3 perlxs - XS language reference manual
9 XS is a language used to create an extension interface
10 between Perl and some C library which one wishes to use with
11 Perl. The XS interface is combined with the library to
12 create a new library which can be linked to Perl. An B<XSUB>
13 is a function in the XS language and is the core component
14 of the Perl application interface.
16 The XS compiler is called B<xsubpp>. This compiler will embed
17 the constructs necessary to let an XSUB, which is really a C
18 function in disguise, manipulate Perl values and creates the
19 glue necessary to let Perl access the XSUB. The compiler
20 uses B<typemaps> to determine how to map C function parameters
21 and variables to Perl values. The default typemap handles
22 many common C types. A supplement typemap must be created
23 to handle special structures and types for the library being
26 See L<perlxstut> for a tutorial on the whole extension creation process.
28 Note: For many extensions, Dave Beazley's SWIG system provides a
29 significantly more convenient mechanism for creating the XS glue
30 code. See L<http://www.cs.utah.edu/~beazley/SWIG> for more
35 Many of the examples which follow will concentrate on creating an interface
36 between Perl and the ONC+ RPC bind library functions. The rpcb_gettime()
37 function is used to demonstrate many features of the XS language. This
38 function has two parameters; the first is an input parameter and the second
39 is an output parameter. The function also returns a status value.
41 bool_t rpcb_gettime(const char *host, time_t *timep);
43 From C this function will be called with the following
49 status = rpcb_gettime( "localhost", &timep );
51 If an XSUB is created to offer a direct translation between this function
52 and Perl, then this XSUB will be used from Perl with the following code.
53 The $status and $timep variables will contain the output of the function.
56 $status = rpcb_gettime( "localhost", $timep );
58 The following XS file shows an XS subroutine, or XSUB, which
59 demonstrates one possible interface to the rpcb_gettime()
60 function. This XSUB represents a direct translation between
61 C and Perl and so preserves the interface even from Perl.
62 This XSUB will be invoked from Perl with the usage shown
63 above. Note that the first three #include statements, for
64 C<EXTERN.h>, C<perl.h>, and C<XSUB.h>, will always be present at the
65 beginning of an XS file. This approach and others will be
66 expanded later in this document.
73 MODULE = RPC PACKAGE = RPC
76 rpcb_gettime(host,timep)
82 Any extension to Perl, including those containing XSUBs,
83 should have a Perl module to serve as the bootstrap which
84 pulls the extension into Perl. This module will export the
85 extension's functions and variables to the Perl program and
86 will cause the extension's XSUBs to be linked into Perl.
87 The following module will be used for most of the examples
88 in this document and should be used from Perl with the C<use>
89 command as shown earlier. Perl modules are explained in
90 more detail later in this document.
96 @ISA = qw(Exporter DynaLoader);
97 @EXPORT = qw( rpcb_gettime );
102 Throughout this document a variety of interfaces to the rpcb_gettime()
103 XSUB will be explored. The XSUBs will take their parameters in different
104 orders or will take different numbers of parameters. In each case the
105 XSUB is an abstraction between Perl and the real C rpcb_gettime()
106 function, and the XSUB must always ensure that the real rpcb_gettime()
107 function is called with the correct parameters. This abstraction will
108 allow the programmer to create a more Perl-like interface to the C
111 =head2 The Anatomy of an XSUB
113 The following XSUB allows a Perl program to access a C library function
114 called sin(). The XSUB will imitate the C function which takes a single
115 argument and returns a single value.
121 When using C pointers the indirection operator C<*> should be considered
122 part of the type and the address operator C<&> should be considered part of
123 the variable, as is demonstrated in the rpcb_gettime() function above. See
124 the section on typemaps for more about handling qualifiers and unary
125 operators in C types.
127 The function name and the return type must be placed on
136 The function body may be indented or left-adjusted. The following example
137 shows a function with its body left-adjusted. Most examples in this
138 document will indent the body.
146 =head2 The Argument Stack
148 The argument stack is used to store the values which are
149 sent as parameters to the XSUB and to store the XSUB's
150 return value. In reality all Perl functions keep their
151 values on this stack at the same time, each limited to its
152 own range of positions on the stack. In this document the
153 first position on that stack which belongs to the active
154 function will be referred to as position 0 for that function.
156 XSUBs refer to their stack arguments with the macro B<ST(x)>, where I<x>
157 refers to a position in this XSUB's part of the stack. Position 0 for that
158 function would be known to the XSUB as ST(0). The XSUB's incoming
159 parameters and outgoing return values always begin at ST(0). For many
160 simple cases the B<xsubpp> compiler will generate the code necessary to
161 handle the argument stack by embedding code fragments found in the
162 typemaps. In more complex cases the programmer must supply the code.
