Commit | Line | Data |
---|---|---|
a0d0e21e LW |
1 | =head1 NAME |
2 | ||
3 | perlcall - Perl calling conventions from C | |
4 | ||
5 | =head1 DESCRIPTION | |
6 | ||
d1b91892 | 7 | The purpose of this document is to show you how to call Perl subroutines |
5f05dabc | 8 | directly from C, i.e., how to write I<callbacks>. |
a0d0e21e | 9 | |
d1b91892 AD |
10 | Apart from discussing the C interface provided by Perl for writing |
11 | callbacks the document uses a series of examples to show how the | |
12 | interface actually works in practice. In addition some techniques for | |
13 | coding callbacks are covered. | |
a0d0e21e | 14 | |
d1b91892 | 15 | Examples where callbacks are necessary include |
a0d0e21e LW |
16 | |
17 | =over 5 | |
18 | ||
d1b91892 | 19 | =item * An Error Handler |
a0d0e21e LW |
20 | |
21 | You have created an XSUB interface to an application's C API. | |
22 | ||
23 | A fairly common feature in applications is to allow you to define a C | |
d1b91892 AD |
24 | function that will be called whenever something nasty occurs. What we |
25 | would like is to be able to specify a Perl subroutine that will be | |
26 | called instead. | |
a0d0e21e | 27 | |
95e84303 | 28 | =item * An Event-Driven Program |
a0d0e21e | 29 | |
d1b91892 | 30 | The classic example of where callbacks are used is when writing an |
b0b54b5e | 31 | event driven program, such as for an X11 application. In this case |
184e9718 | 32 | you register functions to be called whenever specific events occur, |
5f05dabc | 33 | e.g., a mouse button is pressed, the cursor moves into a window or a |
d1b91892 | 34 | menu item is selected. |
a0d0e21e LW |
35 | |
36 | =back | |
37 | ||
d1b91892 AD |
38 | Although the techniques described here are applicable when embedding |
39 | Perl in a C program, this is not the primary goal of this document. | |
40 | There are other details that must be considered and are specific to | |
41 | embedding Perl. For details on embedding Perl in C refer to | |
42 | L<perlembed>. | |
a0d0e21e | 43 | |
d1b91892 | 44 | Before you launch yourself head first into the rest of this document, |
02f6dca1 FC |
45 | it would be a good idea to have read the following two documents--L<perlxs> |
46 | and L<perlguts>. | |
a0d0e21e | 47 | |
4929bf7b | 48 | =head1 THE CALL_ FUNCTIONS |
a0d0e21e | 49 | |
d1b91892 AD |
50 | Although this stuff is easier to explain using examples, you first need |
51 | be aware of a few important definitions. | |
a0d0e21e | 52 | |
d1b91892 AD |
53 | Perl has a number of C functions that allow you to call Perl |
54 | subroutines. They are | |
a0d0e21e | 55 | |
4358a253 SS |
56 | I32 call_sv(SV* sv, I32 flags); |
57 | I32 call_pv(char *subname, I32 flags); | |
58 | I32 call_method(char *methname, I32 flags); | |
5aaab254 | 59 | I32 call_argv(char *subname, I32 flags, char **argv); |
a0d0e21e | 60 | |
4929bf7b | 61 | The key function is I<call_sv>. All the other functions are |
d1b91892 | 62 | fairly simple wrappers which make it easier to call Perl subroutines in |
4929bf7b | 63 | special cases. At the end of the day they will all call I<call_sv> |
5f05dabc | 64 | to invoke the Perl subroutine. |
d1b91892 | 65 | |
4929bf7b | 66 | All the I<call_*> functions have a C<flags> parameter which is |
d1b91892 AD |
67 | used to pass a bit mask of options to Perl. This bit mask operates |
68 | identically for each of the functions. The settings available in the | |
d0554719 | 69 | bit mask are discussed in L</FLAG VALUES>. |
d1b91892 AD |
70 | |
71 | Each of the functions will now be discussed in turn. | |
72 | ||
73 | =over 5 | |
74 | ||
4929bf7b | 75 | =item call_sv |
d1b91892 | 76 | |
02f6dca1 | 77 | I<call_sv> takes two parameters. The first, C<sv>, is an SV*. |
d1b91892 AD |
78 | This allows you to specify the Perl subroutine to be called either as a |
79 | C string (which has first been converted to an SV) or a reference to a | |
d0554719 | 80 | subroutine. The section, L</Using call_sv>, shows how you can make |
4929bf7b | 81 | use of I<call_sv>. |
d1b91892 | 82 | |
4929bf7b | 83 | =item call_pv |
d1b91892 | 84 | |
4929bf7b | 85 | The function, I<call_pv>, is similar to I<call_sv> except it |
d1b91892 | 86 | expects its first parameter to be a C char* which identifies the Perl |
4929bf7b | 87 | subroutine you want to call, e.g., C<call_pv("fred", 0)>. If the |
d1b91892 | 88 | subroutine you want to call is in another package, just include the |
5f05dabc | 89 | package name in the string, e.g., C<"pkg::fred">. |
d1b91892 | 90 | |
4929bf7b | 91 | =item call_method |
d1b91892 | 92 | |
4929bf7b | 93 | The function I<call_method> is used to call a method from a Perl |
d1b91892 AD |
94 | class. The parameter C<methname> corresponds to the name of the method |
95 | to be called. Note that the class that the method belongs to is passed | |
96 | on the Perl stack rather than in the parameter list. This class can be | |
97 | either the name of the class (for a static method) or a reference to an | |
98 | object (for a virtual method). See L<perlobj> for more information on | |
d0554719 | 99 | static and virtual methods and L</Using call_method> for an example |
4929bf7b | 100 | of using I<call_method>. |
d1b91892 | 101 | |
4929bf7b | 102 | =item call_argv |
d1b91892 | 103 | |
4929bf7b | 104 | I<call_argv> calls the Perl subroutine specified by the C string |
d1b91892 | 105 | stored in the C<subname> parameter. It also takes the usual C<flags> |
02f6dca1 | 106 | parameter. The final parameter, C<argv>, consists of a NULL-terminated |
d1b91892 | 107 | list of C strings to be passed as parameters to the Perl subroutine. |
d0554719 | 108 | See L</Using call_argv>. |
d1b91892 AD |
109 | |
110 | =back | |
111 | ||
112 | All the functions return an integer. This is a count of the number of | |
113 | items returned by the Perl subroutine. The actual items returned by the | |
114 | subroutine are stored on the Perl stack. | |
115 | ||
116 | As a general rule you should I<always> check the return value from | |
117 | these functions. Even if you are expecting only a particular number of | |
118 | values to be returned from the Perl subroutine, there is nothing to | |
19799a22 | 119 | stop someone from doing something unexpected--don't say you haven't |
d1b91892 AD |
120 | been warned. |
121 | ||
122 | =head1 FLAG VALUES | |
123 | ||
0b06a753 NT |
124 | The C<flags> parameter in all the I<call_*> functions is one of G_VOID, |
125 | G_SCALAR, or G_ARRAY, which indicate the call context, OR'ed together | |
126 | with a bit mask of any combination of the other G_* symbols defined below. | |
d1b91892 | 127 | |
54310121 | 128 | =head2 G_VOID |
129 | ||
130 | Calls the Perl subroutine in a void context. | |
131 | ||
132 | This flag has 2 effects: | |
133 | ||
134 | =over 5 | |
135 | ||
136 | =item 1. | |
137 | ||
138 | It indicates to the subroutine being called that it is executing in | |
139 | a void context (if it executes I<wantarray> the result will be the | |
140 | undefined value). | |
141 | ||
142 | =item 2. | |
143 | ||
144 | It ensures that nothing is actually returned from the subroutine. | |
145 | ||
146 | =back | |
147 | ||
4929bf7b | 148 | The value returned by the I<call_*> function indicates how many |
02f6dca1 | 149 | items have been returned by the Perl subroutine--in this case it will |
54310121 | 150 | be 0. |
151 | ||
152 | ||
d1b91892 AD |
153 | =head2 G_SCALAR |
154 | ||
155 | Calls the Perl subroutine in a scalar context. This is the default | |
4929bf7b | 156 | context flag setting for all the I<call_*> functions. |
d1b91892 | 157 | |
184e9718 | 158 | This flag has 2 effects: |
d1b91892 AD |
159 | |
160 | =over 5 | |
161 | ||
162 | =item 1. | |
163 | ||
184e9718 | 164 | It indicates to the subroutine being called that it is executing in a |
d1b91892 | 165 | scalar context (if it executes I<wantarray> the result will be false). |
a0d0e21e | 166 | |
d1b91892 AD |
167 | =item 2. |
168 | ||
184e9718 | 169 | It ensures that only a scalar is actually returned from the subroutine. |
d1b91892 AD |
170 | The subroutine can, of course, ignore the I<wantarray> and return a |
171 | list anyway. If so, then only the last element of the list will be | |
172 | returned. | |
173 | ||
174 | =back | |
175 | ||
4929bf7b | 176 | The value returned by the I<call_*> function indicates how many |
d1b91892 AD |
177 | items have been returned by the Perl subroutine - in this case it will |
178 | be either 0 or 1. | |
a0d0e21e | 179 | |
d1b91892 | 180 | If 0, then you have specified the G_DISCARD flag. |
a0d0e21e | 181 | |
d1b91892 | 182 | If 1, then the item actually returned by the Perl subroutine will be |
d0554719 | 183 | stored on the Perl stack - the section L</Returning a Scalar> shows how |
d1b91892 AD |
184 | to access this value on the stack. Remember that regardless of how |
185 | many items the Perl subroutine returns, only the last one will be | |
186 | accessible from the stack - think of the case where only one value is | |
187 | returned as being a list with only one element. Any other items that | |
188 | were returned will not exist by the time control returns from the | |
d0554719 TC |
189 | I<call_*> function. The section L</Returning a List in Scalar |
190 | Context> shows an example of this behavior. | |
a0d0e21e | 191 | |
a0d0e21e | 192 | |
d1b91892 | 193 | =head2 G_ARRAY |
a0d0e21e | 194 | |
d1b91892 | 195 | Calls the Perl subroutine in a list context. |
a0d0e21e | 196 | |
184e9718 | 197 | As with G_SCALAR, this flag has 2 effects: |
a0d0e21e LW |
198 | |
199 | =over 5 | |
200 | ||
d1b91892 AD |
201 | =item 1. |
202 | ||
90fdbbb7 DC |
203 | It indicates to the subroutine being called that it is executing in a |
204 | list context (if it executes I<wantarray> the result will be true). | |
a0d0e21e | 205 | |
d1b91892 | 206 | =item 2. |
a0d0e21e | 207 | |
184e9718 | 208 | It ensures that all items returned from the subroutine will be |
4929bf7b | 209 | accessible when control returns from the I<call_*> function. |
a0d0e21e | 210 | |
d1b91892 | 211 | =back |
a0d0e21e | 212 | |
4929bf7b | 213 | The value returned by the I<call_*> function indicates how many |
d1b91892 | 214 | items have been returned by the Perl subroutine. |
a0d0e21e | 215 | |
184e9718 | 216 | If 0, then you have specified the G_DISCARD flag. |
