
scan-arg-spec := param-arg-spec [keyword-arg-spec] [block-arg-spec]
param-arg-spec := pre-arg-spec [post-arg-spec] / post-arg-spec /
pre-opt-post-arg-spec
pre-arg-spec := num-of-leading-mandatory-args [num-of-optional-args]
post-arg-spec := sym-for-variable-length-args
[num-of-trailing-mandatory-args]
pre-opt-post-arg-spec := num-of-leading-mandatory-args num-of-optional-args
num-of-trailing-mandatory-args
keyword-arg-spec := sym-for-keyword-arg
block-arg-spec := sym-for-block-arg
num-of-leading-mandatory-args := DIGIT ; The number of leading
; mandatory arguments
num-of-optional-args := DIGIT ; The number of optional
; arguments
sym-for-variable-length-args := "*" ; Indicates that variable
; length arguments are
; captured as a ruby array
num-of-trailing-mandatory-args := DIGIT ; The number of trailing
; mandatory arguments
sym-for-keyword-arg := ":" ; Indicates that keyword
; argument captured as a hash.
; If keyword arguments are not
; provided, returns nil.
sym-for-block-arg := "&" ; Indicates that an iterator
; block should be captured if
; given
For example, "12" means that the method requires at least one
argument, and at most receives three (1+2) arguments. So, the format
string must be followed by three variable references, which are to be
assigned to captured arguments. For omitted arguments, variables are
set to +Qnil+. NULL can be put in place of a variable reference, which
means the corresponding captured argument(s) should be just dropped.
The number of given arguments, excluding an option hash or iterator
block, is returned.
<tt>rb_scan_args_kw(int kw_splat, int argc, VALUE *argv, const char *fmt, ...)</tt> ::
The same as +rb_scan_args+, except the _kw_splat_ argument specifies whether
keyword arguments are provided (instead of being determined by the call
from Ruby to the C function). _kw_splat_ should be one of the following
values:
+RB_SCAN_ARGS_PASS_CALLED_KEYWORDS+ :: Same behavior as +rb_scan_args+.
+RB_SCAN_ARGS_KEYWORDS+ :: The final argument should be a hash treated as
keywords.
+RB_SCAN_ARGS_LAST_HASH_KEYWORDS+ :: Treat a final argument as keywords if it
is a hash, and not as keywords otherwise.
<tt>int rb_get_kwargs(VALUE keyword_hash, const ID *table, int required, int optional, VALUE *values)</tt> ::
Retrieves argument VALUEs bound to keywords, which directed by _table_
into _values_, deleting retrieved entries from _keyword_hash_ along
the way. First _required_ number of IDs referred by _table_ are
mandatory, and succeeding _optional_ (<tt>- optional - 1</tt> if
_optional_ is negative) number of IDs are optional. If a
mandatory key is not contained in _keyword_hash_, raises "missing
keyword" ArgumentError. If an optional key is not present in
_keyword_hash_, the corresponding element in _values_ is set to +Qundef+.
If _optional_ is negative, rest of _keyword_hash_ are ignored, otherwise
raises "unknown keyword" ArgumentError.
Be warned, handling keyword arguments in the C API is less efficient
than handling them in Ruby. Consider using a Ruby wrapper method
around a non-keyword C function.
ref: https://bugs.ruby-lang.org/issues/11339
<tt>VALUE rb_extract_keywords(VALUE *original_hash)</tt> ::
Extracts pairs whose key is a symbol into a new hash from a hash
object referred by _original_hash_. If the original hash contains
non-symbol keys, then they are copied to another hash and the new hash
is stored through _original_hash_, else 0 is stored.
=== Invoking Ruby method
<tt>VALUE rb_funcall(VALUE recv, ID mid, int narg, ...)</tt> ::
Invokes a method specified by _mid_ on _recv_. To retrieve _mid_
from a method name, use <tt>rb_intern()</tt>. Able to call even
private/protected methods. Arguments after _narg_ are passe to the
method, and _narg_ should be the number of arguments.
<tt>VALUE rb_funcall2(VALUE recv, ID mid, int argc, VALUE *argv)</tt> ::
<tt>VALUE rb_funcallv(VALUE recv, ID mid, int argc, VALUE *argv)</tt> ::
Invokes a method, passing arguments as an array of values.
Able to call even private/protected methods.