164 =head2 The RETVAL Variable
166 The RETVAL variable is a magic variable which always matches
167 the return type of the C library function. The B<xsubpp> compiler will
168 supply this variable in each XSUB and by default will use it to hold the
169 return value of the C library function being called. In simple cases the
170 value of RETVAL will be placed in ST(0) of the argument stack where it can
171 be received by Perl as the return value of the XSUB.
173 If the XSUB has a return type of C<void> then the compiler will
174 not supply a RETVAL variable for that function. When using
175 the PPCODE: directive the RETVAL variable is not needed, unless used
178 If PPCODE: directive is not used, C<void> return value should be used
179 only for subroutines which do not return a value, I<even if> CODE:
180 directive is used which sets ST(0) explicitly.
182 Older versions of this document recommended to use C<void> return
183 value in such cases. It was discovered that this could lead to
184 segfaults in cases when XSUB was I<truely> C<void>. This practice is
185 now deprecated, and may be not supported at some future version. Use
186 the return value C<SV *> in such cases. (Currently C<xsubpp> contains
187 some heuristic code which tries to disambiguate between "truely-void"
188 and "old-practice-declared-as-void" functions. Hence your code is at
189 mercy of this heuristics unless you use C<SV *> as return value.)
191 =head2 The MODULE Keyword
193 The MODULE keyword is used to start the XS code and to
194 specify the package of the functions which are being
195 defined. All text preceding the first MODULE keyword is
196 considered C code and is passed through to the output
197 untouched. Every XS module will have a bootstrap function
198 which is used to hook the XSUBs into Perl. The package name
199 of this bootstrap function will match the value of the last
200 MODULE statement in the XS source files. The value of
201 MODULE should always remain constant within the same XS
202 file, though this is not required.
204 The following example will start the XS code and will place
205 all functions in a package named RPC.
209 =head2 The PACKAGE Keyword
211 When functions within an XS source file must be separated into packages
212 the PACKAGE keyword should be used. This keyword is used with the MODULE
213 keyword and must follow immediately after it when used.
215 MODULE = RPC PACKAGE = RPC
217 [ XS code in package RPC ]
219 MODULE = RPC PACKAGE = RPCB
221 [ XS code in package RPCB ]
223 MODULE = RPC PACKAGE = RPC
225 [ XS code in package RPC ]
227 Although this keyword is optional and in some cases provides redundant
228 information it should always be used. This keyword will ensure that the
229 XSUBs appear in the desired package.
231 =head2 The PREFIX Keyword
233 The PREFIX keyword designates prefixes which should be
234 removed from the Perl function names. If the C function is
235 C<rpcb_gettime()> and the PREFIX value is C<rpcb_> then Perl will
236 see this function as C<gettime()>.
238 This keyword should follow the PACKAGE keyword when used.
239 If PACKAGE is not used then PREFIX should follow the MODULE
242 MODULE = RPC PREFIX = rpc_
244 MODULE = RPC PACKAGE = RPCB PREFIX = rpcb_
246 =head2 The OUTPUT: Keyword
248 The OUTPUT: keyword indicates that certain function parameters should be
249 updated (new values made visible to Perl) when the XSUB terminates or that
250 certain values should be returned to the calling Perl function. For
251 simple functions, such as the sin() function above, the RETVAL variable is
252 automatically designated as an output value. In more complex functions
253 the B<xsubpp> compiler will need help to determine which variables are output
256 This keyword will normally be used to complement the CODE: keyword.
257 The RETVAL variable is not recognized as an output variable when the
258 CODE: keyword is present. The OUTPUT: keyword is used in this
259 situation to tell the compiler that RETVAL really is an output
262 The OUTPUT: keyword can also be used to indicate that function parameters
263 are output variables. This may be necessary when a parameter has been
264 modified within the function and the programmer would like the update to
268 rpcb_gettime(host,timep)
274 The OUTPUT: keyword will also allow an output parameter to
275 be mapped to a matching piece of code rather than to a
279 rpcb_gettime(host,timep)
283 timep sv_setnv(ST(1), (double)timep);
285 B<xsubpp> emits an automatic C<SvSETMAGIC()> for all parameters in the
286 OUTPUT section of the XSUB, except RETVAL. This is the usually desired
287 behavior, as it takes care of properly invoking 'set' magic on output
288 parameters (needed for hash or array element parameters that must be
289 created if they didn't exist). If for some reason, this behavior is
290 not desired, the OUTPUT section may contain a C<SETMAGIC: DISABLE> line
291 to disable it for the remainder of the parameters in the OUTPUT section.
292 Likewise, C<SETMAGIC: ENABLE> can be used to reenable it for the
293 remainder of the OUTPUT section. See L<perlguts> for more details
296 =head2 The CODE: Keyword
298 This keyword is used in more complicated XSUBs which require
299 special handling for the C function. The RETVAL variable is
300 available but will not be returned unless it is specified
301 under the OUTPUT: keyword.
303 The following XSUB is for a C function which requires special handling of
304 its parameters. The Perl usage is given first.