a0d0e21e | 217 | |
d1b91892 AD |
218 | If not 0, then it will be a count of the number of items returned by |
219 | the subroutine. These items will be stored on the Perl stack. The | |
d0554719 | 220 | section L</Returning a List of Values> gives an example of using the |
d1b91892 AD |
221 | G_ARRAY flag and the mechanics of accessing the returned items from the |
222 | Perl stack. | |
a0d0e21e | 223 | |
d1b91892 | 224 | =head2 G_DISCARD |
a0d0e21e | 225 | |
4929bf7b | 226 | By default, the I<call_*> functions place the items returned from |
d1b91892 AD |
227 | by the Perl subroutine on the stack. If you are not interested in |
228 | these items, then setting this flag will make Perl get rid of them | |
229 | automatically for you. Note that it is still possible to indicate a | |
230 | context to the Perl subroutine by using either G_SCALAR or G_ARRAY. | |
a0d0e21e | 231 | |
d1b91892 | 232 | If you do not set this flag then it is I<very> important that you make |
5f05dabc | 233 | sure that any temporaries (i.e., parameters passed to the Perl |
d1b91892 | 234 | subroutine and values returned from the subroutine) are disposed of |
d0554719 TC |
235 | yourself. The section L</Returning a Scalar> gives details of how to |
236 | dispose of these temporaries explicitly and the section L</Using Perl to | |
237 | Dispose of Temporaries> discusses the specific circumstances where you | |
d1b91892 | 238 | can ignore the problem and let Perl deal with it for you. |
a0d0e21e | 239 | |
d1b91892 | 240 | =head2 G_NOARGS |
a0d0e21e | 241 | |
4929bf7b | 242 | Whenever a Perl subroutine is called using one of the I<call_*> |
d1b91892 AD |
243 | functions, it is assumed by default that parameters are to be passed to |
244 | the subroutine. If you are not passing any parameters to the Perl | |
245 | subroutine, you can save a bit of time by setting this flag. It has | |
246 | the effect of not creating the C<@_> array for the Perl subroutine. | |
a0d0e21e | 247 | |
d1b91892 AD |
248 | Although the functionality provided by this flag may seem |
249 | straightforward, it should be used only if there is a good reason to do | |
02f6dca1 | 250 | so. The reason for being cautious is that, even if you have specified |
d1b91892 AD |
251 | the G_NOARGS flag, it is still possible for the Perl subroutine that |
252 | has been called to think that you have passed it parameters. | |
a0d0e21e | 253 | |
d1b91892 AD |
254 | In fact, what can happen is that the Perl subroutine you have called |
255 | can access the C<@_> array from a previous Perl subroutine. This will | |
4929bf7b | 256 | occur when the code that is executing the I<call_*> function has |
d1b91892 AD |
257 | itself been called from another Perl subroutine. The code below |
258 | illustrates this | |
a0d0e21e | 259 | |
84f709e7 JH |
260 | sub fred |
261 | { print "@_\n" } | |
a0d0e21e | 262 | |
84f709e7 JH |
263 | sub joe |
264 | { &fred } | |
a0d0e21e | 265 | |
4358a253 | 266 | &joe(1,2,3); |
a0d0e21e LW |
267 | |
268 | This will print | |
269 | ||
d1b91892 AD |
270 | 1 2 3 |
271 | ||
272 | What has happened is that C<fred> accesses the C<@_> array which | |
273 | belongs to C<joe>. | |
a0d0e21e | 274 | |
a0d0e21e | 275 | |
54310121 | 276 | =head2 G_EVAL |
a0d0e21e | 277 | |
d1b91892 | 278 | It is possible for the Perl subroutine you are calling to terminate |
5f05dabc | 279 | abnormally, e.g., by calling I<die> explicitly or by not actually |
268118b2 HS |
280 | existing. By default, when either of these events occurs, the |
281 | process will terminate immediately. If you want to trap this | |
d1b91892 AD |
282 | type of event, specify the G_EVAL flag. It will put an I<eval { }> |
283 | around the subroutine call. | |
a0d0e21e | 284 | |
4929bf7b | 285 | Whenever control returns from the I<call_*> function you need to |
d1b91892 AD |
286 | check the C<$@> variable as you would in a normal Perl script. |
287 | ||
4929bf7b | 288 | The value returned from the I<call_*> function is dependent on |
d1b91892 | 289 | what other flags have been specified and whether an error has |
184e9718 | 290 | occurred. Here are all the different cases that can occur: |
d1b91892 AD |
291 | |
292 | =over 5 | |
293 | ||
294 | =item * | |
295 | ||
4929bf7b | 296 | If the I<call_*> function returns normally, then the value |
d1b91892 AD |
297 | returned is as specified in the previous sections. |
298 | ||
299 | =item * | |
300 | ||
301 | If G_DISCARD is specified, the return value will always be 0. | |
302 | ||
303 | =item * | |
304 | ||
305 | If G_ARRAY is specified I<and> an error has occurred, the return value | |
306 | will always be 0. | |
307 | ||
308 | =item * | |
a0d0e21e | 309 | |
d1b91892 AD |
310 | If G_SCALAR is specified I<and> an error has occurred, the return value |
311 | will be 1 and the value on the top of the stack will be I<undef>. This | |
312 | means that if you have already detected the error by checking C<$@> and | |
313 | you want the program to continue, you must remember to pop the I<undef> | |
314 | from the stack. | |
a0d0e21e LW |
315 | |
316 | =back | |
317 | ||
d0554719 | 318 | See L</Using G_EVAL> for details on using G_EVAL. |
d1b91892 | 319 | |
c07a80fd | 320 | =head2 G_KEEPERR |
321 | ||
7ce09284 Z |
322 | Using the G_EVAL flag described above will always set C<$@>: clearing |
323 | it if there was no error, and setting it to describe the error if there | |
324 | was an error in the called code. This is what you want if your intention | |
325 | is to handle possible errors, but sometimes you just want to trap errors | |
326 | and stop them interfering with the rest of the program. | |
327 | ||
328 | This scenario will mostly be applicable to code that is meant to be called | |
329 | from within destructors, asynchronous callbacks, and signal handlers. | |
330 | In such situations, where the code being called has little relation to the | |
331 | surrounding dynamic context, the main program needs to be insulated from | |
332 | errors in the called code, even if they can't be handled intelligently. | |
333 | It may also be useful to do this with code for C<__DIE__> or C<__WARN__> | |
334 | hooks, and C<tie> functions. | |
c07a80fd | 335 | |
336 | The G_KEEPERR flag is meant to be used in conjunction with G_EVAL in | |
be064c4a DM |
337 | I<call_*> functions that are used to implement such code, or with |
338 | C<eval_sv>. This flag has no effect on the C<call_*> functions when | |
339 | G_EVAL is not used. | |
c07a80fd | 340 | |
7ce09284 Z |
341 | When G_KEEPERR is used, any error in the called code will terminate the |
342 | call as usual, and the error will not propagate beyond the call (as usual | |
343 | for G_EVAL), but it will not go into C<$@>. Instead the error will be | |
344 | converted into a warning, prefixed with the string "\t(in cleanup)". | |
345 | This can be disabled using C<no warnings 'misc'>. If there is no error, | |
346 | C<$@> will not be cleared. | |
c07a80fd | 347 | |
be064c4a DM |
348 | Note that the G_KEEPERR flag does not propagate into inner evals; these |
349 | may still set C<$@>. | |
350 | ||
c07a80fd | 351 | The G_KEEPERR flag was introduced in Perl version 5.002. |
352 | ||
d0554719 | 353 | See L</Using G_KEEPERR> for an example of a situation that warrants the |
c07a80fd | 354 | use of this flag. |
355 | ||
54310121 | 356 | =head2 Determining the Context |
d1b91892 AD |
357 | |
358 | As mentioned above, you can determine the context of the currently | |
54310121 | 359 | executing subroutine in Perl with I<wantarray>. The equivalent test |
360 | can be made in C by using the C<GIMME_V> macro, which returns | |
90fdbbb7 | 361 | C<G_ARRAY> if you have been called in a list context, C<G_SCALAR> if |
02f6dca1 | 362 | in a scalar context, or C<G_VOID> if in a void context (i.e., the |
54310121 | 363 | return value will not be used). An older version of this macro is |
364 | called C<GIMME>; in a void context it returns C<G_SCALAR> instead of | |
365 | C<G_VOID>. An example of using the C<GIMME_V> macro is shown in | |
d0554719 | 366 | section L</Using GIMME_V>. |
d1b91892 | 367 | |
a0d0e21e LW |
368 | =head1 EXAMPLES |
369 | ||
02f6dca1 | 370 | Enough of the definition talk! Let's have a few examples. |
a0d0e21e | 371 | |
d1b91892 AD |
372 | Perl provides many macros to assist in accessing the Perl stack. |
373 | Wherever possible, these macros should always be used when interfacing | |
5f05dabc | 374 | to Perl internals. We hope this should make the code less vulnerable |
d1b91892 | 375 | to any changes made to Perl in the future. |
a0d0e21e | 376 | |
d1b91892 | 377 | Another point worth noting is that in the first series of examples I |
4929bf7b | 378 | have made use of only the I<call_pv> function. This has been done |
d1b91892 | 379 | to keep the code simpler and ease you into the topic. Wherever |
4929bf7b GS |
380 | possible, if the choice is between using I<call_pv> and |
381 | I<call_sv>, you should always try to use I<call_sv>. See | |
d0554719 | 382 | L</Using call_sv> for details. |
a0d0e21e | 383 | |
02f6dca1 | 384 | =head2 No Parameters, Nothing Returned |
a0d0e21e | 385 | |
d1b91892 AD |
386 | This first trivial example will call a Perl subroutine, I<PrintUID>, to |
387 | print out the UID of the process. | |
a0d0e21e | 388 | |
84f709e7 JH |
389 | sub PrintUID |
390 | { | |
4358a253 | 391 | print "UID is $<\n"; |
a0d0e21e LW |
392 | } |
393 | ||
d1b91892 | 394 | and here is a C function to call it |
a0d0e21e | 395 | |
d1b91892 | 396 | static void |
a0d0e21e LW |
397 | call_PrintUID() |
398 | { | |
4358a253 | 399 | dSP; |
a0d0e21e | 400 | |
4358a253 SS |
401 | PUSHMARK(SP); |
402 | call_pv("PrintUID", G_DISCARD|G_NOARGS); | |
a0d0e21e LW |
403 | } |
404 | ||
02f6dca1 | 405 | Simple, eh? |
a0d0e21e | 406 | |
02f6dca1 | 407 | A few points to note about this example: |
a0d0e21e LW |
408 | |
409 | =over 5 | |
410 | ||
d1b91892 | 411 | =item 1. |
a0d0e21e | 412 | |
924508f0 | 413 | Ignore C<dSP> and C<PUSHMARK(SP)> for now. They will be discussed in |
d1b91892 | 414 | the next example. |
a0d0e21e LW |
415 | |
416 | =item 2. | |
417 | ||
d1b91892 AD |
418 | We aren't passing any parameters to I<PrintUID> so G_NOARGS can be |
419 | specified. | |
a0d0e21e | 420 | |
d1b91892 | 421 | =item 3. |