<tt>VALUE rb_funcallv_kw(VALUE recv, ID mid, int argc, VALUE *argv, int kw_splat)</tt> ::
Same as rb_funcallv, using _kw_splat_ to determine whether keyword
arguments are passed.
<tt>VALUE rb_funcallv_public(VALUE recv, ID mid, int argc, VALUE *argv)</tt> ::
Invokes a method, passing arguments as an array of values.
Able to call only public methods.
<tt>VALUE rb_funcallv_public_kw(VALUE recv, ID mid, int argc, VALUE *argv, int kw_splat)</tt> ::
Same as +rb_funcallv_public+, using _kw_splat_ to determine whether keyword
arguments are passed.
<tt>VALUE rb_funcall_passing_block(VALUE recv, ID mid, int argc, const VALUE* argv)</tt> ::
Same as +rb_funcallv_public+, except is passes the currently active block as
the block when calling the method.
<tt>VALUE rb_funcall_passing_block_kw(VALUE recv, ID mid, int argc, const VALUE* argv, int kw_splat)</tt> ::
Same as +rb_funcall_passing_block+, using _kw_splat_ to determine whether
keyword arguments are passed.
<tt>VALUE rb_funcall_with_block(VALUE recv, ID mid, int argc, const VALUE *argv, VALUE passed_procval)</tt> ::
Same as +rb_funcallv_public+, except _passed_procval_ specifies the block to
pass to the method.
<tt>VALUE rb_funcall_with_block_kw(VALUE recv, ID mid, int argc, const VALUE *argv, VALUE passed_procval, int kw_splat)</tt> ::
Same as +rb_funcall_with_block+, using _kw_splat_ to determine whether
keyword arguments are passed.
<tt>VALUE rb_eval_string(const char *str)</tt> ::
Compiles and executes the string as a Ruby program.
<tt>ID rb_intern(const char *name)</tt> ::
Returns ID corresponding to the _name_.
<tt>char *rb_id2name(ID id)</tt> ::
Returns the name corresponding ID.
<tt>char *rb_class2name(VALUE klass)</tt> ::
Returns the name of the class.
<tt>int rb_respond_to(VALUE obj, ID id)</tt> ::
Returns true if the object responds to the message specified by _id_.
=== Instance variables
<tt>VALUE rb_iv_get(VALUE obj, const char *name)</tt> ::
Retrieve the value of the instance variable. If the name is not
prefixed by <code>@</code>, that variable shall be inaccessible from Ruby.
<tt>VALUE rb_iv_set(VALUE obj, const char *name, VALUE val)</tt> ::
Sets the value of the instance variable.
=== Control structure
<tt>VALUE rb_block_call(VALUE recv, ID mid, int argc, VALUE * argv, VALUE (*func) (ANYARGS), VALUE data2)</tt> ::
Calls a method on the _recv_, with the method name specified by the
symbol _mid_, with _argc_ arguments in _argv_, supplying _func_ as the
block. When _func_ is called as the block, it will receive the value
from +yield+ as the first argument, and _data2_ as the second argument.
When yielded with multiple values (in C, <tt>rb_yield_values()</tt>,
<tt>rb_yield_values2()</tt> and <tt>rb_yield_splat()</tt>), _data2_ is
packed as an Array, whereas yielded values can be gotten via
_argc_/_argv_ of the third/fourth arguments.
<tt>VALUE rb_block_call_kw(VALUE recv, ID mid, int argc, VALUE * argv, VALUE (*func) (ANYARGS), VALUE data2, int kw_splat)</tt> ::
Same as +rb_funcall_with_block+, using _kw_splat_ to determine whether
keyword arguments are passed.
\[OBSOLETE] <tt>VALUE rb_iterate(VALUE (*func1)(), VALUE arg1, VALUE (*func2)(), VALUE arg2)</tt> ::
Calls the function _func1_, supplying _func2_ as the block. _func1_ will be
called with the argument _arg1_. _func2_ receives the value from +yield+ as
the first argument, _arg2_ as the second argument.
When +rb_iterate+ is used in 1.9, _func1_ has to call some Ruby-level method.
This function is obsolete since 1.9; use +rb_block_call+ instead.
<tt>VALUE rb_yield(VALUE val)</tt> ::
Yields _val_ as a single argument to the block.
<tt>VALUE rb_yield_values(int n, ...)</tt> ::
Yields _n_ number of arguments to the block, using one C argument per Ruby
argument.
