Commit | Line | Data |
---|---|---|
b8d4c8d0 GM |
1 | @c -*-texinfo-*- |
2 | @c This is part of the GNU Emacs Lisp Reference Manual. | |
73b0cd50 | 3 | @c Copyright (C) 1990-1995, 1998, 2001-2011 Free Software Foundation, Inc. |
b8d4c8d0 | 4 | @c See the file elisp.texi for copying conditions. |
6336d8c3 | 5 | @setfilename ../../info/macros |
b8d4c8d0 GM |
6 | @node Macros, Customization, Functions, Top |
7 | @chapter Macros | |
8 | @cindex macros | |
9 | ||
10 | @dfn{Macros} enable you to define new control constructs and other | |
11 | language features. A macro is defined much like a function, but instead | |
12 | of telling how to compute a value, it tells how to compute another Lisp | |
13 | expression which will in turn compute the value. We call this | |
14 | expression the @dfn{expansion} of the macro. | |
15 | ||
16 | Macros can do this because they operate on the unevaluated expressions | |
17 | for the arguments, not on the argument values as functions do. They can | |
18 | therefore construct an expansion containing these argument expressions | |
19 | or parts of them. | |
20 | ||
21 | If you are using a macro to do something an ordinary function could | |
22 | do, just for the sake of speed, consider using an inline function | |
23 | instead. @xref{Inline Functions}. | |
24 | ||
25 | @menu | |
26 | * Simple Macro:: A basic example. | |
27 | * Expansion:: How, when and why macros are expanded. | |
28 | * Compiling Macros:: How macros are expanded by the compiler. | |
29 | * Defining Macros:: How to write a macro definition. | |
30 | * Backquote:: Easier construction of list structure. | |
31 | * Problems with Macros:: Don't evaluate the macro arguments too many times. | |
32 | Don't hide the user's variables. | |
33 | * Indenting Macros:: Specifying how to indent macro calls. | |
34 | @end menu | |
35 | ||
36 | @node Simple Macro | |
37 | @section A Simple Example of a Macro | |
38 | ||
39 | Suppose we would like to define a Lisp construct to increment a | |
40 | variable value, much like the @code{++} operator in C. We would like to | |
41 | write @code{(inc x)} and have the effect of @code{(setq x (1+ x))}. | |
42 | Here's a macro definition that does the job: | |
43 | ||
44 | @findex inc | |
45 | @example | |
46 | @group | |
47 | (defmacro inc (var) | |
48 | (list 'setq var (list '1+ var))) | |
49 | @end group | |
50 | @end example | |
51 | ||
52 | When this is called with @code{(inc x)}, the argument @var{var} is the | |
53 | symbol @code{x}---@emph{not} the @emph{value} of @code{x}, as it would | |
54 | be in a function. The body of the macro uses this to construct the | |
55 | expansion, which is @code{(setq x (1+ x))}. Once the macro definition | |
56 | returns this expansion, Lisp proceeds to evaluate it, thus incrementing | |
57 | @code{x}. | |
58 | ||
59 | @node Expansion | |
60 | @section Expansion of a Macro Call | |
61 | @cindex expansion of macros | |
62 | @cindex macro call | |
63 | ||
64 | A macro call looks just like a function call in that it is a list which | |
65 | starts with the name of the macro. The rest of the elements of the list | |
66 | are the arguments of the macro. | |
67 | ||
68 | Evaluation of the macro call begins like evaluation of a function call | |
69 | except for one crucial difference: the macro arguments are the actual | |
70 | expressions appearing in the macro call. They are not evaluated before | |
71 | they are given to the macro definition. By contrast, the arguments of a | |
72 | function are results of evaluating the elements of the function call | |
73 | list. | |
74 | ||
75 | Having obtained the arguments, Lisp invokes the macro definition just | |
76 | as a function is invoked. The argument variables of the macro are bound | |
77 | to the argument values from the macro call, or to a list of them in the | |
78 | case of a @code{&rest} argument. And the macro body executes and | |
79 | returns its value just as a function body does. | |
80 | ||
81 | The second crucial difference between macros and functions is that the | |