306 $status = rpcb_gettime( "localhost", $timep );
311 rpcb_gettime(host,timep)
315 RETVAL = rpcb_gettime( host, &timep );
320 =head2 The INIT: Keyword
322 The INIT: keyword allows initialization to be inserted into the XSUB before
323 the compiler generates the call to the C function. Unlike the CODE: keyword
324 above, this keyword does not affect the way the compiler handles RETVAL.
327 rpcb_gettime(host,timep)
331 printf("# Host is %s\n", host );
335 =head2 The NO_INIT Keyword
337 The NO_INIT keyword is used to indicate that a function
338 parameter is being used only as an output value. The B<xsubpp>
339 compiler will normally generate code to read the values of
340 all function parameters from the argument stack and assign
341 them to C variables upon entry to the function. NO_INIT
342 will tell the compiler that some parameters will be used for
343 output rather than for input and that they will be handled
344 before the function terminates.
346 The following example shows a variation of the rpcb_gettime() function.
347 This function uses the timep variable only as an output variable and does
348 not care about its initial contents.
351 rpcb_gettime(host,timep)
353 time_t &timep = NO_INIT
357 =head2 Initializing Function Parameters
359 Function parameters are normally initialized with their
360 values from the argument stack. The typemaps contain the
361 code segments which are used to transfer the Perl values to
362 the C parameters. The programmer, however, is allowed to
363 override the typemaps and supply alternate initialization
366 The following code demonstrates how to supply initialization code for
367 function parameters. The initialization code is eval'd by the compiler
368 before it is added to the output so anything which should be interpreted
369 literally, such as double quotes, must be protected with backslashes.
372 rpcb_gettime(host,timep)
373 char *host = (char *)SvPV(ST(0),na);
378 This should not be used to supply default values for parameters. One
379 would normally use this when a function parameter must be processed by
380 another library function before it can be used. Default parameters are
381 covered in the next section.
383 =head2 Default Parameter Values
385 Default values can be specified for function parameters by
386 placing an assignment statement in the parameter list. The
387 default value may be a number or a string. Defaults should
388 always be used on the right-most parameters only.
390 To allow the XSUB for rpcb_gettime() to have a default host
391 value the parameters to the XSUB could be rearranged. The
392 XSUB will then call the real rpcb_gettime() function with
393 the parameters in the correct order. Perl will call this
394 XSUB with either of the following statements.
396 $status = rpcb_gettime( $timep, $host );
398 $status = rpcb_gettime( $timep );
400 The XSUB will look like the code which follows. A CODE:
401 block is used to call the real rpcb_gettime() function with
402 the parameters in the correct order for that function.
405 rpcb_gettime(timep,host="localhost")
407 time_t timep = NO_INIT
409 RETVAL = rpcb_gettime( host, &timep );
414 =head2 The PREINIT: Keyword
416 The PREINIT: keyword allows extra variables to be declared before the
417 typemaps are expanded. If a variable is declared in a CODE: block then that
418 variable will follow any typemap code. This may result in a C syntax
419 error. To force the variable to be declared before the typemap code, place
420 it into a PREINIT: block. The PREINIT: keyword may be used one or more
421 times within an XSUB.
423 The following examples are equivalent, but if the code is using complex
424 typemaps then the first example is safer.
428 time_t timep = NO_INIT
430 char *host = "localhost";
432 RETVAL = rpcb_gettime( host, &timep );
437 A correct, but error-prone example.
441 time_t timep = NO_INIT
443 char *host = "localhost";
444 RETVAL = rpcb_gettime( host, &timep );
449 =head2 The SCOPE: Keyword
451 The SCOPE: keyword allows scoping to be enabled for a particular XSUB. If
452 enabled, the XSUB will invoke ENTER and LEAVE automatically.
454 To support potentially complex type mappings, if a typemap entry used
455 by this XSUB contains a comment like C</*scope*/> then scoping will
456 automatically be enabled for that XSUB.
466 =head2 The INPUT: Keyword
468 The XSUB's parameters are usually evaluated immediately after entering the
469 XSUB. The INPUT: keyword can be used to force those parameters to be
470 evaluated a little later. The INPUT: keyword can be used multiple times
471 within an XSUB and can be used to list one or more input variables. This
472 keyword is used with the PREINIT: keyword.
474 The following example shows how the input parameter C<timep> can be
475 evaluated late, after a PREINIT.
478 rpcb_gettime(host,timep)
485 RETVAL = rpcb_gettime( host, &tt );
491 The next example shows each input parameter evaluated late.
494 rpcb_gettime(host,timep)
505 RETVAL = rpcb_gettime( h, &tt );
511 =head2 Variable-length Parameter Lists
513 XSUBs can have variable-length parameter lists by specifying an ellipsis
514 C<(...)> in the parameter list. This use of the ellipsis is similar to that
515 found in ANSI C. The programmer is able to determine the number of
516 arguments passed to the XSUB by examining the C<items> variable which the
517 B<xsubpp> compiler supplies for all XSUBs. By using this mechanism one can
518 create an XSUB which accepts a list of parameters of unknown length.