a0d0e21e LW |
422 | |
423 | We aren't interested in anything returned from I<PrintUID>, so | |
5f05dabc | 424 | G_DISCARD is specified. Even if I<PrintUID> was changed to |
a0d0e21e | 425 | return some value(s), having specified G_DISCARD will mean that they |
4929bf7b | 426 | will be wiped by the time control returns from I<call_pv>. |
a0d0e21e | 427 | |
d1b91892 | 428 | =item 4. |
a0d0e21e | 429 | |
4929bf7b | 430 | As I<call_pv> is being used, the Perl subroutine is specified as a |
d1b91892 AD |
431 | C string. In this case the subroutine name has been 'hard-wired' into the |
432 | code. | |
a0d0e21e LW |
433 | |
434 | =item 5. | |
435 | ||
d1b91892 | 436 | Because we specified G_DISCARD, it is not necessary to check the value |
4929bf7b | 437 | returned from I<call_pv>. It will always be 0. |
a0d0e21e LW |
438 | |
439 | =back | |
440 | ||
d1b91892 | 441 | =head2 Passing Parameters |
a0d0e21e | 442 | |
d1b91892 | 443 | Now let's make a slightly more complex example. This time we want to |
19799a22 GS |
444 | call a Perl subroutine, C<LeftString>, which will take 2 parameters--a |
445 | string ($s) and an integer ($n). The subroutine will simply | |
446 | print the first $n characters of the string. | |
a0d0e21e | 447 | |
02f6dca1 | 448 | So the Perl subroutine would look like this: |
a0d0e21e | 449 | |
84f709e7 JH |
450 | sub LeftString |
451 | { | |
4358a253 SS |
452 | my($s, $n) = @_; |
453 | print substr($s, 0, $n), "\n"; | |
a0d0e21e LW |
454 | } |
455 | ||
02f6dca1 | 456 | The C function required to call I<LeftString> would look like this: |
a0d0e21e LW |
457 | |
458 | static void | |
459 | call_LeftString(a, b) | |
4358a253 SS |
460 | char * a; |
461 | int b; | |
a0d0e21e | 462 | { |
4358a253 | 463 | dSP; |
a0d0e21e | 464 | |
4358a253 SS |
465 | ENTER; |
466 | SAVETMPS; | |
9b6570b4 | 467 | |
4358a253 | 468 | PUSHMARK(SP); |
d0554719 TC |
469 | EXTEND(SP, 2); |
470 | PUSHs(sv_2mortal(newSVpv(a, 0))); | |
471 | PUSHs(sv_2mortal(newSViv(b))); | |
4358a253 | 472 | PUTBACK; |
a0d0e21e | 473 | |
4929bf7b | 474 | call_pv("LeftString", G_DISCARD); |
9b6570b4 | 475 | |
4358a253 SS |
476 | FREETMPS; |
477 | LEAVE; | |
a0d0e21e LW |
478 | } |
479 | ||
a0d0e21e LW |
480 | Here are a few notes on the C function I<call_LeftString>. |
481 | ||
482 | =over 5 | |
483 | ||
d1b91892 | 484 | =item 1. |
a0d0e21e | 485 | |
d1b91892 AD |
486 | Parameters are passed to the Perl subroutine using the Perl stack. |
487 | This is the purpose of the code beginning with the line C<dSP> and | |
1e62ac33 | 488 | ending with the line C<PUTBACK>. The C<dSP> declares a local copy |
924508f0 GS |
489 | of the stack pointer. This local copy should B<always> be accessed |
490 | as C<SP>. | |
a0d0e21e | 491 | |
d1b91892 | 492 | =item 2. |
a0d0e21e LW |
493 | |
494 | If you are going to put something onto the Perl stack, you need to know | |
19799a22 | 495 | where to put it. This is the purpose of the macro C<dSP>--it declares |
d1b91892 | 496 | and initializes a I<local> copy of the Perl stack pointer. |
a0d0e21e LW |
497 | |
498 | All the other macros which will be used in this example require you to | |
d1b91892 | 499 | have used this macro. |
a0d0e21e | 500 | |
d1b91892 AD |
501 | The exception to this rule is if you are calling a Perl subroutine |
502 | directly from an XSUB function. In this case it is not necessary to | |
19799a22 | 503 | use the C<dSP> macro explicitly--it will be declared for you |
d1b91892 | 504 | automatically. |
a0d0e21e | 505 | |
d1b91892 | 506 | =item 3. |
a0d0e21e LW |
507 | |
508 | Any parameters to be pushed onto the stack should be bracketed by the | |
d1b91892 | 509 | C<PUSHMARK> and C<PUTBACK> macros. The purpose of these two macros, in |
5f05dabc | 510 | this context, is to count the number of parameters you are |
511 | pushing automatically. Then whenever Perl is creating the C<@_> array for the | |
d1b91892 AD |
512 | subroutine, it knows how big to make it. |
513 | ||
514 | The C<PUSHMARK> macro tells Perl to make a mental note of the current | |
515 | stack pointer. Even if you aren't passing any parameters (like the | |
d0554719 | 516 | example shown in the section L</No Parameters, Nothing Returned>) you |
d1b91892 | 517 | must still call the C<PUSHMARK> macro before you can call any of the |
4929bf7b | 518 | I<call_*> functions--Perl still needs to know that there are no |
d1b91892 AD |
519 | parameters. |
520 | ||
521 | The C<PUTBACK> macro sets the global copy of the stack pointer to be | |
02f6dca1 | 522 | the same as our local copy. If we didn't do this, I<call_pv> |
19799a22 | 523 | wouldn't know where the two parameters we pushed were--remember that |
d1b91892 AD |
524 | up to now all the stack pointer manipulation we have done is with our |
525 | local copy, I<not> the global copy. | |
526 | ||
527 | =item 4. | |
528 | ||
d0554719 TC |
529 | Next, we come to EXTEND and PUSHs. This is where the parameters |
530 | actually get pushed onto the stack. In this case we are pushing a | |
531 | string and an integer. | |
532 | ||
533 | Alternatively you can use the XPUSHs() macro, which combines a | |
534 | C<EXTEND(SP, 1)> and C<PUSHs()>. This is less efficient if you're | |
535 | pushing multiple values. | |
a0d0e21e | 536 | |
54310121 | 537 | See L<perlguts/"XSUBs and the Argument Stack"> for details |
d0554719 | 538 | on how the PUSH macros work. |
a0d0e21e | 539 | |
087fe227 | 540 | =item 5. |
a0d0e21e | 541 | |
9b6570b4 AB |
542 | Because we created temporary values (by means of sv_2mortal() calls) |
543 | we will have to tidy up the Perl stack and dispose of mortal SVs. | |
544 | ||
545 | This is the purpose of | |
546 | ||
4358a253 SS |
547 | ENTER; |
548 | SAVETMPS; | |
9b6570b4 AB |
549 | |
550 | at the start of the function, and | |
551 | ||
4358a253 SS |
552 | FREETMPS; |
553 | LEAVE; | |
9b6570b4 AB |
554 | |
555 | at the end. The C<ENTER>/C<SAVETMPS> pair creates a boundary for any | |
556 | temporaries we create. This means that the temporaries we get rid of | |
557 | will be limited to those which were created after these calls. | |
558 | ||
559 | The C<FREETMPS>/C<LEAVE> pair will get rid of any values returned by | |
560 | the Perl subroutine (see next example), plus it will also dump the | |
561 | mortal SVs we have created. Having C<ENTER>/C<SAVETMPS> at the | |
562 | beginning of the code makes sure that no other mortals are destroyed. | |
563 | ||
02f6dca1 | 564 | Think of these macros as working a bit like C<{> and C<}> in Perl |
9b6570b4 AB |
565 | to limit the scope of local variables. |
566 | ||
d0554719 | 567 | See the section L</Using Perl to Dispose of Temporaries> for details of |
9b6570b4 AB |
568 | an alternative to using these macros. |
569 | ||
087fe227 | 570 | =item 6. |
9b6570b4 | 571 | |
087fe227 JA |
572 | Finally, I<LeftString> can now be called via the I<call_pv> function. |
573 | The only flag specified this time is G_DISCARD. Because we are passing | |
574 | 2 parameters to the Perl subroutine this time, we have not specified | |
575 | G_NOARGS. | |
a0d0e21e LW |
576 | |
577 | =back | |
578 | ||
d1b91892 | 579 | =head2 Returning a Scalar |
a0d0e21e | 580 | |
d1b91892 AD |
581 | Now for an example of dealing with the items returned from a Perl |
582 | subroutine. | |
a0d0e21e | 583 | |
5f05dabc | 584 | Here is a Perl subroutine, I<Adder>, that takes 2 integer parameters |
d1b91892 | 585 | and simply returns their sum. |
a0d0e21e | 586 | |
84f709e7 JH |
587 | sub Adder |
588 | { | |
4358a253 SS |
589 | my($a, $b) = @_; |
590 | $a + $b; | |
a0d0e21e LW |
591 | } |
592 | ||
5f05dabc | 593 | Because we are now concerned with the return value from I<Adder>, the C |
d1b91892 | 594 | function required to call it is now a bit more complex. |
a0d0e21e LW |
595 | |
596 | static void | |
597 | call_Adder(a, b) | |
4358a253 SS |
598 | int a; |
599 | int b; | |
a0d0e21e | 600 | { |
4358a253 SS |
601 | dSP; |
602 | int count; | |
a0d0e21e | 603 | |
4358a253 | 604 | ENTER; |
a0d0e21e LW |
605 | SAVETMPS; |
606 | ||
4358a253 | 607 | PUSHMARK(SP); |
d0554719 TC |
608 | EXTEND(SP, 2); |
609 | PUSHs(sv_2mortal(newSViv(a))); | |
610 | PUSHs(sv_2mortal(newSViv(b))); | |
4358a253 | 611 | PUTBACK; |
a0d0e21e | 612 | |
4929bf7b | 613 | count = call_pv("Adder", G_SCALAR); |
a0d0e21e | 614 | |
4358a253 | 615 | SPAGAIN; |
a0d0e21e | 616 | |
d1b91892 | 617 | if (count != 1) |
4358a253 | 618 | croak("Big trouble\n"); |
a0d0e21e | 619 | |
4358a253 | 620 | printf ("The sum of %d and %d is %d\n", a, b, POPi); |
a0d0e21e | 621 | |
4358a253 SS |
622 | PUTBACK; |
623 | FREETMPS; | |
624 | LEAVE; | |
a0d0e21e LW |
625 | } |
626 | ||
a0d0e21e LW |
627 | Points to note this time are |
628 | ||
629 | =over 5 | |
630 | ||
54310121 | 631 | =item 1. |
a0d0e21e | 632 | |
02f6dca1 | 633 | The only flag specified this time was G_SCALAR. That means that the C<@_> |
d1b91892 | 634 | array will be created and that the value returned by I<Adder> will |
4929bf7b | 635 | still exist after the call to I<call_pv>. |
a0d0e21e | 636 | |
a0d0e21e LW |
637 | =item 2. |
638 | ||
a0d0e21e LW |
639 | The purpose of the macro C<SPAGAIN> is to refresh the local copy of the |
640 | stack pointer. This is necessary because it is possible that the memory | |
ed874981 | 641 | allocated to the Perl stack has been reallocated during the |
4929bf7b | 642 | I<call_pv> call. |
a0d0e21e | 643 | |
d1b91892 | 644 | If you are making use of the Perl stack pointer in your code you must |
54310121 | 645 | always refresh the local copy using SPAGAIN whenever you make use |
4929bf7b | 646 | of the I<call_*> functions or any other Perl internal function. |
a0d0e21e | 647 | |
9b6570b4 | 648 | =item 3. |
a0d0e21e | 649 | |
d1b91892 | 650 | Although only a single value was expected to be returned from I<Adder>, |
4929bf7b | 651 | it is still good practice to check the return code from I<call_pv> |
d1b91892 | 652 | anyway. |
a0d0e21e | 653 | |
d1b91892 AD |
654 | Expecting a single value is not quite the same as knowing that there |
655 | will be one. If someone modified I<Adder> to return a list and we | |
656 | didn't check for that possibility and take appropriate action the Perl | |
657 | stack would end up in an inconsistent state. That is something you | |
5f05dabc | 658 | I<really> don't want to happen ever. |