82 | value returned by the macro body is not the value of the macro call. | |
83 | Instead, it is an alternate expression for computing that value, also | |
84 | known as the @dfn{expansion} of the macro. The Lisp interpreter | |
85 | proceeds to evaluate the expansion as soon as it comes back from the | |
86 | macro. | |
87 | ||
88 | Since the expansion is evaluated in the normal manner, it may contain | |
89 | calls to other macros. It may even be a call to the same macro, though | |
90 | this is unusual. | |
91 | ||
92 | You can see the expansion of a given macro call by calling | |
93 | @code{macroexpand}. | |
94 | ||
95 | @defun macroexpand form &optional environment | |
96 | @cindex macro expansion | |
97 | This function expands @var{form}, if it is a macro call. If the result | |
98 | is another macro call, it is expanded in turn, until something which is | |
99 | not a macro call results. That is the value returned by | |
100 | @code{macroexpand}. If @var{form} is not a macro call to begin with, it | |
101 | is returned as given. | |
102 | ||
103 | Note that @code{macroexpand} does not look at the subexpressions of | |
104 | @var{form} (although some macro definitions may do so). Even if they | |
105 | are macro calls themselves, @code{macroexpand} does not expand them. | |
106 | ||
107 | The function @code{macroexpand} does not expand calls to inline functions. | |
108 | Normally there is no need for that, since a call to an inline function is | |
109 | no harder to understand than a call to an ordinary function. | |
110 | ||
111 | If @var{environment} is provided, it specifies an alist of macro | |
112 | definitions that shadow the currently defined macros. Byte compilation | |
113 | uses this feature. | |
114 | ||
115 | @smallexample | |
116 | @group | |
117 | (defmacro inc (var) | |
118 | (list 'setq var (list '1+ var))) | |
119 | @result{} inc | |
120 | @end group | |
121 | ||
122 | @group | |
123 | (macroexpand '(inc r)) | |
124 | @result{} (setq r (1+ r)) | |
125 | @end group | |
126 | ||
127 | @group | |
128 | (defmacro inc2 (var1 var2) | |
129 | (list 'progn (list 'inc var1) (list 'inc var2))) | |
130 | @result{} inc2 | |
131 | @end group | |
132 | ||
133 | @group | |
134 | (macroexpand '(inc2 r s)) | |
135 | @result{} (progn (inc r) (inc s)) ; @r{@code{inc} not expanded here.} | |
136 | @end group | |
137 | @end smallexample | |
138 | @end defun | |
139 | ||
140 | ||
141 | @defun macroexpand-all form &optional environment | |
142 | @code{macroexpand-all} expands macros like @code{macroexpand}, but | |
143 | will look for and expand all macros in @var{form}, not just at the | |
144 | top-level. If no macros are expanded, the return value is @code{eq} | |
145 | to @var{form}. | |
146 | ||
147 | Repeating the example used for @code{macroexpand} above with | |
148 | @code{macroexpand-all}, we see that @code{macroexpand-all} @emph{does} | |
149 | expand the embedded calls to @code{inc}: | |
150 | ||
151 | @smallexample | |
152 | (macroexpand-all '(inc2 r s)) | |
153 | @result{} (progn (setq r (1+ r)) (setq s (1+ s))) | |
154 | @end smallexample | |
155 | ||
156 | @end defun | |
157 | ||
158 | @node Compiling Macros | |
159 | @section Macros and Byte Compilation | |
160 | @cindex byte-compiling macros | |
161 | ||
162 | You might ask why we take the trouble to compute an expansion for a | |
163 | macro and then evaluate the expansion. Why not have the macro body | |
164 | produce the desired results directly? The reason has to do with | |
165 | compilation. | |
166 | ||
167 | When a macro call appears in a Lisp program being compiled, the Lisp | |
168 | compiler calls the macro definition just as the interpreter would, and | |
169 | receives an expansion. But instead of evaluating this expansion, it | |
170 | compiles the expansion as if it had appeared directly in the program. | |
171 | As a result, the compiled code produces the value and side effects | |
172 | intended for the macro, but executes at full compiled speed. This would | |