520 The I<host> parameter for the rpcb_gettime() XSUB can be
521 optional so the ellipsis can be used to indicate that the
522 XSUB will take a variable number of parameters. Perl should
523 be able to call this XSUB with either of the following statements.
525 $status = rpcb_gettime( $timep, $host );
527 $status = rpcb_gettime( $timep );
529 The XS code, with ellipsis, follows.
532 rpcb_gettime(timep, ...)
533 time_t timep = NO_INIT
535 char *host = "localhost";
538 host = (char *)SvPV(ST(1), na);
539 RETVAL = rpcb_gettime( host, &timep );
544 =head2 The C_ARGS: Keyword
546 The C_ARGS: keyword allows creating of XSUBS which have different
547 calling sequence from Perl than from C, without a need to write
548 CODE: or CPPCODE: section. The contents of the C_ARGS: paragraph is
549 put as the argument to the called C function without any change.
551 For example, suppose that C function is declared as
553 symbolic nth_derivative(int n, symbolic function, int flags);
555 and that the default flags are kept in a global C variable
556 C<default_flags>. Suppose that you want to create an interface which
559 $second_deriv = $function->nth_derivative(2);
561 To do this, declare the XSUB as
564 nth_derivative(function, n)
568 n, function, default_flags
570 =head2 The PPCODE: Keyword
572 The PPCODE: keyword is an alternate form of the CODE: keyword and is used
573 to tell the B<xsubpp> compiler that the programmer is supplying the code to
574 control the argument stack for the XSUBs return values. Occasionally one
575 will want an XSUB to return a list of values rather than a single value.
576 In these cases one must use PPCODE: and then explicitly push the list of
577 values on the stack. The PPCODE: and CODE: keywords are not used
578 together within the same XSUB.
580 The following XSUB will call the C rpcb_gettime() function
581 and will return its two output values, timep and status, to
582 Perl as a single list.
591 status = rpcb_gettime( host, &timep );
593 PUSHs(sv_2mortal(newSViv(status)));
594 PUSHs(sv_2mortal(newSViv(timep)));
596 Notice that the programmer must supply the C code necessary
597 to have the real rpcb_gettime() function called and to have
598 the return values properly placed on the argument stack.
600 The C<void> return type for this function tells the B<xsubpp> compiler that
601 the RETVAL variable is not needed or used and that it should not be created.
602 In most scenarios the void return type should be used with the PPCODE:
605 The EXTEND() macro is used to make room on the argument
606 stack for 2 return values. The PPCODE: directive causes the
607 B<xsubpp> compiler to create a stack pointer available as C<SP>, and it
608 is this pointer which is being used in the EXTEND() macro.
609 The values are then pushed onto the stack with the PUSHs()
612 Now the rpcb_gettime() function can be used from Perl with
613 the following statement.
615 ($status, $timep) = rpcb_gettime("localhost");
617 When handling output parameters with a PPCODE section, be sure to handle
618 'set' magic properly. See L<perlguts> for details about 'set' magic.
620 =head2 Returning Undef And Empty Lists
622 Occasionally the programmer will want to return simply
623 C<undef> or an empty list if a function fails rather than a
624 separate status value. The rpcb_gettime() function offers
625 just this situation. If the function succeeds we would like
626 to have it return the time and if it fails we would like to
627 have undef returned. In the following Perl code the value
628 of $timep will either be undef or it will be a valid time.
630 $timep = rpcb_gettime( "localhost" );
632 The following XSUB uses the C<SV *> return type as a mnemonic only,
633 and uses a CODE: block to indicate to the compiler
634 that the programmer has supplied all the necessary code. The
635 sv_newmortal() call will initialize the return value to undef, making that
636 the default return value.
645 ST(0) = sv_newmortal();
646 if( rpcb_gettime( host, &timep ) )
647 sv_setnv( ST(0), (double)timep);
649 The next example demonstrates how one would place an explicit undef in the
650 return value, should the need arise.
659 ST(0) = sv_newmortal();
660 if( rpcb_gettime( host, &timep ) ){
661 sv_setnv( ST(0), (double)timep);
667 To return an empty list one must use a PPCODE: block and
668 then not push return values on the stack.
676 if( rpcb_gettime( host, &timep ) )
677 PUSHs(sv_2mortal(newSViv(timep)));
679 /* Nothing pushed on stack, so an empty */
680 /* list is implicitly returned. */
683 Some people may be inclined to include an explicit C<return> in the above
684 XSUB, rather than letting control fall through to the end. In those
685 situations C<XSRETURN_EMPTY> should be used, instead. This will ensure that
686 the XSUB stack is properly adjusted. Consult L<perlguts/"API LISTING"> for
687 other C<XSRETURN> macros.