a0d0e21e | 659 | |
9b6570b4 | 660 | =item 4. |
a0d0e21e | 661 | |
d1b91892 AD |
662 | The C<POPi> macro is used here to pop the return value from the stack. |
663 | In this case we wanted an integer, so C<POPi> was used. | |
a0d0e21e LW |
664 | |
665 | ||
d1b91892 AD |
666 | Here is the complete list of POP macros available, along with the types |
667 | they return. | |
a0d0e21e | 668 | |
d1b91892 | 669 | POPs SV |
d0554719 TC |
670 | POPp pointer (PV) |
671 | POPpbytex pointer to bytes (PV) | |
672 | POPn double (NV) | |
673 | POPi integer (IV) | |
674 | POPu unsigned integer (UV) | |
d1b91892 | 675 | POPl long |
d0554719 TC |
676 | POPul unsigned long |
677 | ||
678 | Since these macros have side-effects don't use them as arguments to | |
679 | macros that may evaluate their argument several times, for example: | |
680 | ||
681 | /* Bad idea, don't do this */ | |
682 | STRLEN len; | |
683 | const char *s = SvPV(POPs, len); | |
684 | ||
685 | Instead, use a temporary: | |
686 | ||
687 | STRLEN len; | |
688 | SV *sv = POPs; | |
689 | const char *s = SvPV(sv, len); | |
690 | ||
691 | or a macro that guarantees it will evaluate its arguments only once: | |
692 | ||
693 | STRLEN len; | |
694 | const char *s = SvPVx(POPs, len); | |
a0d0e21e | 695 | |
9b6570b4 | 696 | =item 5. |
a0d0e21e | 697 | |
d1b91892 AD |
698 | The final C<PUTBACK> is used to leave the Perl stack in a consistent |
699 | state before exiting the function. This is necessary because when we | |
700 | popped the return value from the stack with C<POPi> it updated only our | |
701 | local copy of the stack pointer. Remember, C<PUTBACK> sets the global | |
702 | stack pointer to be the same as our local copy. | |
a0d0e21e LW |
703 | |
704 | =back | |
705 | ||
706 | ||
02f6dca1 | 707 | =head2 Returning a List of Values |
a0d0e21e | 708 | |
d1b91892 AD |
709 | Now, let's extend the previous example to return both the sum of the |
710 | parameters and the difference. | |
a0d0e21e | 711 | |
d1b91892 | 712 | Here is the Perl subroutine |
a0d0e21e | 713 | |
84f709e7 JH |
714 | sub AddSubtract |
715 | { | |
4358a253 SS |
716 | my($a, $b) = @_; |
717 | ($a+$b, $a-$b); | |
a0d0e21e LW |
718 | } |
719 | ||
a0d0e21e LW |
720 | and this is the C function |
721 | ||
722 | static void | |
723 | call_AddSubtract(a, b) | |
4358a253 SS |
724 | int a; |
725 | int b; | |
a0d0e21e | 726 | { |
4358a253 SS |
727 | dSP; |
728 | int count; | |
a0d0e21e | 729 | |
4358a253 | 730 | ENTER; |
a0d0e21e LW |
731 | SAVETMPS; |
732 | ||
4358a253 | 733 | PUSHMARK(SP); |
d0554719 TC |
734 | EXTEND(SP, 2); |
735 | PUSHs(sv_2mortal(newSViv(a))); | |
736 | PUSHs(sv_2mortal(newSViv(b))); | |
4358a253 | 737 | PUTBACK; |
a0d0e21e | 738 | |
4929bf7b | 739 | count = call_pv("AddSubtract", G_ARRAY); |
a0d0e21e | 740 | |
4358a253 | 741 | SPAGAIN; |
a0d0e21e | 742 | |
d1b91892 | 743 | if (count != 2) |
4358a253 | 744 | croak("Big trouble\n"); |
a0d0e21e | 745 | |
4358a253 SS |
746 | printf ("%d - %d = %d\n", a, b, POPi); |
747 | printf ("%d + %d = %d\n", a, b, POPi); | |
a0d0e21e | 748 | |
4358a253 SS |
749 | PUTBACK; |
750 | FREETMPS; | |
751 | LEAVE; | |
a0d0e21e LW |
752 | } |
753 | ||
d1b91892 AD |
754 | If I<call_AddSubtract> is called like this |
755 | ||
4358a253 | 756 | call_AddSubtract(7, 4); |
d1b91892 AD |
757 | |
758 | then here is the output | |
759 | ||
760 | 7 - 4 = 3 | |
761 | 7 + 4 = 11 | |
a0d0e21e LW |
762 | |
763 | Notes | |
764 | ||
765 | =over 5 | |
766 | ||
767 | =item 1. | |
768 | ||
90fdbbb7 | 769 | We wanted list context, so G_ARRAY was used. |
a0d0e21e LW |
770 | |
771 | =item 2. | |
772 | ||
d1b91892 AD |
773 | Not surprisingly C<POPi> is used twice this time because we were |
774 | retrieving 2 values from the stack. The important thing to note is that | |
775 | when using the C<POP*> macros they come off the stack in I<reverse> | |
776 | order. | |
a0d0e21e LW |
777 | |
778 | =back | |
779 | ||
d0554719 | 780 | =head2 Returning a List in Scalar Context |
d1b91892 AD |
781 | |
782 | Say the Perl subroutine in the previous section was called in a scalar | |
783 | context, like this | |
784 | ||
785 | static void | |
786 | call_AddSubScalar(a, b) | |
4358a253 SS |
787 | int a; |
788 | int b; | |
d1b91892 | 789 | { |
4358a253 SS |
790 | dSP; |
791 | int count; | |
792 | int i; | |
d1b91892 | 793 | |
4358a253 | 794 | ENTER; |
d1b91892 AD |
795 | SAVETMPS; |
796 | ||
4358a253 | 797 | PUSHMARK(SP); |
d0554719 TC |
798 | EXTEND(SP, 2); |
799 | PUSHs(sv_2mortal(newSViv(a))); | |
800 | PUSHs(sv_2mortal(newSViv(b))); | |
4358a253 | 801 | PUTBACK; |
d1b91892 | 802 | |
4929bf7b | 803 | count = call_pv("AddSubtract", G_SCALAR); |
d1b91892 | 804 | |
4358a253 | 805 | SPAGAIN; |
d1b91892 | 806 | |
4358a253 | 807 | printf ("Items Returned = %d\n", count); |
d1b91892 | 808 | |
4358a253 SS |
809 | for (i = 1; i <= count; ++i) |
810 | printf ("Value %d = %d\n", i, POPi); | |
d1b91892 | 811 | |
4358a253 SS |
812 | PUTBACK; |
813 | FREETMPS; | |
814 | LEAVE; | |
d1b91892 AD |
815 | } |
816 | ||
817 | The other modification made is that I<call_AddSubScalar> will print the | |
818 | number of items returned from the Perl subroutine and their value (for | |
819 | simplicity it assumes that they are integer). So if | |
820 | I<call_AddSubScalar> is called | |
821 | ||
4358a253 | 822 | call_AddSubScalar(7, 4); |
d1b91892 AD |
823 | |
824 | then the output will be | |
825 | ||
826 | Items Returned = 1 | |
827 | Value 1 = 3 | |
828 | ||
829 | In this case the main point to note is that only the last item in the | |
48052fe5 | 830 | list is returned from the subroutine. I<AddSubtract> actually made it back to |
d1b91892 AD |
831 | I<call_AddSubScalar>. |
832 | ||
833 | ||
02f6dca1 | 834 | =head2 Returning Data from Perl via the Parameter List |
a0d0e21e | 835 | |
48052fe5 FC |
836 | It is also possible to return values directly via the parameter |
837 | list--whether it is actually desirable to do it is another matter entirely. | |
a0d0e21e | 838 | |
d1b91892 AD |
839 | The Perl subroutine, I<Inc>, below takes 2 parameters and increments |
840 | each directly. | |
a0d0e21e | 841 | |
84f709e7 JH |
842 | sub Inc |
843 | { | |
4358a253 SS |
844 | ++ $_[0]; |
845 | ++ $_[1]; | |
a0d0e21e LW |
846 | } |
847 | ||
848 | and here is a C function to call it. | |
849 | ||
850 | static void | |
851 | call_Inc(a, b) | |
4358a253 SS |
852 | int a; |
853 | int b; | |
a0d0e21e | 854 | { |
4358a253 SS |
855 | dSP; |
856 | int count; | |
857 | SV * sva; | |
858 | SV * svb; | |
a0d0e21e | 859 | |
4358a253 | 860 | ENTER; |
a0d0e21e LW |
861 | SAVETMPS; |
862 | ||
4358a253 SS |
863 | sva = sv_2mortal(newSViv(a)); |
864 | svb = sv_2mortal(newSViv(b)); | |
a0d0e21e | 865 | |
4358a253 | 866 | PUSHMARK(SP); |
d0554719 TC |
867 | EXTEND(SP, 2); |
868 | PUSHs(sva); | |
869 | PUSHs(svb); | |
4358a253 | 870 | PUTBACK; |
a0d0e21e | 871 | |
4929bf7b | 872 | count = call_pv("Inc", G_DISCARD); |
a0d0e21e LW |
873 | |
874 | if (count != 0) | |
d1b91892 | 875 | croak ("call_Inc: expected 0 values from 'Inc', got %d\n", |
4358a253 | 876 | count); |
a0d0e21e | 877 | |
4358a253 SS |
878 | printf ("%d + 1 = %d\n", a, SvIV(sva)); |
879 | printf ("%d + 1 = %d\n", b, SvIV(svb)); | |
a0d0e21e | 880 | |
4358a253 SS |
881 | FREETMPS; |
882 | LEAVE; | |
a0d0e21e LW |
883 | } |
884 | ||
d1b91892 | 885 | To be able to access the two parameters that were pushed onto the stack |
4929bf7b | 886 | after they return from I<call_pv> it is necessary to make a note |
19799a22 | 887 | of their addresses--thus the two variables C<sva> and C<svb>. |
a0d0e21e | 888 | |
d1b91892 AD |
889 | The reason this is necessary is that the area of the Perl stack which |
890 | held them will very likely have been overwritten by something else by | |
4929bf7b | 891 | the time control returns from I<call_pv>. |
a0d0e21e LW |
892 | |
893 | ||
894 | ||
895 | ||
d1b91892 | 896 | =head2 Using G_EVAL |
a0d0e21e | 897 | |
d1b91892 AD |
898 | Now an example using G_EVAL. Below is a Perl subroutine which computes |
899 | the difference of its 2 parameters. If this would result in a negative | |
900 | result, the subroutine calls I<die>. | |
a0d0e21e | 901 | |
84f709e7 JH |
902 | sub Subtract |
903 | { | |
4358a253 | 904 | my ($a, $b) = @_; |
a0d0e21e | 905 | |
4358a253 | 906 | die "death can be fatal\n" if $a < $b; |
a0d0e21e | 907 | |
4358a253 | 908 | $a - $b; |
a0d0e21e LW |
909 | } |
910 | ||
911 | and some C to call it | |
912 | ||
e46aa1dd KW |
913 | static void |
914 | call_Subtract(a, b) | |
915 | int a; | |
916 | int b; | |
917 | { | |
918 | dSP; | |
919 | int count; | |
920 | SV *err_tmp; | |
921 | ||
922 | ENTER; | |
923 | SAVETMPS; | |
924 | ||
925 | PUSHMARK(SP); | |
926 | EXTEND(SP, 2); | |
927 | PUSHs(sv_2mortal(newSViv(a))); | |
928 | PUSHs(sv_2mortal(newSViv(b))); | |
929 | PUTBACK; | |
930 | ||
931 | count = call_pv("Subtract", G_EVAL|G_SCALAR); | |
932 | ||
933 | SPAGAIN; | |
934 | ||
935 | /* Check the eval first */ | |
936 | err_tmp = ERRSV; | |
937 | if (SvTRUE(err_tmp)) | |
938 | { | |
939 | printf ("Uh oh - %s\n", SvPV_nolen(err_tmp)); | |
940 | POPs; | |
941 | } | |
942 | else | |
943 | { | |
944 | if (count != 1) | |
945 | croak("call_Subtract: wanted 1 value from 'Subtract', got %d\n", | |
946 | count); | |
947 | ||
948 | printf ("%d - %d = %d\n", a, b, POPi); | |
949 | } | |
950 | ||
951 | PUTBACK; | |
952 | FREETMPS; | |
953 | LEAVE; | |
954 | } | |
a0d0e21e LW |
955 | |
956 | If I<call_Subtract> is called thus | |
957 | ||
d1b91892 | 958 | call_Subtract(4, 5) |
a0d0e21e LW |
959 | |
960 | the following will be printed | |
961 | ||
d1b91892 | 962 | Uh oh - death can be fatal |
a0d0e21e LW |
963 | |
964 | Notes | |
965 | ||
966 | =over 5 | |
967 | ||
968 | =item 1. | |
969 | ||
d1b91892 AD |
970 | We want to be able to catch the I<die> so we have used the G_EVAL |
971 | flag. Not specifying this flag would mean that the program would | |
972 | terminate immediately at the I<die> statement in the subroutine | |
973 | I<Subtract>. | |
a0d0e21e LW |
974 | |
975 | =item 2. | |
976 | ||
54310121 | 977 | The code |
a0d0e21e | 978 | |
d0554719 TC |
979 | err_tmp = ERRSV; |
980 | if (SvTRUE(err_tmp)) | |
d1b91892 | 981 | { |
d0554719 | 982 | printf ("Uh oh - %s\n", SvPV_nolen(err_tmp)); |
4358a253 | 983 | POPs; |
d1b91892 | 984 | } |