173 | not work if the macro body computed the value and side effects | |
174 | itself---they would be computed at compile time, which is not useful. | |
175 | ||
176 | In order for compilation of macro calls to work, the macros must | |
177 | already be defined in Lisp when the calls to them are compiled. The | |
178 | compiler has a special feature to help you do this: if a file being | |
179 | compiled contains a @code{defmacro} form, the macro is defined | |
ebfbce67 CY |
180 | temporarily for the rest of the compilation of that file. |
181 | ||
182 | Byte-compiling a file also executes any @code{require} calls at | |
183 | top-level in the file, so you can ensure that necessary macro | |
184 | definitions are available during compilation by requiring the files | |
185 | that define them (@pxref{Named Features}). To avoid loading the macro | |
186 | definition files when someone @emph{runs} the compiled program, write | |
b8d4c8d0 GM |
187 | @code{eval-when-compile} around the @code{require} calls (@pxref{Eval |
188 | During Compile}). | |
189 | ||
190 | @node Defining Macros | |
191 | @section Defining Macros | |
192 | ||
193 | A Lisp macro is a list whose @sc{car} is @code{macro}. Its @sc{cdr} should | |
194 | be a function; expansion of the macro works by applying the function | |
195 | (with @code{apply}) to the list of unevaluated argument-expressions | |
196 | from the macro call. | |
197 | ||
198 | It is possible to use an anonymous Lisp macro just like an anonymous | |
199 | function, but this is never done, because it does not make sense to pass | |
200 | an anonymous macro to functionals such as @code{mapcar}. In practice, | |
201 | all Lisp macros have names, and they are usually defined with the | |
202 | special form @code{defmacro}. | |
203 | ||
204 | @defspec defmacro name argument-list body-forms@dots{} | |
205 | @code{defmacro} defines the symbol @var{name} as a macro that looks | |
206 | like this: | |
207 | ||
208 | @example | |
209 | (macro lambda @var{argument-list} . @var{body-forms}) | |
210 | @end example | |
211 | ||
212 | (Note that the @sc{cdr} of this list is a function---a lambda expression.) | |
213 | This macro object is stored in the function cell of @var{name}. The | |
214 | value returned by evaluating the @code{defmacro} form is @var{name}, but | |
215 | usually we ignore this value. | |
216 | ||
217 | The shape and meaning of @var{argument-list} is the same as in a | |
218 | function, and the keywords @code{&rest} and @code{&optional} may be used | |
219 | (@pxref{Argument List}). Macros may have a documentation string, but | |
220 | any @code{interactive} declaration is ignored since macros cannot be | |
221 | called interactively. | |
222 | @end defspec | |
223 | ||
224 | The body of the macro definition can include a @code{declare} form, | |
225 | which can specify how @key{TAB} should indent macro calls, and how to | |
226 | step through them for Edebug. | |
227 | ||
228 | @defmac declare @var{specs}@dots{} | |
229 | @anchor{Definition of declare} | |
230 | A @code{declare} form is used in a macro definition to specify various | |
231 | additional information about it. Two kinds of specification are | |
232 | currently supported: | |
233 | ||
234 | @table @code | |
235 | @item (debug @var{edebug-form-spec}) | |
236 | Specify how to step through macro calls for Edebug. | |
237 | @xref{Instrumenting Macro Calls}. | |
238 | ||
239 | @item (indent @var{indent-spec}) | |
240 | Specify how to indent calls to this macro. @xref{Indenting Macros}, | |
241 | for more details. | |
242 | @end table | |
243 | ||
244 | A @code{declare} form only has its special effect in the body of a | |
245 | @code{defmacro} form if it immediately follows the documentation | |
246 | string, if present, or the argument list otherwise. (Strictly | |
247 | speaking, @emph{several} @code{declare} forms can follow the | |
248 | documentation string or argument list, but since a @code{declare} form | |
249 | can have several @var{specs}, they can always be combined into a | |