<tt>VALUE rb_yield_values2(int n, VALUE *argv)</tt> ::
Yields _n_ number of arguments to the block, with all Ruby arguments in the
C argv array.
<tt>VALUE rb_yield_values_kw(int n, VALUE *argv, int kw_splat)</tt> ::
Same as +rb_yield_values2+, using _kw_splat_ to determine whether
keyword arguments are passed.
<tt>VALUE rb_yield_splat(VALUE args)</tt> ::
Same as +rb_yield_values2+, except arguments are specified by the Ruby
array +args+.
<tt>VALUE rb_yield_splat_kw(VALUE args, int kw_splat)</tt> ::
Same as +rb_yield_splat+, using _kw_splat_ to determine whether
keyword arguments are passed.
<tt>VALUE rb_rescue(VALUE (*func1)(ANYARGS), VALUE arg1, VALUE (*func2)(ANYARGS), VALUE arg2)</tt> ::
Calls the function _func1_, with _arg1_ as the argument. If an exception
occurs during _func1_, it calls _func2_ with _arg2_ as the first argument
and the exception object as the second argument. The return value of
<tt>rb_rescue()</tt> is the return value from _func1_ if no exception
occurs, from _func2_ otherwise.
<tt>VALUE rb_ensure(VALUE (*func1)(ANYARGS), VALUE arg1, VALUE (*func2)(ANYARGS), VALUE arg2)</tt> ::
Calls the function _func1_ with _arg1_ as the argument, then calls _func2_
with _arg2_ if execution terminated. The return value from
<tt>rb_ensure()</tt> is that of _func1 when no exception occurred.
<tt>VALUE rb_protect(VALUE (*func) (VALUE), VALUE arg, int *state)</tt> ::
Calls the function _func_ with _arg_ as the argument. If no exception
occurred during _func_, it returns the result of _func_ and
<tt>*state</tt> is zero.
Otherwise, it returns +Qnil+ and sets <tt>*state</tt> to nonzero. If
state is NULL, it is not set in both cases. You have to clear the error
info with <tt>rb_set_errinfo(Qnil)</tt> when ignoring the caught exception.
<tt>void rb_jump_tag(int state)</tt> ::
Continues the exception caught by <tt>rb_protect()</tt> and <tt>rb_eval_string_protect()</tt>.
_state_ must be the returned value from those functions. This function
never return to the caller.
<tt>void rb_iter_break()</tt> ::
Exits from the current innermost block. This function never return to
the caller.
<tt>void rb_iter_break_value(VALUE value)</tt> ::
Exits from the current innermost block with the value. The block will
return the given argument value. This function never return to the
caller.
=== Exceptions and errors
<tt>void rb_warn(const char *fmt, ...)</tt> ::
Prints a warning message according to a printf-like format.
<tt>void rb_warning(const char *fmt, ...)</tt> ::
Prints a warning message according to a printf-like format, if
_$VERBOSE_ is true.
<tt>void rb_raise(rb_eRuntimeError, const char *fmt, ...)</tt> ::
Raises RuntimeError. The _fmt_ is a format string just like printf().
<tt>void rb_raise(VALUE exception, const char *fmt, ...)</tt> ::
Raises a class exception. The _fmt_ is a format string just like printf().
<tt>void rb_fatal(const char *fmt, ...)</tt> ::
Raises a fatal error, terminates the interpreter. No exception handling
will be done for fatal errors, but ensure blocks will be executed.
<tt>void rb_bug(const char *fmt, ...)</tt> ::
Terminates the interpreter immediately. This function should be
called under the situation caused by the bug in the interpreter. No
exception handling nor ensure execution will be done.
Note: In the format string, <tt>"%"PRIsVALUE</tt> can be used for Object#to_s
(or Object#inspect if <tt>+</tt> flag is set) output (and related argument
must be a VALUE). Since it conflicts with <tt>"%i"</tt>, for integers in
format strings, use <tt>"%d"</tt>.
=== Threading
As of Ruby 1.9, Ruby supports native 1:1 threading with one kernel
thread per Ruby Thread object. Currently, there is a GVL (Global VM Lock)
which prevents simultaneous execution of Ruby code which may be released
by the +rb_thread_call_without_gvl+ and +rb_thread_call_without_gvl2+ functions.
These functions are tricky-to-use and documented in thread.c; do not
use them before reading comments in thread.c.