a0d0e21e | 985 | |
d1b91892 | 986 | is the direct equivalent of this bit of Perl |
a0d0e21e | 987 | |
4358a253 | 988 | print "Uh oh - $@\n" if $@; |
a0d0e21e | 989 | |
d0554719 TC |
990 | C<PL_errgv> is a perl global of type C<GV *> that points to the symbol |
991 | table entry containing the error. C<ERRSV> therefore refers to the C | |
3da3c74d | 992 | equivalent of C<$@>. We use a local temporary, C<err_tmp>, since |
d0554719 TC |
993 | C<ERRSV> is a macro that calls a function, and C<SvTRUE(ERRSV)> would |
994 | end up calling that function multiple times. | |
c07a80fd | 995 | |
d1b91892 | 996 | =item 3. |
a0d0e21e | 997 | |
d1b91892 | 998 | Note that the stack is popped using C<POPs> in the block where |
d0554719 | 999 | C<SvTRUE(err_tmp)> is true. This is necessary because whenever a |
4929bf7b | 1000 | I<call_*> function invoked with G_EVAL|G_SCALAR returns an error, |
5f05dabc | 1001 | the top of the stack holds the value I<undef>. Because we want the |
d1b91892 | 1002 | program to continue after detecting this error, it is essential that |
6c818a50 | 1003 | the stack be tidied up by removing the I<undef>. |
a0d0e21e LW |
1004 | |
1005 | =back | |
1006 | ||
1007 | ||
c07a80fd | 1008 | =head2 Using G_KEEPERR |
1009 | ||
1010 | Consider this rather facetious example, where we have used an XS | |
1011 | version of the call_Subtract example above inside a destructor: | |
1012 | ||
1013 | package Foo; | |
84f709e7 | 1014 | sub new { bless {}, $_[0] } |
54310121 | 1015 | sub Subtract { |
84f709e7 | 1016 | my($a,$b) = @_; |
4358a253 | 1017 | die "death can be fatal" if $a < $b; |
84f709e7 | 1018 | $a - $b; |
c07a80fd | 1019 | } |
84f709e7 JH |
1020 | sub DESTROY { call_Subtract(5, 4); } |
1021 | sub foo { die "foo dies"; } | |
c07a80fd | 1022 | |
1023 | package main; | |
7ce09284 Z |
1024 | { |
1025 | my $foo = Foo->new; | |
1026 | eval { $foo->foo }; | |
1027 | } | |
c07a80fd | 1028 | print "Saw: $@" if $@; # should be, but isn't |
1029 | ||
1030 | This example will fail to recognize that an error occurred inside the | |
1031 | C<eval {}>. Here's why: the call_Subtract code got executed while perl | |
7ce09284 | 1032 | was cleaning up temporaries when exiting the outer braced block, and because |
4929bf7b | 1033 | call_Subtract is implemented with I<call_pv> using the G_EVAL |
c07a80fd | 1034 | flag, it promptly reset C<$@>. This results in the failure of the |
1035 | outermost test for C<$@>, and thereby the failure of the error trap. | |
1036 | ||
4929bf7b | 1037 | Appending the G_KEEPERR flag, so that the I<call_pv> call in |
c07a80fd | 1038 | call_Subtract reads: |
1039 | ||
4929bf7b | 1040 | count = call_pv("Subtract", G_EVAL|G_SCALAR|G_KEEPERR); |
c07a80fd | 1041 | |
1042 | will preserve the error and restore reliable error handling. | |
1043 | ||
4929bf7b | 1044 | =head2 Using call_sv |
a0d0e21e | 1045 | |
d1b91892 AD |
1046 | In all the previous examples I have 'hard-wired' the name of the Perl |
1047 | subroutine to be called from C. Most of the time though, it is more | |
1048 | convenient to be able to specify the name of the Perl subroutine from | |
d0554719 TC |
1049 | within the Perl script, and you'll want to use |
1050 | L<call_sv|perlapi/call_sv>. | |
a0d0e21e LW |
1051 | |
1052 | Consider the Perl code below | |
1053 | ||
84f709e7 JH |
1054 | sub fred |
1055 | { | |
4358a253 | 1056 | print "Hello there\n"; |
d1b91892 AD |
1057 | } |
1058 | ||
4358a253 | 1059 | CallSubPV("fred"); |
d1b91892 AD |
1060 | |
1061 | Here is a snippet of XSUB which defines I<CallSubPV>. | |
1062 | ||
1063 | void | |
1064 | CallSubPV(name) | |
1065 | char * name | |
1066 | CODE: | |
4358a253 SS |
1067 | PUSHMARK(SP); |
1068 | call_pv(name, G_DISCARD|G_NOARGS); | |
a0d0e21e | 1069 | |
54310121 | 1070 | That is fine as far as it goes. The thing is, the Perl subroutine |
94a37a2f BF |
1071 | can be specified as only a string, however, Perl allows references |
1072 | to subroutines and anonymous subroutines. | |
4929bf7b | 1073 | This is where I<call_sv> is useful. |
d1b91892 AD |
1074 | |
1075 | The code below for I<CallSubSV> is identical to I<CallSubPV> except | |
1076 | that the C<name> parameter is now defined as an SV* and we use | |
4929bf7b | 1077 | I<call_sv> instead of I<call_pv>. |
d1b91892 AD |
1078 | |
1079 | void | |
1080 | CallSubSV(name) | |
1081 | SV * name | |
1082 | CODE: | |
4358a253 SS |
1083 | PUSHMARK(SP); |
1084 | call_sv(name, G_DISCARD|G_NOARGS); | |
a0d0e21e | 1085 | |
1d45ec27 | 1086 | Because we are using an SV to call I<fred> the following can all be used: |
a0d0e21e | 1087 | |
4358a253 SS |
1088 | CallSubSV("fred"); |
1089 | CallSubSV(\&fred); | |
1090 | $ref = \&fred; | |
1091 | CallSubSV($ref); | |
1092 | CallSubSV( sub { print "Hello there\n" } ); | |
a0d0e21e | 1093 | |
4929bf7b | 1094 | As you can see, I<call_sv> gives you much greater flexibility in |
d1b91892 AD |
1095 | how you can specify the Perl subroutine. |
1096 | ||
1d45ec27 | 1097 | You should note that, if it is necessary to store the SV (C<name> in the |
d1b91892 | 1098 | example above) which corresponds to the Perl subroutine so that it can |
5f05dabc | 1099 | be used later in the program, it not enough just to store a copy of the |
1d45ec27 | 1100 | pointer to the SV. Say the code above had been like this: |
d1b91892 | 1101 | |
4358a253 | 1102 | static SV * rememberSub; |
d1b91892 AD |
1103 | |
1104 | void | |
1105 | SaveSub1(name) | |
1106 | SV * name | |
1107 | CODE: | |
4358a253 | 1108 | rememberSub = name; |
d1b91892 AD |
1109 | |
1110 | void | |
1111 | CallSavedSub1() | |
1112 | CODE: | |
4358a253 SS |
1113 | PUSHMARK(SP); |
1114 | call_sv(rememberSub, G_DISCARD|G_NOARGS); | |
a0d0e21e | 1115 | |
1d45ec27 | 1116 | The reason this is wrong is that, by the time you come to use the |
d1b91892 AD |
1117 | pointer C<rememberSub> in C<CallSavedSub1>, it may or may not still refer |
1118 | to the Perl subroutine that was recorded in C<SaveSub1>. This is | |
1d45ec27 | 1119 | particularly true for these cases: |
a0d0e21e | 1120 | |
4358a253 SS |
1121 | SaveSub1(\&fred); |
1122 | CallSavedSub1(); | |
a0d0e21e | 1123 | |
4358a253 SS |
1124 | SaveSub1( sub { print "Hello there\n" } ); |
1125 | CallSavedSub1(); | |
a0d0e21e | 1126 | |
1d45ec27 | 1127 | By the time each of the C<SaveSub1> statements above has been executed, |
54310121 | 1128 | the SV*s which corresponded to the parameters will no longer exist. |
d1b91892 | 1129 | Expect an error message from Perl of the form |
a0d0e21e | 1130 | |
d1b91892 | 1131 | Can't use an undefined value as a subroutine reference at ... |
a0d0e21e | 1132 | |
d1b91892 | 1133 | for each of the C<CallSavedSub1> lines. |
a0d0e21e | 1134 | |
54310121 | 1135 | Similarly, with this code |
a0d0e21e | 1136 | |
4358a253 SS |
1137 | $ref = \&fred; |
1138 | SaveSub1($ref); | |
1139 | $ref = 47; | |
1140 | CallSavedSub1(); | |
a0d0e21e | 1141 | |
54310121 | 1142 | you can expect one of these messages (which you actually get is dependent on |
1143 | the version of Perl you are using) | |
a0d0e21e | 1144 | |
d1b91892 AD |
1145 | Not a CODE reference at ... |
1146 | Undefined subroutine &main::47 called ... | |
a0d0e21e | 1147 | |
19799a22 | 1148 | The variable $ref may have referred to the subroutine C<fred> |
d1b91892 | 1149 | whenever the call to C<SaveSub1> was made but by the time |
5f05dabc | 1150 | C<CallSavedSub1> gets called it now holds the number C<47>. Because we |
d1b91892 | 1151 | saved only a pointer to the original SV in C<SaveSub1>, any changes to |
19799a22 | 1152 | $ref will be tracked by the pointer C<rememberSub>. This means that |
d1b91892 AD |
1153 | whenever C<CallSavedSub1> gets called, it will attempt to execute the |
1154 | code which is referenced by the SV* C<rememberSub>. In this case | |
1155 | though, it now refers to the integer C<47>, so expect Perl to complain | |
1156 | loudly. | |
a0d0e21e | 1157 | |
1d45ec27 | 1158 | A similar but more subtle problem is illustrated with this code: |
a0d0e21e | 1159 | |
4358a253 SS |
1160 | $ref = \&fred; |
1161 | SaveSub1($ref); | |
1162 | $ref = \&joe; | |
1163 | CallSavedSub1(); | |
a0d0e21e | 1164 | |
1d45ec27 | 1165 | This time whenever C<CallSavedSub1> gets called it will execute the Perl |
54310121 | 1166 | subroutine C<joe> (assuming it exists) rather than C<fred> as was |
d1b91892 | 1167 | originally requested in the call to C<SaveSub1>. |
a0d0e21e | 1168 | |
d1b91892 | 1169 | To get around these problems it is necessary to take a full copy of the |
1d45ec27 | 1170 | SV. The code below shows C<SaveSub2> modified to do that. |
a0d0e21e | 1171 | |
d0554719 | 1172 | /* this isn't thread-safe */ |
4358a253 | 1173 | static SV * keepSub = (SV*)NULL; |
d1b91892 AD |
1174 | |
1175 | void | |
1176 | SaveSub2(name) | |
1177 | SV * name | |
1178 | CODE: | |
1179 | /* Take a copy of the callback */ | |
1180 | if (keepSub == (SV*)NULL) | |
1181 | /* First time, so create a new SV */ | |
4358a253 | 1182 | keepSub = newSVsv(name); |
d1b91892 AD |
1183 | else |
1184 | /* Been here before, so overwrite */ | |
4358a253 | 1185 | SvSetSV(keepSub, name); |
d1b91892 AD |
1186 | |
1187 | void | |
1188 | CallSavedSub2() | |
1189 | CODE: | |
4358a253 SS |
1190 | PUSHMARK(SP); |
1191 | call_sv(keepSub, G_DISCARD|G_NOARGS); | |
d1b91892 | 1192 | |
5f05dabc | 1193 | To avoid creating a new SV every time C<SaveSub2> is called, |
d1b91892 AD |
1194 | the function first checks to see if it has been called before. If not, |
1195 | then space for a new SV is allocated and the reference to the Perl | |
1d45ec27 FC |
1196 | subroutine C<name> is copied to the variable C<keepSub> in one |
1197 | operation using C<newSVsv>. Thereafter, whenever C<SaveSub2> is called, | |
d1b91892 AD |
1198 | the existing SV, C<keepSub>, is overwritten with the new value using |
1199 | C<SvSetSV>. | |
1200 | ||
d0554719 TC |
1201 | Note: using a static or global variable to store the SV isn't |
1202 | thread-safe. You can either use the C<MY_CXT> mechanism documented in | |
1203 | L<perlxs/Safely Storing Static Data in XS> which is fast, or store the | |
1204 | values in perl global variables, using get_sv(), which is much slower. | |
1205 | ||
4929bf7b | 1206 | =head2 Using call_argv |
d1b91892 AD |
1207 | |
1208 | Here is a Perl subroutine which prints whatever parameters are passed | |