250 | single form.) When used at other places in a @code{defmacro} form, or | |
251 | outside a @code{defmacro} form, @code{declare} just returns @code{nil} | |
252 | without evaluating any @var{specs}. | |
253 | @end defmac | |
254 | ||
255 | No macro absolutely needs a @code{declare} form, because that form | |
256 | has no effect on how the macro expands, on what the macro means in the | |
257 | program. It only affects secondary features: indentation and Edebug. | |
258 | ||
259 | @node Backquote | |
260 | @section Backquote | |
261 | @cindex backquote (list substitution) | |
262 | @cindex ` (list substitution) | |
263 | @findex ` | |
264 | ||
265 | Macros often need to construct large list structures from a mixture of | |
266 | constants and nonconstant parts. To make this easier, use the @samp{`} | |
267 | syntax (usually called @dfn{backquote}). | |
268 | ||
269 | Backquote allows you to quote a list, but selectively evaluate | |
270 | elements of that list. In the simplest case, it is identical to the | |
271 | special form @code{quote} (@pxref{Quoting}). For example, these | |
272 | two forms yield identical results: | |
273 | ||
274 | @example | |
275 | @group | |
276 | `(a list of (+ 2 3) elements) | |
277 | @result{} (a list of (+ 2 3) elements) | |
278 | @end group | |
279 | @group | |
280 | '(a list of (+ 2 3) elements) | |
281 | @result{} (a list of (+ 2 3) elements) | |
282 | @end group | |
283 | @end example | |
284 | ||
285 | @findex , @r{(with backquote)} | |
286 | The special marker @samp{,} inside of the argument to backquote | |
287 | indicates a value that isn't constant. Backquote evaluates the | |
288 | argument of @samp{,} and puts the value in the list structure: | |
289 | ||
290 | @example | |
291 | @group | |
292 | (list 'a 'list 'of (+ 2 3) 'elements) | |
293 | @result{} (a list of 5 elements) | |
294 | @end group | |
295 | @group | |
296 | `(a list of ,(+ 2 3) elements) | |
297 | @result{} (a list of 5 elements) | |
298 | @end group | |
299 | @end example | |
300 | ||
301 | Substitution with @samp{,} is allowed at deeper levels of the list | |
302 | structure also. For example: | |
303 | ||
304 | @example | |
305 | @group | |
306 | (defmacro t-becomes-nil (variable) | |
307 | `(if (eq ,variable t) | |
308 | (setq ,variable nil))) | |
309 | @end group | |
310 | ||
311 | @group | |
312 | (t-becomes-nil foo) | |
313 | @equiv{} (if (eq foo t) (setq foo nil)) | |
314 | @end group | |
315 | @end example | |
316 | ||
317 | @findex ,@@ @r{(with backquote)} | |
318 | @cindex splicing (with backquote) | |
319 | You can also @dfn{splice} an evaluated value into the resulting list, | |
320 | using the special marker @samp{,@@}. The elements of the spliced list | |
321 | become elements at the same level as the other elements of the resulting | |
322 | list. The equivalent code without using @samp{`} is often unreadable. | |
323 | Here are some examples: | |
324 | ||
325 | @example | |
326 | @group | |
327 | (setq some-list '(2 3)) | |
328 | @result{} (2 3) | |
329 | @end group | |
330 | @group | |
331 | (cons 1 (append some-list '(4) some-list)) | |
332 | @result{} (1 2 3 4 2 3) | |
333 | @end group | |
334 | @group | |
335 | `(1 ,@@some-list 4 ,@@some-list) | |
336 | @result{} (1 2 3 4 2 3) | |
337 | @end group | |
338 | ||
339 | @group | |
340 | (setq list '(hack foo bar)) | |
341 | @result{} (hack foo bar) | |
342 | @end group | |
343 | @group | |
344 | (cons 'use | |
345 | (cons 'the | |
346 | (cons 'words (append (cdr list) '(as elements))))) | |
347 | @result{} (use the words foo bar as elements) | |
348 | @end group | |
349 | @group | |
350 | `(use the words ,@@(cdr list) as elements) | |
351 | @result{} (use the words foo bar as elements) | |
352 | @end group | |
353 | @end example | |
354 | ||
b8d4c8d0 GM |
355 | @node Problems with Macros |
356 | @section Common Problems Using Macros | |
357 | ||
358 | The basic facts of macro expansion have counterintuitive consequences. | |
359 | This section describes some important consequences that can lead to | |