689 =head2 The REQUIRE: Keyword
691 The REQUIRE: keyword is used to indicate the minimum version of the
692 B<xsubpp> compiler needed to compile the XS module. An XS module which
693 contains the following statement will compile with only B<xsubpp> version
698 =head2 The CLEANUP: Keyword
700 This keyword can be used when an XSUB requires special cleanup procedures
701 before it terminates. When the CLEANUP: keyword is used it must follow
702 any CODE:, PPCODE:, or OUTPUT: blocks which are present in the XSUB. The
703 code specified for the cleanup block will be added as the last statements
706 =head2 The BOOT: Keyword
708 The BOOT: keyword is used to add code to the extension's bootstrap
709 function. The bootstrap function is generated by the B<xsubpp> compiler and
710 normally holds the statements necessary to register any XSUBs with Perl.
711 With the BOOT: keyword the programmer can tell the compiler to add extra
712 statements to the bootstrap function.
714 This keyword may be used any time after the first MODULE keyword and should
715 appear on a line by itself. The first blank line after the keyword will
716 terminate the code block.
719 # The following message will be printed when the
720 # bootstrap function executes.
721 printf("Hello from the bootstrap!\n");
723 =head2 The VERSIONCHECK: Keyword
725 The VERSIONCHECK: keyword corresponds to B<xsubpp>'s C<-versioncheck> and
726 C<-noversioncheck> options. This keyword overrides the command line
727 options. Version checking is enabled by default. When version checking is
728 enabled the XS module will attempt to verify that its version matches the
729 version of the PM module.
731 To enable version checking:
735 To disable version checking:
737 VERSIONCHECK: DISABLE
739 =head2 The PROTOTYPES: Keyword
741 The PROTOTYPES: keyword corresponds to B<xsubpp>'s C<-prototypes> and
742 C<-noprototypes> options. This keyword overrides the command line options.
743 Prototypes are enabled by default. When prototypes are enabled XSUBs will
744 be given Perl prototypes. This keyword may be used multiple times in an XS
745 module to enable and disable prototypes for different parts of the module.
747 To enable prototypes:
751 To disable prototypes:
755 =head2 The PROTOTYPE: Keyword
757 This keyword is similar to the PROTOTYPES: keyword above but can be used to
758 force B<xsubpp> to use a specific prototype for the XSUB. This keyword
759 overrides all other prototype options and keywords but affects only the
760 current XSUB. Consult L<perlsub/Prototypes> for information about Perl
764 rpcb_gettime(timep, ...)
765 time_t timep = NO_INIT
768 char *host = "localhost";
771 host = (char *)SvPV(ST(1), na);
772 RETVAL = rpcb_gettime( host, &timep );
777 =head2 The ALIAS: Keyword
779 The ALIAS: keyword allows an XSUB to have two or more unique Perl names
780 and to know which of those names was used when it was invoked. The Perl
781 names may be fully-qualified with package names. Each alias is given an
782 index. The compiler will setup a variable called C<ix> which contain the
783 index of the alias which was used. When the XSUB is called with its
784 declared name C<ix> will be 0.
786 The following example will create aliases C<FOO::gettime()> and
787 C<BAR::getit()> for this function.
790 rpcb_gettime(host,timep)
797 printf("# ix = %d\n", ix );
801 =head2 The INTERFACE: Keyword
803 This keyword declares the current XSUB as a keeper of the given
804 calling signature. If some text follows this keyword, it is
805 considered as a list of functions which have this signature, and
806 should be attached to XSUBs.
808 Say, if you have 4 functions multiply(), divide(), add(), subtract() all
811 symbolic f(symbolic, symbolic);
813 you code them all by using XSUB
816 interface_s_ss(arg1, arg2)
823 The advantage of this approach comparing to ALIAS: keyword is that one
824 can attach an extra function remainder() at runtime by using
826 CV *mycv = newXSproto("Symbolic::remainder",
827 XS_Symbolic_interface_s_ss, __FILE__, "$$");
828 XSINTERFACE_FUNC_SET(mycv, remainder);
830 (This example supposes that there was no INTERFACE_MACRO: section,
831 otherwise one needs to use something else instead of
832 C<XSINTERFACE_FUNC_SET>.)
834 =head2 The INTERFACE_MACRO: Keyword
836 This keyword allows one to define an INTERFACE using a different way
837 to extract a function pointer from an XSUB. The text which follows
838 this keyword should give the name of macros which would extract/set a
839 function pointer. The extractor macro is given return type, C<CV*>,
840 and C<XSANY.any_dptr> for this C<CV*>. The setter macro is given cv,
841 and the function pointer.