<tt>void rb_thread_schedule(void)</tt> ::
Give the scheduler a hint to pass execution to another thread.
=== Input/Output (IO) on a single file descriptor
<tt>int rb_io_wait_readable(int fd)</tt> ::
Wait indefinitely for the given FD to become readable, allowing other
threads to be scheduled. Returns a true value if a read may be
performed, false if there is an unrecoverable error.
<tt>int rb_io_wait_writable(int fd)</tt> ::
Like +rb_io_wait_readable+, but for writability.
<tt>int rb_wait_for_single_fd(int fd, int events, struct timeval *timeout)</tt> ::
Allows waiting on a single FD for one or multiple events with a
specified timeout.
_events_ is a mask of any combination of the following values:
+RB_WAITFD_IN+ :: wait for readability of normal data
+RB_WAITFD_OUT+ :: wait for writability
+RB_WAITFD_PRI+ :: wait for readability of urgent data
Use a NULL _timeout_ to wait indefinitely.
=== I/O multiplexing
Ruby supports I/O multiplexing based on the select(2) system call.
The Linux select_tut(2) manpage
<http://man7.org/linux/man-pages/man2/select_tut.2.html>
provides a good overview on how to use select(2), and the Ruby API has
analogous functions and data structures to the well-known select API.
Understanding of select(2) is required to understand this section.
<tt>typedef struct rb_fdset_t</tt> ::
The data structure which wraps the fd_set bitmap used by select(2).
This allows Ruby to use FD sets larger than that allowed by
historic limitations on modern platforms.
<tt>void rb_fd_init(rb_fdset_t *)</tt> ::
Initializes the +rb_fdset_t+, it must be initialized before other rb_fd_*
operations. Analogous to calling malloc(3) to allocate an fd_set.
<tt>void rb_fd_term(rb_fdset_t *)</tt> ::
Destroys the +rb_fdset_t+, releasing any memory and resources it used.
It must be reinitialized using +rb_fd_init+ before future use.
Analogous to calling free(3) to release memory for an fd_set.
<tt>void rb_fd_zero(rb_fdset_t *)</tt> ::
Clears all FDs from the +rb_fdset_t+, analogous to FD_ZERO(3).
<tt>void rb_fd_set(int fd, rb_fdset_t *)</tt> ::
Adds a given FD in the +rb_fdset_t+, analogous to FD_SET(3).
<tt>void rb_fd_clr(int fd, rb_fdset_t *)</tt> ::
Removes a given FD from the +rb_fdset_t+, analogous to FD_CLR(3).
<tt>int rb_fd_isset(int fd, const rb_fdset_t *)</tt> ::
Returns true if a given _fd_ is set in the +rb_fdset_t+, false if not.
Analogous to FD_ISSET(3).
<tt>int rb_thread_fd_select(int nfds, rb_fdset_t *readfds, rb_fdset_t *writefds, rb_fdset_t *exceptfds, struct timeval *timeout)</tt> ::
Analogous to the select(2) system call, but allows other Ruby
threads to be scheduled while waiting.
When only waiting on a single FD, favor +rb_io_wait_readable+,
+rb_io_wait_writable+, or +rb_wait_for_single_fd+ functions since
they can be optimized for specific platforms (currently, only Linux).
=== Initialize and start the interpreter
The embedding API functions are below (not needed for extension libraries):
<tt>void ruby_init()</tt> ::
Initializes the interpreter.
<tt>void *ruby_options(int argc, char **argv)</tt> ::
Process command line arguments for the interpreter.
And compiles the Ruby source to execute.
It returns an opaque pointer to the compiled source
or an internal special value.
<tt>int ruby_run_node(void *n)</tt> ::
Runs the given compiled source and exits this process.
It returns EXIT_SUCCESS if successfully runs the source.
Otherwise, it returns other value.
<tt>void ruby_script(char *name)</tt> ::
Specifies the name of the script (_$0_).
=== Hooks for the interpreter events
<tt>void rb_add_event_hook(rb_event_hook_func_t func, rb_event_flag_t events, VALUE data)</tt> ::
Adds a hook function for the specified interpreter events.
events should be OR'ed value of:
RUBY_EVENT_LINE
RUBY_EVENT_CLASS
RUBY_EVENT_END
RUBY_EVENT_CALL
RUBY_EVENT_RETURN
RUBY_EVENT_C_CALL
RUBY_EVENT_C_RETURN
RUBY_EVENT_RAISE
RUBY_EVENT_ALL
The definition of rb_event_hook_func_t is below:
typedef void (*rb_event_hook_func_t)(rb_event_t event, VALUE data,
VALUE self, ID id, VALUE klass)
The third argument _data_ to <code>rb_add_event_hook()</code> is passed to the hook
function as the second argument, which was the pointer to the current
NODE in 1.8. See +RB_EVENT_HOOKS_HAVE_CALLBACK_DATA+ below.