360 | trouble, and rules to follow to avoid trouble. | |
361 | ||
362 | @menu | |
363 | * Wrong Time:: Do the work in the expansion, not in the macro. | |
364 | * Argument Evaluation:: The expansion should evaluate each macro arg once. | |
365 | * Surprising Local Vars:: Local variable bindings in the expansion | |
366 | require special care. | |
367 | * Eval During Expansion:: Don't evaluate them; put them in the expansion. | |
368 | * Repeated Expansion:: Avoid depending on how many times expansion is done. | |
369 | @end menu | |
370 | ||
371 | @node Wrong Time | |
372 | @subsection Wrong Time | |
373 | ||
374 | The most common problem in writing macros is doing some of the | |
375 | real work prematurely---while expanding the macro, rather than in the | |
376 | expansion itself. For instance, one real package had this macro | |
377 | definition: | |
378 | ||
379 | @example | |
380 | (defmacro my-set-buffer-multibyte (arg) | |
381 | (if (fboundp 'set-buffer-multibyte) | |
382 | (set-buffer-multibyte arg))) | |
383 | @end example | |
384 | ||
385 | With this erroneous macro definition, the program worked fine when | |
386 | interpreted but failed when compiled. This macro definition called | |
387 | @code{set-buffer-multibyte} during compilation, which was wrong, and | |
388 | then did nothing when the compiled package was run. The definition | |
389 | that the programmer really wanted was this: | |
390 | ||
391 | @example | |
392 | (defmacro my-set-buffer-multibyte (arg) | |
393 | (if (fboundp 'set-buffer-multibyte) | |
394 | `(set-buffer-multibyte ,arg))) | |
395 | @end example | |
396 | ||
397 | @noindent | |
398 | This macro expands, if appropriate, into a call to | |
399 | @code{set-buffer-multibyte} that will be executed when the compiled | |
400 | program is actually run. | |
401 | ||
402 | @node Argument Evaluation | |
403 | @subsection Evaluating Macro Arguments Repeatedly | |
404 | ||
405 | When defining a macro you must pay attention to the number of times | |
406 | the arguments will be evaluated when the expansion is executed. The | |
407 | following macro (used to facilitate iteration) illustrates the problem. | |
408 | This macro allows us to write a simple ``for'' loop such as one might | |
409 | find in Pascal. | |
410 | ||
411 | @findex for | |
412 | @smallexample | |
413 | @group | |
414 | (defmacro for (var from init to final do &rest body) | |
415 | "Execute a simple \"for\" loop. | |
416 | For example, (for i from 1 to 10 do (print i))." | |
417 | (list 'let (list (list var init)) | |
418 | (cons 'while (cons (list '<= var final) | |
419 | (append body (list (list 'inc var))))))) | |
420 | @end group | |
421 | @result{} for | |
422 | ||
423 | @group | |
424 | (for i from 1 to 3 do | |
425 | (setq square (* i i)) | |
426 | (princ (format "\n%d %d" i square))) | |
427 | @expansion{} | |
428 | @end group | |
429 | @group | |
430 | (let ((i 1)) | |
431 | (while (<= i 3) | |
432 | (setq square (* i i)) | |
433 | (princ (format "\n%d %d" i square)) | |
434 | (inc i))) | |
435 | @end group | |
436 | @group | |
437 | ||
438 | @print{}1 1 | |
439 | @print{}2 4 | |
440 | @print{}3 9 | |
441 | @result{} nil | |
442 | @end group | |
443 | @end smallexample | |
444 | ||
445 | @noindent | |
446 | The arguments @code{from}, @code{to}, and @code{do} in this macro are | |
447 | ``syntactic sugar''; they are entirely ignored. The idea is that you | |
448 | will write noise words (such as @code{from}, @code{to}, and @code{do}) | |
449 | in those positions in the macro call. | |
450 | ||
451 | Here's an equivalent definition simplified through use of backquote: | |
452 | ||
453 | @smallexample | |
454 | @group | |
455 | (defmacro for (var from init to final do &rest body) | |
456 | "Execute a simple \"for\" loop. | |
457 | For example, (for i from 1 to 10 do (print i))." | |
458 | `(let ((,var ,init)) | |
459 | (while (<= ,var ,final) | |
460 | ,@@body | |
461 | (inc ,var)))) | |
462 | @end group | |
463 | @end smallexample | |
464 | ||