843 The default value is C<XSINTERFACE_FUNC> and C<XSINTERFACE_FUNC_SET>.
844 An INTERFACE keyword with an empty list of functions can be omitted if
845 INTERFACE_MACRO keyword is used.
847 Suppose that in the previous example functions pointers for
848 multiply(), divide(), add(), subtract() are kept in a global C array
849 C<fp[]> with offsets being C<multiply_off>, C<divide_off>, C<add_off>,
850 C<subtract_off>. Then one can use
852 #define XSINTERFACE_FUNC_BYOFFSET(ret,cv,f) \
853 ((XSINTERFACE_CVT(ret,))fp[CvXSUBANY(cv).any_i32])
854 #define XSINTERFACE_FUNC_BYOFFSET_set(cv,f) \
855 CvXSUBANY(cv).any_i32 = CAT2( f, _off )
860 interface_s_ss(arg1, arg2)
864 XSINTERFACE_FUNC_BYOFFSET
865 XSINTERFACE_FUNC_BYOFFSET_set
872 =head2 The INCLUDE: Keyword
874 This keyword can be used to pull other files into the XS module. The other
875 files may have XS code. INCLUDE: can also be used to run a command to
876 generate the XS code to be pulled into the module.
878 The file F<Rpcb1.xsh> contains our C<rpcb_gettime()> function:
881 rpcb_gettime(host,timep)
887 The XS module can use INCLUDE: to pull that file into it.
891 If the parameters to the INCLUDE: keyword are followed by a pipe (C<|>) then
892 the compiler will interpret the parameters as a command.
894 INCLUDE: cat Rpcb1.xsh |
896 =head2 The CASE: Keyword
898 The CASE: keyword allows an XSUB to have multiple distinct parts with each
899 part acting as a virtual XSUB. CASE: is greedy and if it is used then all
900 other XS keywords must be contained within a CASE:. This means nothing may
901 precede the first CASE: in the XSUB and anything following the last CASE: is
902 included in that case.
904 A CASE: might switch via a parameter of the XSUB, via the C<ix> ALIAS:
905 variable (see L<"The ALIAS: Keyword">), or maybe via the C<items> variable
906 (see L<"Variable-length Parameter Lists">). The last CASE: becomes the
907 B<default> case if it is not associated with a conditional. The following
908 example shows CASE switched via C<ix> with a function C<rpcb_gettime()>
909 having an alias C<x_gettime()>. When the function is called as
910 C<rpcb_gettime()> its parameters are the usual C<(char *host, time_t *timep)>,
911 but when the function is called as C<x_gettime()> its parameters are
912 reversed, C<(time_t *timep, char *host)>.
920 # 'a' is timep, 'b' is host
924 RETVAL = rpcb_gettime( b, &a );
929 # 'a' is host, 'b' is timep
936 That function can be called with either of the following statements. Note
937 the different argument lists.
939 $status = rpcb_gettime( $host, $timep );
941 $status = x_gettime( $timep, $host );
943 =head2 The & Unary Operator
945 The & unary operator is used to tell the compiler that it should dereference
946 the object when it calls the C function. This is used when a CODE: block is
947 not used and the object is a not a pointer type (the object is an C<int> or
948 C<long> but not a C<int*> or C<long*>).
950 The following XSUB will generate incorrect C code. The xsubpp compiler will
951 turn this into code which calls C<rpcb_gettime()> with parameters C<(char
952 *host, time_t timep)>, but the real C<rpcb_gettime()> wants the C<timep>
953 parameter to be of type C<time_t*> rather than C<time_t>.
956 rpcb_gettime(host,timep)
962 That problem is corrected by using the C<&> operator. The xsubpp compiler
963 will now turn this into code which calls C<rpcb_gettime()> correctly with
964 parameters C<(char *host, time_t *timep)>. It does this by carrying the
965 C<&> through, so the function call looks like C<rpcb_gettime(host, &timep)>.
968 rpcb_gettime(host,timep)
974 =head2 Inserting Comments and C Preprocessor Directives
976 C preprocessor directives are allowed within BOOT:, PREINIT: INIT:,
977 CODE:, PPCODE:, and CLEANUP: blocks, as well as outside the functions.
978 Comments are allowed anywhere after the MODULE keyword. The compiler
979 will pass the preprocessor directives through untouched and will remove
982 Comments can be added to XSUBs by placing a C<#> as the first
983 non-whitespace of a line. Care should be taken to avoid making the
984 comment look like a C preprocessor directive, lest it be interpreted as
985 such. The simplest way to prevent this is to put whitespace in front of
988 If you use preprocessor directives to choose one of two
989 versions of a function, use
992 #else /* ... version2 */
1002 because otherwise xsubpp will believe that you made a duplicate
1003 definition of the function. Also, put a blank line before the
1004 #else/#endif so it will not be seen as part of the function body.
1006 =head2 Using XS With C++
1008 If a function is defined as a C++ method then it will assume
1009 its first argument is an object pointer. The object pointer
1010 will be stored in a variable called THIS. The object should
1011 have been created by C++ with the new() function and should
1012 be blessed by Perl with the sv_setref_pv() macro. The
1013 blessing of the object by Perl can be handled by a typemap. An example
1014 typemap is shown at the end of this section.