<tt>int rb_remove_event_hook(rb_event_hook_func_t func)</tt> ::
Removes the specified hook function.
=== Memory usage
<tt>void rb_gc_adjust_memory_usage(ssize_t diff)</tt> ::
Adjusts the amount of registered external memory. You can tell GC how
much memory is used by an external library by this function. Calling
this function with positive _diff_ means the memory usage is increased;
new memory block is allocated or a block is reallocated as larger
size. Calling this function with negative _diff_ means the memory usage
is decreased; a memory block is freed or a block is reallocated as
smaller size. This function may trigger the GC.
=== Macros for compatibility
Some macros to check API compatibilities are available by default.
+NORETURN_STYLE_NEW+ ::
Means that +NORETURN+ macro is functional style instead of prefix.
+HAVE_RB_DEFINE_ALLOC_FUNC+ ::
Means that function <tt>rb_define_alloc_func()</tt> is provided, that means the
allocation framework is used. This is the same as the result of
<tt>have_func("rb_define_alloc_func", "ruby.h")</tt>.
+HAVE_RB_REG_NEW_STR+ ::
Means that function <tt>rb_reg_new_str()</tt> is provided, that creates Regexp
object from String object. This is the same as the result of
<tt>have_func("rb_reg_new_str", "ruby.h")</tt>.
+HAVE_RB_IO_T+ ::
Means that type +rb_io_t+ is provided.
+HAVE_RB_LEN_T+ ::
Means that types +rb_len_t+ and +rb_ulen_t+ are provided, along with the
macros for them such as +LEN2NUM+, +NUM2LEN+ and +PRIdLEN+.
+USE_SYMBOL_AS_METHOD_NAME+ ::
Means that Symbols will be returned as method names, e.g.,
Module#methods, \#singleton_methods and so on.
<tt>HAVE_RUBY_*_H</tt> ::
Defined in ruby.h and means corresponding header is available. For
instance, when +HAVE_RUBY_ST_H+ is defined you should use ruby/st.h not
mere st.h.
Header files corresponding to these macros may be <tt>#include</tt>
directly from extension libraries.
+RB_EVENT_HOOKS_HAVE_CALLBACK_DATA+ ::
Means that <code>rb_add_event_hook()</code> takes the third argument _data_, to be
passed to the given event hook function.
=== Defining backward compatible macros for keyword argument functions
Most ruby C extensions are designed to support multiple Ruby versions.
In order to correctly support Ruby 2.7+ in regards to keyword
argument separation, C extensions need to use <code>*_kw</code>
functions. However, these functions do not exist in Ruby 2.6 and
below, so in those cases macros should be defined to allow you to use
the same code on multiple Ruby versions. Here are example macros
you can use in extensions that support Ruby 2.6 (or below) when using
the <code>*_kw</code> functions introduced in Ruby 2.7.