1209 | to it. | |
1210 | ||
84f709e7 JH |
1211 | sub PrintList |
1212 | { | |
4358a253 | 1213 | my(@list) = @_; |
d1b91892 | 1214 | |
84f709e7 | 1215 | foreach (@list) { print "$_\n" } |
d1b91892 AD |
1216 | } |
1217 | ||
1d45ec27 | 1218 | And here is an example of I<call_argv> which will call |
d1b91892 AD |
1219 | I<PrintList>. |
1220 | ||
4358a253 | 1221 | static char * words[] = {"alpha", "beta", "gamma", "delta", NULL}; |
d1b91892 AD |
1222 | |
1223 | static void | |
1224 | call_PrintList() | |
1225 | { | |
4358a253 | 1226 | call_argv("PrintList", G_DISCARD, words); |
d1b91892 AD |
1227 | } |
1228 | ||
1229 | Note that it is not necessary to call C<PUSHMARK> in this instance. | |
4929bf7b | 1230 | This is because I<call_argv> will do it for you. |
d1b91892 | 1231 | |
4929bf7b | 1232 | =head2 Using call_method |
a0d0e21e | 1233 | |
1d45ec27 | 1234 | Consider the following Perl code: |
a0d0e21e | 1235 | |
d1b91892 | 1236 | { |
4358a253 | 1237 | package Mine; |
84f709e7 JH |
1238 | |
1239 | sub new | |
1240 | { | |
4358a253 | 1241 | my($type) = shift; |
84f709e7 JH |
1242 | bless [@_] |
1243 | } | |
1244 | ||
1245 | sub Display | |
1246 | { | |
4358a253 SS |
1247 | my ($self, $index) = @_; |
1248 | print "$index: $$self[$index]\n"; | |
84f709e7 JH |
1249 | } |
1250 | ||
1251 | sub PrintID | |
1252 | { | |
4358a253 SS |
1253 | my($class) = @_; |
1254 | print "This is Class $class version 1.0\n"; | |
84f709e7 | 1255 | } |
d1b91892 AD |
1256 | } |
1257 | ||
5f05dabc | 1258 | It implements just a very simple class to manage an array. Apart from |
d1b91892 | 1259 | the constructor, C<new>, it declares methods, one static and one |
5f05dabc | 1260 | virtual. The static method, C<PrintID>, prints out simply the class |
d1b91892 | 1261 | name and a version number. The virtual method, C<Display>, prints out a |
1d45ec27 | 1262 | single element of the array. Here is an all-Perl example of using it. |
d1b91892 | 1263 | |
797f796a | 1264 | $a = Mine->new('red', 'green', 'blue'); |
4358a253 | 1265 | $a->Display(1); |
797f796a | 1266 | Mine->PrintID; |
a0d0e21e | 1267 | |
d1b91892 | 1268 | will print |
a0d0e21e | 1269 | |
d1b91892 | 1270 | 1: green |
54310121 | 1271 | This is Class Mine version 1.0 |
a0d0e21e | 1272 | |
d1b91892 | 1273 | Calling a Perl method from C is fairly straightforward. The following |
1d45ec27 | 1274 | things are required: |
a0d0e21e | 1275 | |
d1b91892 AD |
1276 | =over 5 |
1277 | ||
1278 | =item * | |
1279 | ||
1d45ec27 FC |
1280 | A reference to the object for a virtual method or the name of the class |
1281 | for a static method | |
d1b91892 AD |
1282 | |
1283 | =item * | |
1284 | ||
1d45ec27 | 1285 | The name of the method |
d1b91892 AD |
1286 | |
1287 | =item * | |
1288 | ||
1d45ec27 | 1289 | Any other parameters specific to the method |
d1b91892 AD |
1290 | |
1291 | =back | |
1292 | ||
1293 | Here is a simple XSUB which illustrates the mechanics of calling both | |
1294 | the C<PrintID> and C<Display> methods from C. | |
1295 | ||
1296 | void | |
1297 | call_Method(ref, method, index) | |
1298 | SV * ref | |
1299 | char * method | |
1300 | int index | |
1301 | CODE: | |
924508f0 | 1302 | PUSHMARK(SP); |
d0554719 TC |
1303 | EXTEND(SP, 2); |
1304 | PUSHs(ref); | |
1305 | PUSHs(sv_2mortal(newSViv(index))); | |
d1b91892 AD |
1306 | PUTBACK; |
1307 | ||
4358a253 | 1308 | call_method(method, G_DISCARD); |
d1b91892 AD |
1309 | |
1310 | void | |
1311 | call_PrintID(class, method) | |
1312 | char * class | |
1313 | char * method | |
1314 | CODE: | |
924508f0 | 1315 | PUSHMARK(SP); |
4358a253 | 1316 | XPUSHs(sv_2mortal(newSVpv(class, 0))); |
d1b91892 AD |
1317 | PUTBACK; |
1318 | ||
4358a253 | 1319 | call_method(method, G_DISCARD); |
d1b91892 AD |
1320 | |
1321 | ||
1d45ec27 | 1322 | So the methods C<PrintID> and C<Display> can be invoked like this: |
d1b91892 | 1323 | |
797f796a | 1324 | $a = Mine->new('red', 'green', 'blue'); |
4358a253 SS |
1325 | call_Method($a, 'Display', 1); |
1326 | call_PrintID('Mine', 'PrintID'); | |
d1b91892 | 1327 | |
1d45ec27 | 1328 | The only thing to note is that, in both the static and virtual methods, |
19799a22 | 1329 | the method name is not passed via the stack--it is used as the first |
4929bf7b | 1330 | parameter to I<call_method>. |
d1b91892 | 1331 | |
54310121 | 1332 | =head2 Using GIMME_V |
d1b91892 | 1333 | |
54310121 | 1334 | Here is a trivial XSUB which prints the context in which it is |
d1b91892 AD |
1335 | currently executing. |
1336 | ||
1337 | void | |
1338 | PrintContext() | |
1339 | CODE: | |
1c23e2bd | 1340 | U8 gimme = GIMME_V; |
54310121 | 1341 | if (gimme == G_VOID) |
4358a253 | 1342 | printf ("Context is Void\n"); |
54310121 | 1343 | else if (gimme == G_SCALAR) |
4358a253 | 1344 | printf ("Context is Scalar\n"); |
d1b91892 | 1345 | else |
4358a253 | 1346 | printf ("Context is Array\n"); |
d1b91892 | 1347 | |
1d45ec27 | 1348 | And here is some Perl to test it. |
d1b91892 | 1349 | |
4358a253 SS |
1350 | PrintContext; |
1351 | $a = PrintContext; | |
1352 | @a = PrintContext; | |
d1b91892 AD |
1353 | |
1354 | The output from that will be | |
1355 | ||
54310121 | 1356 | Context is Void |
d1b91892 AD |
1357 | Context is Scalar |
1358 | Context is Array | |
1359 | ||
02f6dca1 | 1360 | =head2 Using Perl to Dispose of Temporaries |
d1b91892 AD |
1361 | |
1362 | In the examples given to date, any temporaries created in the callback | |
4929bf7b | 1363 | (i.e., parameters passed on the stack to the I<call_*> function or |
1d45ec27 | 1364 | values returned via the stack) have been freed by one of these methods: |
d1b91892 AD |
1365 | |
1366 | =over 5 | |
1367 | ||
1368 | =item * | |
1369 | ||
1d45ec27 | 1370 | Specifying the G_DISCARD flag with I<call_*> |
d1b91892 AD |
1371 | |
1372 | =item * | |
1373 | ||
1d45ec27 | 1374 | Explicitly using the C<ENTER>/C<SAVETMPS>--C<FREETMPS>/C<LEAVE> pairing |
d1b91892 AD |
1375 | |
1376 | =back | |
1377 | ||
1378 | There is another method which can be used, namely letting Perl do it | |
1379 | for you automatically whenever it regains control after the callback | |
1380 | has terminated. This is done by simply not using the | |
1381 | ||
4358a253 SS |
1382 | ENTER; |
1383 | SAVETMPS; | |
d1b91892 | 1384 | ... |
4358a253 SS |
1385 | FREETMPS; |
1386 | LEAVE; | |
d1b91892 AD |
1387 | |
1388 | sequence in the callback (and not, of course, specifying the G_DISCARD | |
1389 | flag). | |
1390 | ||
1391 | If you are going to use this method you have to be aware of a possible | |
1392 | memory leak which can arise under very specific circumstances. To | |
1393 | explain these circumstances you need to know a bit about the flow of | |
1394 | control between Perl and the callback routine. | |
1395 | ||
1396 | The examples given at the start of the document (an error handler and | |
1397 | an event driven program) are typical of the two main sorts of flow | |
1398 | control that you are likely to encounter with callbacks. There is a | |
1399 | very important distinction between them, so pay attention. | |
1400 | ||
1401 | In the first example, an error handler, the flow of control could be as | |
1402 | follows. You have created an interface to an external library. | |
1403 | Control can reach the external library like this | |
1404 | ||
1405 | perl --> XSUB --> external library | |
1406 | ||
1407 | Whilst control is in the library, an error condition occurs. You have | |
1408 | previously set up a Perl callback to handle this situation, so it will | |
1409 | get executed. Once the callback has finished, control will drop back to | |
1410 | Perl again. Here is what the flow of control will be like in that | |
1411 | situation | |
1412 | ||
1413 | perl --> XSUB --> external library | |
1414 | ... | |
1415 | error occurs | |
1416 | ... | |
4929bf7b | 1417 | external library --> call_* --> perl |
d1b91892 | 1418 | | |
4929bf7b | 1419 | perl <-- XSUB <-- external library <-- call_* <----+ |
d1b91892 | 1420 | |
4929bf7b | 1421 | After processing of the error using I<call_*> is completed, |
d1b91892 AD |
1422 | control reverts back to Perl more or less immediately. |
1423 | ||
1424 | In the diagram, the further right you go the more deeply nested the | |
1425 | scope is. It is only when control is back with perl on the extreme | |
1426 | left of the diagram that you will have dropped back to the enclosing | |
1427 | scope and any temporaries you have left hanging around will be freed. | |
1428 | ||
1429 | In the second example, an event driven program, the flow of control | |
1430 | will be more like this | |
1431 | ||
1432 | perl --> XSUB --> event handler | |
1433 | ... | |
4929bf7b | 1434 | event handler --> call_* --> perl |
d1b91892 | 1435 | | |
4929bf7b | 1436 | event handler <-- call_* <----+ |
d1b91892 | 1437 | ... |
4929bf7b | 1438 | event handler --> call_* --> perl |
d1b91892 | 1439 | | |
4929bf7b | 1440 | event handler <-- call_* <----+ |
d1b91892 | 1441 | ... |
4929bf7b | 1442 | event handler --> call_* --> perl |
d1b91892 | 1443 | | |
4929bf7b | 1444 | event handler <-- call_* <----+ |
d1b91892 AD |
1445 | |
1446 | In this case the flow of control can consist of only the repeated | |
1447 | sequence | |
1448 | ||
4929bf7b | 1449 | event handler --> call_* --> perl |
d1b91892 | 1450 | |
54310121 | 1451 | for practically the complete duration of the program. This means that |
1452 | control may I<never> drop back to the surrounding scope in Perl at the | |
1453 | extreme left. | |
d1b91892 AD |
1454 | |
1455 | So what is the big problem? Well, if you are expecting Perl to tidy up | |
1456 | those temporaries for you, you might be in for a long wait. For Perl | |
5f05dabc | 1457 | to dispose of your temporaries, control must drop back to the |
d1b91892 | 1458 | enclosing scope at some stage. In the event driven scenario that may |
1d45ec27 | 1459 | never happen. This means that, as time goes on, your program will |
d1b91892 AD |
1460 | create more and more temporaries, none of which will ever be freed. As |
1461 | each of these temporaries consumes some memory your program will | |
19799a22 | 1462 | eventually consume all the available memory in your system--kapow! |
d1b91892 | 1463 | |
19799a22 | 1464 | So here is the bottom line--if you are sure that control will revert |