465 | Both forms of this definition (with backquote and without) suffer from | |
466 | the defect that @var{final} is evaluated on every iteration. If | |
467 | @var{final} is a constant, this is not a problem. If it is a more | |
468 | complex form, say @code{(long-complex-calculation x)}, this can slow | |
469 | down the execution significantly. If @var{final} has side effects, | |
470 | executing it more than once is probably incorrect. | |
471 | ||
472 | @cindex macro argument evaluation | |
473 | A well-designed macro definition takes steps to avoid this problem by | |
474 | producing an expansion that evaluates the argument expressions exactly | |
475 | once unless repeated evaluation is part of the intended purpose of the | |
476 | macro. Here is a correct expansion for the @code{for} macro: | |
477 | ||
478 | @smallexample | |
479 | @group | |
480 | (let ((i 1) | |
481 | (max 3)) | |
482 | (while (<= i max) | |
483 | (setq square (* i i)) | |
484 | (princ (format "%d %d" i square)) | |
485 | (inc i))) | |
486 | @end group | |
487 | @end smallexample | |
488 | ||
489 | Here is a macro definition that creates this expansion: | |
490 | ||
491 | @smallexample | |
492 | @group | |
493 | (defmacro for (var from init to final do &rest body) | |
494 | "Execute a simple for loop: (for i from 1 to 10 do (print i))." | |
495 | `(let ((,var ,init) | |
496 | (max ,final)) | |
497 | (while (<= ,var max) | |
498 | ,@@body | |
499 | (inc ,var)))) | |
500 | @end group | |
501 | @end smallexample | |
502 | ||
503 | Unfortunately, this fix introduces another problem, | |
504 | described in the following section. | |
505 | ||
506 | @node Surprising Local Vars | |
507 | @subsection Local Variables in Macro Expansions | |
508 | ||
509 | @ifnottex | |
510 | In the previous section, the definition of @code{for} was fixed as | |
511 | follows to make the expansion evaluate the macro arguments the proper | |
512 | number of times: | |
513 | ||
514 | @smallexample | |
515 | @group | |
516 | (defmacro for (var from init to final do &rest body) | |
517 | "Execute a simple for loop: (for i from 1 to 10 do (print i))." | |
518 | @end group | |
519 | @group | |
520 | `(let ((,var ,init) | |
521 | (max ,final)) | |
522 | (while (<= ,var max) | |
523 | ,@@body | |
524 | (inc ,var)))) | |
525 | @end group | |
526 | @end smallexample | |
527 | @end ifnottex | |
528 | ||
529 | The new definition of @code{for} has a new problem: it introduces a | |
530 | local variable named @code{max} which the user does not expect. This | |
531 | causes trouble in examples such as the following: | |
532 | ||
533 | @smallexample | |
534 | @group | |
535 | (let ((max 0)) | |
536 | (for x from 0 to 10 do | |
537 | (let ((this (frob x))) | |
538 | (if (< max this) | |
539 | (setq max this))))) | |
540 | @end group | |
541 | @end smallexample | |
542 | ||
543 | @noindent | |
544 | The references to @code{max} inside the body of the @code{for}, which | |
545 | are supposed to refer to the user's binding of @code{max}, really access | |
546 | the binding made by @code{for}. | |
547 | ||
548 | The way to correct this is to use an uninterned symbol instead of | |
549 | @code{max} (@pxref{Creating Symbols}). The uninterned symbol can be | |
550 | bound and referred to just like any other symbol, but since it is | |
551 | created by @code{for}, we know that it cannot already appear in the | |
552 | user's program. Since it is not interned, there is no way the user can | |
553 | put it into the program later. It will never appear anywhere except | |
554 | where put by @code{for}. Here is a definition of @code{for} that works | |
555 | this way: | |
556 | ||
557 | @smallexample | |
558 | @group | |
559 | (defmacro for (var from init to final do &rest body) | |
560 | "Execute a simple for loop: (for i from 1 to 10 do (print i))." | |
561 | (let ((tempvar (make-symbol "max"))) | |
562 | `(let ((,var ,init) | |
563 | (,tempvar ,final)) | |
564 | (while (<= ,var ,tempvar) | |
565 | ,@@body | |
566 | (inc ,var))))) | |
567 | @end group | |
568 | @end smallexample | |
569 | ||