#ifndef RB_PASS_KEYWORDS
/* Only define macros on Ruby <2.7 */
#define rb_funcallv_kw(o, m, c, v, kw) rb_funcallv(o, m, c, v)
#define rb_funcallv_public_kw(o, m, c, v, kw) rb_funcallv_public(o, m, c, v)
#define rb_funcall_passing_block_kw(o, m, c, v, kw) rb_funcall_passing_block(o, m, c, v)
#define rb_funcall_with_block_kw(o, m, c, v, b, kw) rb_funcall_with_block(o, m, c, v, b)
#define rb_scan_args_kw(kw, c, v, s, ...) rb_scan_args(c, v, s, __VA_ARGS__)
#define rb_call_super_kw(c, v, kw) rb_call_super(c, v)
#define rb_yield_values_kw(c, v, kw) rb_yield_values2(c, v)
#define rb_yield_splat_kw(a, kw) rb_yield_splat(a)
#define rb_block_call_kw(o, m, c, v, f, p, kw) rb_block_call(o, m, c, v, f, p)
#define rb_fiber_resume_kw(o, c, v, kw) rb_fiber_resume(o, c, v)
#define rb_fiber_yield_kw(c, v, kw) rb_fiber_yield(c, v)
#define rb_enumeratorize_with_size_kw(o, m, c, v, f, kw) rb_enumeratorize_with_size(o, m, c, v, f)
#define SIZED_ENUMERATOR_KW(obj, argc, argv, size_fn, kw_splat) \
rb_enumeratorize_with_size((obj), ID2SYM(rb_frame_this_func()), \
(argc), (argv), (size_fn))
#define RETURN_SIZED_ENUMERATOR_KW(obj, argc, argv, size_fn, kw_splat) do { \
if (!rb_block_given_p()) \
return SIZED_ENUMERATOR(obj, argc, argv, size_fn); \
} while (0)
#define RETURN_ENUMERATOR_KW(obj, argc, argv, kw_splat) RETURN_SIZED_ENUMERATOR(obj, argc, argv, 0)
#define rb_check_funcall_kw(o, m, c, v, kw) rb_check_funcall(o, m, c, v)
#define rb_obj_call_init_kw(o, c, v, kw) rb_obj_call_init(o, c, v)
#define rb_class_new_instance_kw(c, v, k, kw) rb_class_new_instance(c, v, k)
#define rb_proc_call_kw(p, a, kw) rb_proc_call(p, a)
#define rb_proc_call_with_block_kw(p, c, v, b, kw) rb_proc_call_with_block(p, c, v, b)
#define rb_method_call_kw(c, v, m, kw) rb_method_call(c, v, m)
#define rb_method_call_with_block_kw(c, v, m, b, kw) rb_method_call_with_block(c, v, m, b)
#define rb_eval_cmd_kw(c, a, kw) rb_eval_cmd(c, a, 0)
#endif
== Appendix C. Functions available for use in extconf.rb
See documentation for {mkmf}[rdoc-ref:MakeMakefile].
== Appendix D. Generational GC
Ruby 2.1 introduced a generational garbage collector (called RGenGC).
RGenGC (mostly) keeps compatibility.
Generally, the use of the technique called write barriers is required in
extension libraries for {generational GC
}[https://en.wikipedia.org/wiki/Garbage_collection_%28computer_science%29].
RGenGC works fine without write barriers in extension libraries.
If your library adheres to the following tips, performance can be further
improved. Especially, the "{Don't touch pointers
directly}[rdoc-ref:@Don-27t+touch+pointers+directly]" section is important.
=== Incompatibility
You can't write <tt>RBASIC(obj)->klass</tt> field directly because it is
const value now.
Basically you should not write this field because MRI expects it to be
an immutable field, but if you want to do it in your extension you can
use the following functions:
<tt>VALUE rb_obj_hide(VALUE obj)</tt> ::
Clear +RBasic::klass+ field. The object will be an internal object.
ObjectSpace::each_object can't find this object.
<tt>VALUE rb_obj_reveal(VALUE obj, VALUE klass)</tt> ::
Reset +RBasic::klass+ to be _klass_.
We expect the _klass_ is hidden class by <tt>rb_obj_hide()</tt>.
=== Write barriers
RGenGC doesn't require write barriers to support generational GC.
However, caring about write barrier can improve the performance of
RGenGC. Please check the following tips.
==== Don't touch pointers directly
In MRI (include/ruby/ruby.h), some macros acquire pointers to the
internal data structures such as <tt>RARRAY_PTR()</tt>,
<tt>RSTRUCT_PTR()</tt> and so on.
DO NOT USE THESE MACROS and instead use the corresponding C-APIs such as
<tt>rb_ary_aref()</tt>, <tt>rb_ary_store()</tt> and so on.
==== Consider whether to insert write barriers
You don't need to care about write barriers if you only use built-in
types.
If you support +T_DATA+ objects, you may consider using write barriers.
Inserting write barriers into +T_DATA+ objects only works with the
following type objects: (a) long-lived objects, (b) when a huge number
of objects are generated and \(c) container-type objects that have
references to other objects. If your extension provides such a type of
+T_DATA+ objects, consider inserting write barriers.
a. short-lived objects don't become old generation objects.
b. only a few oldgen objects don't have performance impact.
c. only a few references don't have performance impact.
Inserting write barriers is a very difficult hack, it is easy to
introduce critical bugs. And inserting write barriers has several areas
of overhead. Basically we don't recommend you insert write barriers.