1016 If the method is defined as static it will call the C++
1017 function using the class::method() syntax. If the method is not static
1018 the function will be called using the THIS-E<gt>method() syntax.
1020 The next examples will use the following C++ class.
1027 void set_blue( int );
1033 The XSUBs for the blue() and set_blue() methods are defined with the class
1034 name but the parameter for the object (THIS, or "self") is implicit and is
1041 color::set_blue( val )
1044 Both functions will expect an object as the first parameter. The xsubpp
1045 compiler will call that object C<THIS> and will use it to call the specified
1046 method. So in the C++ code the blue() and set_blue() methods will be called
1047 in the following manner.
1049 RETVAL = THIS->blue();
1051 THIS->set_blue( val );
1053 If the function's name is B<DESTROY> then the C++ C<delete> function will be
1054 called and C<THIS> will be given as its parameter.
1059 The C++ code will call C<delete>.
1063 If the function's name is B<new> then the C++ C<new> function will be called
1064 to create a dynamic C++ object. The XSUB will expect the class name, which
1065 will be kept in a variable called C<CLASS>, to be given as the first
1071 The C++ code will call C<new>.
1073 RETVAL = new color();
1075 The following is an example of a typemap that could be used for this C++
1082 # The Perl object is blessed into 'CLASS', which should be a
1083 # char* having the name of the package for the blessing.
1085 sv_setref_pv( $arg, CLASS, (void*)$var );
1089 if( sv_isobject($arg) && (SvTYPE(SvRV($arg)) == SVt_PVMG) )
1090 $var = ($type)SvIV((SV*)SvRV( $arg ));
1092 warn( \"${Package}::$func_name() -- $var is not a blessed SV reference\" );
1096 =head2 Interface Strategy
1098 When designing an interface between Perl and a C library a straight
1099 translation from C to XS is often sufficient. The interface will often be
1100 very C-like and occasionally nonintuitive, especially when the C function
1101 modifies one of its parameters. In cases where the programmer wishes to
1102 create a more Perl-like interface the following strategy may help to
1103 identify the more critical parts of the interface.
1105 Identify the C functions which modify their parameters. The XSUBs for
1106 these functions may be able to return lists to Perl, or may be
1107 candidates to return undef or an empty list in case of failure.
1109 Identify which values are used by only the C and XSUB functions
1110 themselves. If Perl does not need to access the contents of the value
1111 then it may not be necessary to provide a translation for that value
1114 Identify the pointers in the C function parameter lists and return
1115 values. Some pointers can be handled in XS with the & unary operator on
1116 the variable name while others will require the use of the * operator on
1117 the type name. In general it is easier to work with the & operator.
1119 Identify the structures used by the C functions. In many
1120 cases it may be helpful to use the T_PTROBJ typemap for
1121 these structures so they can be manipulated by Perl as
1124 =head2 Perl Objects And C Structures
1126 When dealing with C structures one should select either
1127 B<T_PTROBJ> or B<T_PTRREF> for the XS type. Both types are
1128 designed to handle pointers to complex objects. The
1129 T_PTRREF type will allow the Perl object to be unblessed
1130 while the T_PTROBJ type requires that the object be blessed.
1131 By using T_PTROBJ one can achieve a form of type-checking
1132 because the XSUB will attempt to verify that the Perl object
1133 is of the expected type.
1135 The following XS code shows the getnetconfigent() function which is used
1136 with ONC+ TIRPC. The getnetconfigent() function will return a pointer to a
1137 C structure and has the C prototype shown below. The example will
1138 demonstrate how the C pointer will become a Perl reference. Perl will
1139 consider this reference to be a pointer to a blessed object and will
1140 attempt to call a destructor for the object. A destructor will be
1141 provided in the XS source to free the memory used by getnetconfigent().
1142 Destructors in XS can be created by specifying an XSUB function whose name
1143 ends with the word B<DESTROY>. XS destructors can be used to free memory
1144 which may have been malloc'd by another XSUB.
1146 struct netconfig *getnetconfigent(const char *netid);
1148 A C<typedef> will be created for C<struct netconfig>. The Perl
1149 object will be blessed in a class matching the name of the C
1150 type, with the tag C<Ptr> appended, and the name should not
1151 have embedded spaces if it will be a Perl package name. The
1152 destructor will be placed in a class corresponding to the
1153 class of the object and the PREFIX keyword will be used to
1154 trim the name to the word DESTROY as Perl will expect.
1156 typedef struct netconfig Netconfig;
1158 MODULE = RPC PACKAGE = RPC
1161 getnetconfigent(netid)
1164 MODULE = RPC PACKAGE = NetconfigPtr PREFIX = rpcb_
1167 rpcb_DESTROY(netconf)
1170 printf("Now in NetconfigPtr::DESTROY\n");
1173 This example requires the following typemap entry. Consult the typemap
1174 section for more information about adding new typemaps for an extension.