570 | @noindent | |
571 | This creates an uninterned symbol named @code{max} and puts it in the | |
572 | expansion instead of the usual interned symbol @code{max} that appears | |
573 | in expressions ordinarily. | |
574 | ||
575 | @node Eval During Expansion | |
576 | @subsection Evaluating Macro Arguments in Expansion | |
577 | ||
578 | Another problem can happen if the macro definition itself | |
579 | evaluates any of the macro argument expressions, such as by calling | |
580 | @code{eval} (@pxref{Eval}). If the argument is supposed to refer to the | |
581 | user's variables, you may have trouble if the user happens to use a | |
582 | variable with the same name as one of the macro arguments. Inside the | |
583 | macro body, the macro argument binding is the most local binding of this | |
584 | variable, so any references inside the form being evaluated do refer to | |
585 | it. Here is an example: | |
586 | ||
587 | @example | |
588 | @group | |
589 | (defmacro foo (a) | |
590 | (list 'setq (eval a) t)) | |
591 | @result{} foo | |
592 | @end group | |
593 | @group | |
594 | (setq x 'b) | |
595 | (foo x) @expansion{} (setq b t) | |
596 | @result{} t ; @r{and @code{b} has been set.} | |
597 | ;; @r{but} | |
598 | (setq a 'c) | |
599 | (foo a) @expansion{} (setq a t) | |
600 | @result{} t ; @r{but this set @code{a}, not @code{c}.} | |
601 | ||
602 | @end group | |
603 | @end example | |
604 | ||
605 | It makes a difference whether the user's variable is named @code{a} or | |
606 | @code{x}, because @code{a} conflicts with the macro argument variable | |
607 | @code{a}. | |
608 | ||
609 | Another problem with calling @code{eval} in a macro definition is that | |
610 | it probably won't do what you intend in a compiled program. The | |
cc060ff7 | 611 | byte compiler runs macro definitions while compiling the program, when |
b8d4c8d0 GM |
612 | the program's own computations (which you might have wished to access |
613 | with @code{eval}) don't occur and its local variable bindings don't | |
614 | exist. | |
615 | ||
616 | To avoid these problems, @strong{don't evaluate an argument expression | |
617 | while computing the macro expansion}. Instead, substitute the | |
618 | expression into the macro expansion, so that its value will be computed | |
619 | as part of executing the expansion. This is how the other examples in | |
620 | this chapter work. | |
621 | ||
622 | @node Repeated Expansion | |
623 | @subsection How Many Times is the Macro Expanded? | |
624 | ||
625 | Occasionally problems result from the fact that a macro call is | |
626 | expanded each time it is evaluated in an interpreted function, but is | |
627 | expanded only once (during compilation) for a compiled function. If the | |
628 | macro definition has side effects, they will work differently depending | |
629 | on how many times the macro is expanded. | |
630 | ||
631 | Therefore, you should avoid side effects in computation of the | |
632 | macro expansion, unless you really know what you are doing. | |
633 | ||
634 | One special kind of side effect can't be avoided: constructing Lisp | |
635 | objects. Almost all macro expansions include constructed lists; that is | |
636 | the whole point of most macros. This is usually safe; there is just one | |
637 | case where you must be careful: when the object you construct is part of a | |
638 | quoted constant in the macro expansion. | |
639 | ||
640 | If the macro is expanded just once, in compilation, then the object is | |
641 | constructed just once, during compilation. But in interpreted | |
642 | execution, the macro is expanded each time the macro call runs, and this | |
643 | means a new object is constructed each time. | |
644 | ||
645 | In most clean Lisp code, this difference won't matter. It can matter | |
646 | only if you perform side-effects on the objects constructed by the macro | |
647 | definition. Thus, to avoid trouble, @strong{avoid side effects on | |
648 | objects constructed by macro definitions}. Here is an example of how | |