Please carefully consider the risks.
==== Combine with built-in types
Please consider utilizing built-in types. Most built-in types support
write barrier, so you can use them to avoid manually inserting write
barriers.
For example, if your +T_DATA+ has references to other objects, then you
can move these references to Array. A +T_DATA+ object only has a reference
to an array object. Or you can also use a Struct object to gather a
+T_DATA+ object (without any references) and an Array that contains
references.
With use of such techniques, you don't need to insert write barriers
anymore.
==== Insert write barriers
\[AGAIN] Inserting write barriers is a very difficult hack, and it is
easy to introduce critical bugs. And inserting write barriers has
several areas of overhead. Basically we don't recommend you insert write
barriers. Please carefully consider the risks.
Before inserting write barriers, you need to know about RGenGC algorithm
(gc.c will help you). Macros and functions to insert write barriers are
available in include/ruby/ruby.h. An example is available in iseq.c.
For a complete guide for RGenGC and write barriers, please refer to
<https://bugs.ruby-lang.org/projects/ruby-master/wiki/RGenGC>.
== Appendix E. +RB_GC_GUARD+ to protect from premature GC
C Ruby currently uses conservative garbage collection, thus VALUE
variables must remain visible on the stack or registers to ensure any
associated data remains usable. Optimizing C compilers are not designed
with conservative garbage collection in mind, so they may optimize away
the original VALUE even if the code depends on data associated with that
VALUE.
The following example illustrates the use of +RB_GC_GUARD+ to ensure
the contents of _sptr_ remain valid while the second invocation of
+rb_str_new_cstr+ is running.
VALUE s, w;
const char *sptr;
s = rb_str_new_cstr("hello world!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!");
sptr = RSTRING_PTR(s);
w = rb_str_new_cstr(sptr + 6); /* Possible GC invocation */
RB_GC_GUARD(s); /* ensure s (and thus sptr) do not get GC-ed */
In the above example, +RB_GC_GUARD+ must be placed *after* the last use of
_sptr_. Placing +RB_GC_GUARD+ before dereferencing _sptr_ would be of no use.
+RB_GC_GUARD+ is only effective on the VALUE data type, not converted C
data types.
+RB_GC_GUARD+ would not be necessary at all in the above example if
non-inlined function calls are made on the _s_ VALUE after _sptr_ is
dereferenced. Thus, in the above example, calling any un-inlined
function on _s_ such as:
rb_str_modify(s);
Will ensure _s_ stays on the stack or register to prevent a
GC invocation from prematurely freeing it.
Using the +RB_GC_GUARD+ macro is preferable to using the "volatile"
keyword in C. +RB_GC_GUARD+ has the following advantages:
1. the intent of the macro use is clear
2. +RB_GC_GUARD+ only affects its call site, "volatile" generates some
extra code every time the variable is used, hurting optimization.
3. "volatile" implementations may be buggy/inconsistent in some
compilers and architectures. +RB_GC_GUARD+ is customizable for broken
systems/compilers without negatively affecting other systems.
== Appendix F. Ractor support
Ractor(s) are the parallel execution mechanism introduced in Ruby 3.0. All
ractors can run in parallel on a different OS thread (using an underlying system
provided thread), so the C extension should be thread-safe. A C extension that
can run in multiple ractors is called "Ractor-safe".
Ractor safety around C extensions has the following properties:
1. By default, all C extensions are recognized as Ractor-unsafe.
2. Ractor-unsafe C-methods may only be called from the main Ractor. If invoked
by a non-main Ractor, then a Ractor::UnsafeError is raised.
3. If an extension desires to be marked as Ractor-safe the extension should call
<tt>rb_ext_ractor_safe(true)</tt> at the +Init_+ function for the extension,
and all defined methods will be marked as Ractor-safe.
To make a "Ractor-safe" C extension, we need to check the following points:
1. Do not share unshareable objects between ractors
For example, C's global variable can lead sharing an unshareable objects
between ractors.
VALUE g_var;
VALUE set(VALUE self, VALUE v){ return g_var = v; }
VALUE get(VALUE self){ return g_var; }
<tt>set()</tt> and <tt>get()</tt> pair can share an unshareable objects
using _g_var_, and it is Ractor-unsafe.
Not only using global variables directly, some indirect data structure
such as global +st_table+ can share the objects, so please take care.