d1b91892 | 1465 | back to the enclosing Perl scope fairly quickly after the end of your |
5f05dabc | 1466 | callback, then it isn't absolutely necessary to dispose explicitly of |
d1b91892 AD |
1467 | any temporaries you may have created. Mind you, if you are at all |
1468 | uncertain about what to do, it doesn't do any harm to tidy up anyway. | |
1469 | ||
1470 | ||
02f6dca1 | 1471 | =head2 Strategies for Storing Callback Context Information |
d1b91892 AD |
1472 | |
1473 | ||
1474 | Potentially one of the trickiest problems to overcome when designing a | |
1475 | callback interface can be figuring out how to store the mapping between | |
1476 | the C callback function and the Perl equivalent. | |
1477 | ||
1478 | To help understand why this can be a real problem first consider how a | |
1479 | callback is set up in an all C environment. Typically a C API will | |
1480 | provide a function to register a callback. This will expect a pointer | |
1481 | to a function as one of its parameters. Below is a call to a | |
1482 | hypothetical function C<register_fatal> which registers the C function | |
1483 | to get called when a fatal error occurs. | |
1484 | ||
4358a253 | 1485 | register_fatal(cb1); |
d1b91892 AD |
1486 | |
1487 | The single parameter C<cb1> is a pointer to a function, so you must | |
1488 | have defined C<cb1> in your code, say something like this | |
1489 | ||
1490 | static void | |
1491 | cb1() | |
1492 | { | |
4358a253 SS |
1493 | printf ("Fatal Error\n"); |
1494 | exit(1); | |
d1b91892 AD |
1495 | } |
1496 | ||
1497 | Now change that to call a Perl subroutine instead | |
1498 | ||
1499 | static SV * callback = (SV*)NULL; | |
1500 | ||
1501 | static void | |
1502 | cb1() | |
1503 | { | |
4358a253 | 1504 | dSP; |
d1b91892 | 1505 | |
4358a253 | 1506 | PUSHMARK(SP); |
d1b91892 AD |
1507 | |
1508 | /* Call the Perl sub to process the callback */ | |
4358a253 | 1509 | call_sv(callback, G_DISCARD); |
d1b91892 AD |
1510 | } |
1511 | ||
1512 | ||
1513 | void | |
1514 | register_fatal(fn) | |
1515 | SV * fn | |
1516 | CODE: | |
1517 | /* Remember the Perl sub */ | |
1518 | if (callback == (SV*)NULL) | |
4358a253 | 1519 | callback = newSVsv(fn); |
d1b91892 | 1520 | else |
4358a253 | 1521 | SvSetSV(callback, fn); |
d1b91892 AD |
1522 | |
1523 | /* register the callback with the external library */ | |
4358a253 | 1524 | register_fatal(cb1); |
d1b91892 AD |
1525 | |
1526 | where the Perl equivalent of C<register_fatal> and the callback it | |
1527 | registers, C<pcb1>, might look like this | |
1528 | ||
1529 | # Register the sub pcb1 | |
4358a253 | 1530 | register_fatal(\&pcb1); |
d1b91892 | 1531 | |
84f709e7 JH |
1532 | sub pcb1 |
1533 | { | |
4358a253 | 1534 | die "I'm dying...\n"; |
d1b91892 AD |
1535 | } |
1536 | ||
1537 | The mapping between the C callback and the Perl equivalent is stored in | |
1538 | the global variable C<callback>. | |
1539 | ||
5f05dabc | 1540 | This will be adequate if you ever need to have only one callback |
d1b91892 AD |
1541 | registered at any time. An example could be an error handler like the |
1542 | code sketched out above. Remember though, repeated calls to | |
1543 | C<register_fatal> will replace the previously registered callback | |
1544 | function with the new one. | |
1545 | ||
1546 | Say for example you want to interface to a library which allows asynchronous | |
1547 | file i/o. In this case you may be able to register a callback whenever | |
1548 | a read operation has completed. To be of any use we want to be able to | |
1549 | call separate Perl subroutines for each file that is opened. As it | |
1550 | stands, the error handler example above would not be adequate as it | |
1551 | allows only a single callback to be defined at any time. What we | |
1552 | require is a means of storing the mapping between the opened file and | |
1553 | the Perl subroutine we want to be called for that file. | |
1554 | ||
1555 | Say the i/o library has a function C<asynch_read> which associates a C | |
19799a22 | 1556 | function C<ProcessRead> with a file handle C<fh>--this assumes that it |
d1b91892 AD |
1557 | has also provided some routine to open the file and so obtain the file |
1558 | handle. | |
1559 | ||
1560 | asynch_read(fh, ProcessRead) | |
1561 | ||
1562 | This may expect the C I<ProcessRead> function of this form | |
1563 | ||
1564 | void | |
1565 | ProcessRead(fh, buffer) | |
4358a253 SS |
1566 | int fh; |
1567 | char * buffer; | |
d1b91892 | 1568 | { |
54310121 | 1569 | ... |
d1b91892 AD |
1570 | } |
1571 | ||
1572 | To provide a Perl interface to this library we need to be able to map | |
1573 | between the C<fh> parameter and the Perl subroutine we want called. A | |
1574 | hash is a convenient mechanism for storing this mapping. The code | |
1575 | below shows a possible implementation | |
1576 | ||
4358a253 | 1577 | static HV * Mapping = (HV*)NULL; |
a0d0e21e | 1578 | |
d1b91892 AD |
1579 | void |
1580 | asynch_read(fh, callback) | |
1581 | int fh | |
1582 | SV * callback | |
1583 | CODE: | |
1584 | /* If the hash doesn't already exist, create it */ | |
1585 | if (Mapping == (HV*)NULL) | |
4358a253 | 1586 | Mapping = newHV(); |
d1b91892 AD |
1587 | |
1588 | /* Save the fh -> callback mapping */ | |
4358a253 | 1589 | hv_store(Mapping, (char*)&fh, sizeof(fh), newSVsv(callback), 0); |
d1b91892 AD |
1590 | |
1591 | /* Register with the C Library */ | |
4358a253 | 1592 | asynch_read(fh, asynch_read_if); |
d1b91892 AD |
1593 | |
1594 | and C<asynch_read_if> could look like this | |
1595 | ||
1596 | static void | |
1597 | asynch_read_if(fh, buffer) | |
4358a253 SS |
1598 | int fh; |
1599 | char * buffer; | |
d1b91892 | 1600 | { |
4358a253 SS |
1601 | dSP; |
1602 | SV ** sv; | |
d1b91892 AD |
1603 | |
1604 | /* Get the callback associated with fh */ | |
4358a253 | 1605 | sv = hv_fetch(Mapping, (char*)&fh , sizeof(fh), FALSE); |
d1b91892 | 1606 | if (sv == (SV**)NULL) |
4358a253 | 1607 | croak("Internal error...\n"); |
d1b91892 | 1608 | |
4358a253 | 1609 | PUSHMARK(SP); |
d0554719 TC |
1610 | EXTEND(SP, 2); |
1611 | PUSHs(sv_2mortal(newSViv(fh))); | |
1612 | PUSHs(sv_2mortal(newSVpv(buffer, 0))); | |
4358a253 | 1613 | PUTBACK; |
d1b91892 AD |
1614 | |
1615 | /* Call the Perl sub */ | |
4358a253 | 1616 | call_sv(*sv, G_DISCARD); |
d1b91892 AD |
1617 | } |
1618 | ||
1619 | For completeness, here is C<asynch_close>. This shows how to remove | |
1620 | the entry from the hash C<Mapping>. | |
1621 | ||
1622 | void | |
1623 | asynch_close(fh) | |
1624 | int fh | |
1625 | CODE: | |
1626 | /* Remove the entry from the hash */ | |
4358a253 | 1627 | (void) hv_delete(Mapping, (char*)&fh, sizeof(fh), G_DISCARD); |
a0d0e21e | 1628 | |
d1b91892 | 1629 | /* Now call the real asynch_close */ |
4358a253 | 1630 | asynch_close(fh); |
a0d0e21e | 1631 | |
d1b91892 AD |
1632 | So the Perl interface would look like this |
1633 | ||
84f709e7 JH |
1634 | sub callback1 |
1635 | { | |
4358a253 | 1636 | my($handle, $buffer) = @_; |
d1b91892 | 1637 | } |
a0d0e21e | 1638 | |
d1b91892 | 1639 | # Register the Perl callback |
4358a253 | 1640 | asynch_read($fh, \&callback1); |
a0d0e21e | 1641 | |
4358a253 | 1642 | asynch_close($fh); |
d1b91892 AD |
1643 | |
1644 | The mapping between the C callback and Perl is stored in the global | |
1645 | hash C<Mapping> this time. Using a hash has the distinct advantage that | |
1646 | it allows an unlimited number of callbacks to be registered. | |
1647 | ||
1648 | What if the interface provided by the C callback doesn't contain a | |
1649 | parameter which allows the file handle to Perl subroutine mapping? Say | |
1650 | in the asynchronous i/o package, the callback function gets passed only | |
1651 | the C<buffer> parameter like this | |
1652 | ||
1653 | void | |
1654 | ProcessRead(buffer) | |
4358a253 | 1655 | char * buffer; |
d1b91892 AD |
1656 | { |
1657 | ... | |
1658 | } | |
a0d0e21e | 1659 | |
d1b91892 AD |
1660 | Without the file handle there is no straightforward way to map from the |
1661 | C callback to the Perl subroutine. | |
a0d0e21e | 1662 | |
54310121 | 1663 | In this case a possible way around this problem is to predefine a |
d1b91892 AD |
1664 | series of C functions to act as the interface to Perl, thus |
1665 | ||
1666 | #define MAX_CB 3 | |
1667 | #define NULL_HANDLE -1 | |
4358a253 | 1668 | typedef void (*FnMap)(); |
d1b91892 AD |
1669 | |
1670 | struct MapStruct { | |
4358a253 SS |
1671 | FnMap Function; |
1672 | SV * PerlSub; | |
1673 | int Handle; | |
1674 | }; | |
d1b91892 | 1675 | |
4358a253 SS |
1676 | static void fn1(); |
1677 | static void fn2(); | |
1678 | static void fn3(); | |
d1b91892 AD |
1679 | |
1680 | static struct MapStruct Map [MAX_CB] = | |
1681 | { | |
1682 | { fn1, NULL, NULL_HANDLE }, | |
1683 | { fn2, NULL, NULL_HANDLE }, | |
1684 | { fn3, NULL, NULL_HANDLE } | |
4358a253 | 1685 | }; |
d1b91892 AD |
1686 | |
1687 | static void | |
1688 | Pcb(index, buffer) | |
4358a253 SS |
1689 | int index; |
1690 | char * buffer; | |
d1b91892 | 1691 | { |
4358a253 | 1692 | dSP; |
d1b91892 | 1693 | |
4358a253 SS |
1694 | PUSHMARK(SP); |
1695 | XPUSHs(sv_2mortal(newSVpv(buffer, 0))); | |
1696 | PUTBACK; | |
d1b91892 AD |
1697 | |
1698 | /* Call the Perl sub */ | |
4358a253 | 1699 | call_sv(Map[index].PerlSub, G_DISCARD); |
d1b91892 AD |
1700 | } |
1701 | ||
1702 | static void | |
1703 | fn1(buffer) | |
4358a253 | 1704 | char * buffer; |
d1b91892 | 1705 | { |
4358a253 | 1706 | Pcb(0, buffer); |
d1b91892 AD |
1707 | } |
1708 | ||
1709 | static void | |
1710 | fn2(buffer) | |
4358a253 | 1711 | char * buffer; |
d1b91892 | 1712 | { |
4358a253 | 1713 | Pcb(1, buffer); |
d1b91892 AD |
1714 | } |
1715 | ||
1716 | static void | |
1717 | fn3(buffer) | |
4358a253 | 1718 | char * buffer; |
d1b91892 | 1719 | { |
4358a253 | 1720 | Pcb(2, buffer); |
d1b91892 AD |
1721 | } |
1722 | ||
1723 | void | |
1724 | array_asynch_read(fh, callback) | |
1725 | int fh | |
1726 | SV * callback | |
1727 | CODE: | |
4358a253 SS |
1728 | int index; |
1729 | int null_index = MAX_CB; | |
d1b91892 AD |
1730 | |
1731 | /* Find the same handle or an empty entry */ | |
4358a253 | 1732 | for (index = 0; index < MAX_CB; ++index) |
d1b91892 AD |
1733 | { |
1734 | if (Map[index].Handle == fh) | |
4358a253 | 1735 | break; |
d1b91892 AD |
1736 | |