649 | such side effects can get you into trouble: | |
650 | ||
651 | @lisp | |
652 | @group | |
653 | (defmacro empty-object () | |
654 | (list 'quote (cons nil nil))) | |
655 | @end group | |
656 | ||
657 | @group | |
658 | (defun initialize (condition) | |
659 | (let ((object (empty-object))) | |
660 | (if condition | |
661 | (setcar object condition)) | |
662 | object)) | |
663 | @end group | |
664 | @end lisp | |
665 | ||
666 | @noindent | |
667 | If @code{initialize} is interpreted, a new list @code{(nil)} is | |
668 | constructed each time @code{initialize} is called. Thus, no side effect | |
669 | survives between calls. If @code{initialize} is compiled, then the | |
670 | macro @code{empty-object} is expanded during compilation, producing a | |
671 | single ``constant'' @code{(nil)} that is reused and altered each time | |
672 | @code{initialize} is called. | |
673 | ||
674 | One way to avoid pathological cases like this is to think of | |
675 | @code{empty-object} as a funny kind of constant, not as a memory | |
676 | allocation construct. You wouldn't use @code{setcar} on a constant such | |
677 | as @code{'(nil)}, so naturally you won't use it on @code{(empty-object)} | |
678 | either. | |
679 | ||
680 | @node Indenting Macros | |
681 | @section Indenting Macros | |
682 | ||
683 | You can use the @code{declare} form in the macro definition to | |
a86cd395 | 684 | specify how to @key{TAB} should indent calls to the macro. You |
b8d4c8d0 GM |
685 | write it like this: |
686 | ||
687 | @example | |
688 | (declare (indent @var{indent-spec})) | |
689 | @end example | |
690 | ||
691 | @noindent | |
692 | Here are the possibilities for @var{indent-spec}: | |
693 | ||
694 | @table @asis | |
695 | @item @code{nil} | |
696 | This is the same as no property---use the standard indentation pattern. | |
697 | @item @code{defun} | |
698 | Handle this function like a @samp{def} construct: treat the second | |
699 | line as the start of a @dfn{body}. | |
700 | @item an integer, @var{number} | |
701 | The first @var{number} arguments of the function are | |
702 | @dfn{distinguished} arguments; the rest are considered the body | |
703 | of the expression. A line in the expression is indented according to | |
704 | whether the first argument on it is distinguished or not. If the | |
705 | argument is part of the body, the line is indented @code{lisp-body-indent} | |
706 | more columns than the open-parenthesis starting the containing | |
707 | expression. If the argument is distinguished and is either the first | |
708 | or second argument, it is indented @emph{twice} that many extra columns. | |
709 | If the argument is distinguished and not the first or second argument, | |
710 | the line uses the standard pattern. | |
711 | @item a symbol, @var{symbol} | |
712 | @var{symbol} should be a function name; that function is called to | |
713 | calculate the indentation of a line within this expression. The | |
714 | function receives two arguments: | |
715 | @table @asis | |
716 | @item @var{state} | |
717 | The value returned by @code{parse-partial-sexp} (a Lisp primitive for | |
718 | indentation and nesting computation) when it parses up to the | |
719 | beginning of this line. | |
720 | @item @var{pos} | |
721 | The position at which the line being indented begins. | |
722 | @end table | |
723 | @noindent | |
724 | It should return either a number, which is the number of columns of | |
725 | indentation for that line, or a list whose car is such a number. The | |
726 | difference between returning a number and returning a list is that a | |
727 | number says that all following lines at the same nesting level should | |
728 | be indented just like this one; a list says that following lines might | |
729 | call for different indentations. This makes a difference when the | |
730 | indentation is being computed by @kbd{C-M-q}; if the value is a | |
731 | number, @kbd{C-M-q} need not recalculate indentation for the following | |
732 | lines until the end of the list. | |
733 | @end table |