Note that class and module objects are shareable objects, so you can
keep the code <tt>cFoo = rb_define_class(...)</tt> with C's global
variables.
2. Check the thread-safety of the extension
An extension should be thread-safe. For example, the following code is
not thread-safe:
bool g_called = false;
VALUE call(VALUE self) {
if (g_called) rb_raise("recursive call is not allowed.");
g_called = true;
VALUE ret = do_something();
g_called = false;
return ret;
}
because _g_called_ global variable should be synchronized by other
ractor's threads. To avoid such data-race, some synchronization should
be used. Check include/ruby/thread_native.h and include/ruby/atomic.h.
With Ractors, all objects given as method parameters and the receiver (self)
are guaranteed to be from the current Ractor or to be shareable. As a
consequence, it is easier to make code ractor-safe than to make code generally
thread-safe. For example, we don't need to lock an array object to access the
element of it.
3. Check the thread-safety of any used library
If the extension relies on an external library, such as a function
<tt>foo()</tt> from a library libfoo, the function libfoo <tt>foo()</tt>
should be thread safe.
4. Make an object shareable
This is not required to make an extension Ractor-safe.
If an extension provides special objects defined by +rb_data_type_t+,
consider these objects can become shareable or not.
+RUBY_TYPED_FROZEN_SHAREABLE+ flag indicates that these objects can be
shareable objects if the object is frozen. This means that if the object is
frozen, the mutation of wrapped data is not allowed.
5. Others
There are possibly other points or requirements which must be considered in
the making of a Ractor-safe extension. This document will be extended as
they are discovered.
== Appendix G. Embedded TypedData
Here is an example of how to use +RUBY_TYPED_EMBEDDABLE+:
struct my_data {
struct timespec created_at;
size_t buffer_capa;
char *buffer;
};
static void
my_data_free(void *ptr)
{
struct my_data *data = (struct my_data *)ptr;
// Deliberately don't free `ptr` if it is embeddable.
// Only auxiliary memory needs to be freed.
ruby_xfree(data->buffer);
}
static size_t
my_data_size(const void *ptr)
{
const struct my_data *data = (const struct my_data *)ptr;
// We don't need to account for `sizeof(struct my_data)` because it is embedded inside the Ruby object.
// Only auxiliary memory needs to be reported.
return data->buffer_capa;
}
static const rb_data_type_t my_type = {
.wrap_struct_name = "my_type",
.function = {
.dfree = my_data_free,
.dsize = my_data_size,
},
.flags = RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_EMBEDDABLE,
};
static VALUE
my_data_alloc(VALUE klass)
{
struct my_data *data;
VALUE obj = TypedData_Make_Struct(klass, struct my_data, &my_type, data);
// It is fine to pass pointers into the embedded struct, for as long as
// the called function won't use it after the Ruby object has left the stack.
clock_gettime(CLOCK_REALTIME, &data->created_at);
data->buffer_capa = 1024;
data->buffer = ZALLOC_N(char, data->buffer_capa);
return obj;
}
static VALUE
my_data_read(VALUE self)
{
struct my_data *data;
TypedData_Get_Struct(self, struct my_data, &my_type, data);
// `self` is received from `rb_define_method` so `RB_GC_GUARD` isn't necessary.
return rb_str_new(data->buffer, data->buffer_capa);
}
static VALUE
my_data_m_parse(VALUE klass)
{
VALUE my_data_obj = my_data_alloc(klass);
VALUE result = my_data_read(my_data_obj);
// `my_data_obj` was allocated from C, `RB_GC_GUARD` must be used to
// ensure the compiler will keep its reference on the stack.
RB_GC_GUARD(my_data_obj);
return result;
}
void
Init_my_data(void)
{
VALUE cMyData = rb_define_class("MyData", rb_cObject);
rb_define_method(cMyData, "read", my_data_read, 0);
rb_define_singleton_method(cMyData, "parse", my_data_m_parse, 0);
}
== Appendix H. Efficent Extensions
The Ruby native extension API is large and some parts are kept for compatibility but are discouraged or "soft-deprecated".
See https://github.com/ruby/ruby/blob/master/doc/efficient_extensions.md for a list of those and tips on which APIs to use instead.
It also documents what to use for extensions to be efficient and avoid disabling Ruby VM optimizations.
--
Local variables:
fill-column: 78
end:
++
View Git Blame Copy Permalink