1737 | if (Map[index].Handle == NULL_HANDLE) | |
4358a253 | 1738 | null_index = index; |
d1b91892 AD |
1739 | } |
1740 | ||
1741 | if (index == MAX_CB && null_index == MAX_CB) | |
4358a253 | 1742 | croak ("Too many callback functions registered\n"); |
d1b91892 AD |
1743 | |
1744 | if (index == MAX_CB) | |
4358a253 | 1745 | index = null_index; |
d1b91892 AD |
1746 | |
1747 | /* Save the file handle */ | |
4358a253 | 1748 | Map[index].Handle = fh; |
d1b91892 AD |
1749 | |
1750 | /* Remember the Perl sub */ | |
1751 | if (Map[index].PerlSub == (SV*)NULL) | |
4358a253 | 1752 | Map[index].PerlSub = newSVsv(callback); |
d1b91892 | 1753 | else |
4358a253 | 1754 | SvSetSV(Map[index].PerlSub, callback); |
d1b91892 | 1755 | |
4358a253 | 1756 | asynch_read(fh, Map[index].Function); |
d1b91892 AD |
1757 | |
1758 | void | |
1759 | array_asynch_close(fh) | |
1760 | int fh | |
1761 | CODE: | |
4358a253 | 1762 | int index; |
d1b91892 AD |
1763 | |
1764 | /* Find the file handle */ | |
4358a253 | 1765 | for (index = 0; index < MAX_CB; ++ index) |
d1b91892 | 1766 | if (Map[index].Handle == fh) |
4358a253 | 1767 | break; |
d1b91892 AD |
1768 | |
1769 | if (index == MAX_CB) | |
4358a253 | 1770 | croak ("could not close fh %d\n", fh); |
d1b91892 | 1771 | |
4358a253 SS |
1772 | Map[index].Handle = NULL_HANDLE; |
1773 | SvREFCNT_dec(Map[index].PerlSub); | |
1774 | Map[index].PerlSub = (SV*)NULL; | |
d1b91892 | 1775 | |
4358a253 | 1776 | asynch_close(fh); |
d1b91892 | 1777 | |
5f05dabc | 1778 | In this case the functions C<fn1>, C<fn2>, and C<fn3> are used to |
d1b91892 | 1779 | remember the Perl subroutine to be called. Each of the functions holds |
4a6725af | 1780 | a separate hard-wired index which is used in the function C<Pcb> to |
d1b91892 AD |
1781 | access the C<Map> array and actually call the Perl subroutine. |
1782 | ||
1783 | There are some obvious disadvantages with this technique. | |
1784 | ||
1785 | Firstly, the code is considerably more complex than with the previous | |
1786 | example. | |
1787 | ||
4a6725af | 1788 | Secondly, there is a hard-wired limit (in this case 3) to the number of |
d1b91892 AD |
1789 | callbacks that can exist simultaneously. The only way to increase the |
1790 | limit is by modifying the code to add more functions and then | |
54310121 | 1791 | recompiling. None the less, as long as the number of functions is |
d1b91892 AD |
1792 | chosen with some care, it is still a workable solution and in some |
1793 | cases is the only one available. | |
1794 | ||
1795 | To summarize, here are a number of possible methods for you to consider | |
1796 | for storing the mapping between C and the Perl callback | |
1797 | ||
1798 | =over 5 | |
1799 | ||
1800 | =item 1. Ignore the problem - Allow only 1 callback | |
1801 | ||
1802 | For a lot of situations, like interfacing to an error handler, this may | |
1803 | be a perfectly adequate solution. | |
1804 | ||
1805 | =item 2. Create a sequence of callbacks - hard wired limit | |
1806 | ||
1807 | If it is impossible to tell from the parameters passed back from the C | |
1808 | callback what the context is, then you may need to create a sequence of C | |
1809 | callback interface functions, and store pointers to each in an array. | |
1810 | ||
1811 | =item 3. Use a parameter to map to the Perl callback | |
1812 | ||
1813 | A hash is an ideal mechanism to store the mapping between C and Perl. | |
1814 | ||
1815 | =back | |
a0d0e21e | 1816 | |
a0d0e21e LW |
1817 | |
1818 | =head2 Alternate Stack Manipulation | |
1819 | ||
a0d0e21e | 1820 | |
d1b91892 AD |
1821 | Although I have made use of only the C<POP*> macros to access values |
1822 | returned from Perl subroutines, it is also possible to bypass these | |
8e07c86e | 1823 | macros and read the stack using the C<ST> macro (See L<perlxs> for a |
d1b91892 AD |
1824 | full description of the C<ST> macro). |
1825 | ||
1d45ec27 | 1826 | Most of the time the C<POP*> macros should be adequate; the main |
d1b91892 AD |
1827 | problem with them is that they force you to process the returned values |
1828 | in sequence. This may not be the most suitable way to process the | |
1829 | values in some cases. What we want is to be able to access the stack in | |
1830 | a random order. The C<ST> macro as used when coding an XSUB is ideal | |
1831 | for this purpose. | |
1832 | ||
d0554719 | 1833 | The code below is the example given in the section L</Returning a List |
1d45ec27 | 1834 | of Values> recoded to use C<ST> instead of C<POP*>. |
d1b91892 AD |
1835 | |
1836 | static void | |
1837 | call_AddSubtract2(a, b) | |
4358a253 SS |
1838 | int a; |
1839 | int b; | |
d1b91892 | 1840 | { |
4358a253 SS |
1841 | dSP; |
1842 | I32 ax; | |
1843 | int count; | |
d1b91892 | 1844 | |
4358a253 | 1845 | ENTER; |
d1b91892 AD |
1846 | SAVETMPS; |
1847 | ||
4358a253 | 1848 | PUSHMARK(SP); |
d0554719 TC |
1849 | EXTEND(SP, 2); |
1850 | PUSHs(sv_2mortal(newSViv(a))); | |
1851 | PUSHs(sv_2mortal(newSViv(b))); | |
4358a253 | 1852 | PUTBACK; |
d1b91892 | 1853 | |
4929bf7b | 1854 | count = call_pv("AddSubtract", G_ARRAY); |
d1b91892 | 1855 | |
4358a253 SS |
1856 | SPAGAIN; |
1857 | SP -= count; | |
1858 | ax = (SP - PL_stack_base) + 1; | |
d1b91892 AD |
1859 | |
1860 | if (count != 2) | |
4358a253 | 1861 | croak("Big trouble\n"); |
a0d0e21e | 1862 | |
4358a253 SS |
1863 | printf ("%d + %d = %d\n", a, b, SvIV(ST(0))); |
1864 | printf ("%d - %d = %d\n", a, b, SvIV(ST(1))); | |
d1b91892 | 1865 | |
4358a253 SS |
1866 | PUTBACK; |
1867 | FREETMPS; | |
1868 | LEAVE; | |
d1b91892 AD |
1869 | } |
1870 | ||
1871 | Notes | |
1872 | ||
1873 | =over 5 | |
1874 | ||
1875 | =item 1. | |
1876 | ||
1877 | Notice that it was necessary to define the variable C<ax>. This is | |
1878 | because the C<ST> macro expects it to exist. If we were in an XSUB it | |
1879 | would not be necessary to define C<ax> as it is already defined for | |
1d45ec27 | 1880 | us. |
d1b91892 AD |
1881 | |
1882 | =item 2. | |
1883 | ||
1884 | The code | |
1885 | ||
4358a253 SS |
1886 | SPAGAIN; |
1887 | SP -= count; | |
1888 | ax = (SP - PL_stack_base) + 1; | |
d1b91892 AD |
1889 | |
1890 | sets the stack up so that we can use the C<ST> macro. | |
1891 | ||
1892 | =item 3. | |
1893 | ||
1894 | Unlike the original coding of this example, the returned | |
1895 | values are not accessed in reverse order. So C<ST(0)> refers to the | |
54310121 | 1896 | first value returned by the Perl subroutine and C<ST(count-1)> |
d1b91892 AD |
1897 | refers to the last. |
1898 | ||
1899 | =back | |
a0d0e21e | 1900 | |
02f6dca1 | 1901 | =head2 Creating and Calling an Anonymous Subroutine in C |
8f183262 | 1902 | |
4929bf7b | 1903 | As we've already shown, C<call_sv> can be used to invoke an |
c2611fb3 GS |
1904 | anonymous subroutine. However, our example showed a Perl script |
1905 | invoking an XSUB to perform this operation. Let's see how it can be | |
8f183262 DM |
1906 | done inside our C code: |
1907 | ||
8f183262 DM |
1908 | ... |
1909 | ||
e46aa1dd KW |
1910 | SV *cvrv |
1911 | = eval_pv("sub { | |
1912 | print 'You will not find me cluttering any namespace!' | |
1913 | }", TRUE); | |
8f183262 DM |
1914 | |
1915 | ... | |
1916 | ||
4929bf7b | 1917 | call_sv(cvrv, G_VOID|G_NOARGS); |
8f183262 | 1918 | |
4929bf7b GS |
1919 | C<eval_pv> is used to compile the anonymous subroutine, which |
1920 | will be the return value as well (read more about C<eval_pv> in | |
4a4eefd0 | 1921 | L<perlapi/eval_pv>). Once this code reference is in hand, it |
8f183262 DM |
1922 | can be mixed in with all the previous examples we've shown. |
1923 | ||
9850bf21 RH |
1924 | =head1 LIGHTWEIGHT CALLBACKS |
1925 | ||
1926 | Sometimes you need to invoke the same subroutine repeatedly. | |
1927 | This usually happens with a function that acts on a list of | |
1928 | values, such as Perl's built-in sort(). You can pass a | |
1929 | comparison function to sort(), which will then be invoked | |
1930 | for every pair of values that needs to be compared. The first() | |
1931 | and reduce() functions from L<List::Util> follow a similar | |
1932 | pattern. | |
1933 | ||
1934 | In this case it is possible to speed up the routine (often | |
1935 | quite substantially) by using the lightweight callback API. | |
1936 | The idea is that the calling context only needs to be | |
1937 | created and destroyed once, and the sub can be called | |
1938 | arbitrarily many times in between. | |
1939 | ||
ac036724 | 1940 | It is usual to pass parameters using global variables (typically |
1941 | $_ for one parameter, or $a and $b for two parameters) rather | |
9850bf21 RH |
1942 | than via @_. (It is possible to use the @_ mechanism if you know |
1943 | what you're doing, though there is as yet no supported API for | |
1944 | it. It's also inherently slower.) | |
1945 | ||
1946 | The pattern of macro calls is like this: | |
1947 | ||
82f35e8b | 1948 | dMULTICALL; /* Declare local variables */ |
1c23e2bd | 1949 | U8 gimme = G_SCALAR; /* context of the call: G_SCALAR, |
782a81f5 | 1950 | * G_ARRAY, or G_VOID */ |
9850bf21 | 1951 | |
82f35e8b RH |
1952 | PUSH_MULTICALL(cv); /* Set up the context for calling cv, |
1953 | and set local vars appropriately */ | |
9850bf21 RH |
1954 | |
1955 | /* loop */ { | |
1956 | /* set the value(s) af your parameter variables */ | |
1957 | MULTICALL; /* Make the actual call */ | |
1958 | } /* end of loop */ | |
1959 | ||
1960 | POP_MULTICALL; /* Tear down the calling context */ | |
1961 | ||
1962 | For some concrete examples, see the implementation of the | |
1963 | first() and reduce() functions of List::Util 1.18. There you | |
1964 | will also find a header file that emulates the multicall API | |
1965 | on older versions of perl. | |
1966 | ||
a0d0e21e LW |
1967 | =head1 SEE ALSO |
1968 | ||
8e07c86e | 1969 | L<perlxs>, L<perlguts>, L<perlembed> |
a0d0e21e LW |
1970 | |
1971 | =head1 AUTHOR | |
1972 | ||
0536e0eb | 1973 | Paul Marquess |
a0d0e21e | 1974 | |
d1b91892 AD |
1975 | Special thanks to the following people who assisted in the creation of |
1976 | the document. | |
a0d0e21e | 1977 | |
c07a80fd | 1978 | Jeff Okamoto, Tim Bunce, Nick Gianniotis, Steve Kelem, Gurusamy Sarathy |
1979 | and Larry Wall. | |
a0d0e21e LW |
1980 | |
1981 | =head1 DATE | |
1982 | ||
d0554719 | 1983 | Last updated for perl 5.23.1. |