1177 Netconfig * T_PTROBJ
1179 This example will be used with the following Perl statements.
1182 $netconf = getnetconfigent("udp");
1184 When Perl destroys the object referenced by $netconf it will send the
1185 object to the supplied XSUB DESTROY function. Perl cannot determine, and
1186 does not care, that this object is a C struct and not a Perl object. In
1187 this sense, there is no difference between the object created by the
1188 getnetconfigent() XSUB and an object created by a normal Perl subroutine.
1192 The typemap is a collection of code fragments which are used by the B<xsubpp>
1193 compiler to map C function parameters and values to Perl values. The
1194 typemap file may consist of three sections labeled C<TYPEMAP>, C<INPUT>, and
1195 C<OUTPUT>. The INPUT section tells the compiler how to translate Perl values
1196 into variables of certain C types. The OUTPUT section tells the compiler
1197 how to translate the values from certain C types into values Perl can
1198 understand. The TYPEMAP section tells the compiler which of the INPUT and
1199 OUTPUT code fragments should be used to map a given C type to a Perl value.
1200 Each of the sections of the typemap must be preceded by one of the TYPEMAP,
1201 INPUT, or OUTPUT keywords.
1203 The default typemap in the C<ext> directory of the Perl source contains many
1204 useful types which can be used by Perl extensions. Some extensions define
1205 additional typemaps which they keep in their own directory. These
1206 additional typemaps may reference INPUT and OUTPUT maps in the main
1207 typemap. The B<xsubpp> compiler will allow the extension's own typemap to
1208 override any mappings which are in the default typemap.
1210 Most extensions which require a custom typemap will need only the TYPEMAP
1211 section of the typemap file. The custom typemap used in the
1212 getnetconfigent() example shown earlier demonstrates what may be the typical
1213 use of extension typemaps. That typemap is used to equate a C structure
1214 with the T_PTROBJ typemap. The typemap used by getnetconfigent() is shown
1215 here. Note that the C type is separated from the XS type with a tab and
1216 that the C unary operator C<*> is considered to be a part of the C type name.
1219 Netconfig *<tab>T_PTROBJ
1221 Here's a more complicated example: suppose that you wanted C<struct
1222 netconfig> to be blessed into the class C<Net::Config>. One way to do
1223 this is to use underscores (_) to separate package names, as follows:
1225 typedef struct netconfig * Net_Config;
1227 And then provide a typemap entry C<T_PTROBJ_SPECIAL> that maps underscores to
1228 double-colons (::), and declare C<Net_Config> to be of that type:
1232 Net_Config T_PTROBJ_SPECIAL
1236 if (sv_derived_from($arg, \"${(my $ntt=$ntype)=~s/_/::/g;\$ntt}\")) {
1237 IV tmp = SvIV((SV*)SvRV($arg));
1241 croak(\"$var is not of type ${(my $ntt=$ntype)=~s/_/::/g;\$ntt}\")
1245 sv_setref_pv($arg, \"${(my $ntt=$ntype)=~s/_/::/g;\$ntt}\",
1248 The INPUT and OUTPUT sections substitute underscores for double-colons
1249 on the fly, giving the desired effect. This example demonstrates some
1250 of the power and versatility of the typemap facility.
1254 File C<RPC.xs>: Interface to some ONC+ RPC bind library functions.
1260 #include <rpc/rpc.h>
1262 typedef struct netconfig Netconfig;
1264 MODULE = RPC PACKAGE = RPC
1267 rpcb_gettime(host="localhost")
1272 ST(0) = sv_newmortal();
1273 if( rpcb_gettime( host, &timep ) )
1274 sv_setnv( ST(0), (double)timep );
1277 getnetconfigent(netid="udp")
1280 MODULE = RPC PACKAGE = NetconfigPtr PREFIX = rpcb_
1283 rpcb_DESTROY(netconf)
1286 printf("NetconfigPtr::DESTROY\n");
1289 File C<typemap>: Custom typemap for RPC.xs.
1292 Netconfig * T_PTROBJ
1294 File C<RPC.pm>: Perl module for the RPC extension.
1300 @ISA = qw(Exporter DynaLoader);
1301 @EXPORT = qw(rpcb_gettime getnetconfigent);
1306 File C<rpctest.pl>: Perl test program for the RPC extension.
1310 $netconf = getnetconfigent();
1311 $a = rpcb_gettime();
1312 print "time = $a\n";
1313 print "netconf = $netconf\n";
1315 $netconf = getnetconfigent("tcp");
1316 $a = rpcb_gettime("poplar");
1317 print "time = $a\n";
1318 print "netconf = $netconf\n";
1323 This document covers features supported by C<xsubpp> 1.935.
1327 Dean Roehrich <F<roehrich@cray.com>>