xref: /freebsd/contrib/ncurses/form/fty_regex.c (revision 21817992b3314c908ab50f0bb88d2ee750b9c4ac)
1 /****************************************************************************
2  * Copyright 2018-2020,2021 Thomas E. Dickey                                *
3  * Copyright 1998-2012,2015 Free Software Foundation, Inc.                  *
4  *                                                                          *
5  * Permission is hereby granted, free of charge, to any person obtaining a  *
6  * copy of this software and associated documentation files (the            *
7  * "Software"), to deal in the Software without restriction, including      *
8  * without limitation the rights to use, copy, modify, merge, publish,      *
9  * distribute, distribute with modifications, sublicense, and/or sell       *
10  * copies of the Software, and to permit persons to whom the Software is    *
11  * furnished to do so, subject to the following conditions:                 *
12  *                                                                          *
13  * The above copyright notice and this permission notice shall be included  *
14  * in all copies or substantial portions of the Software.                   *
15  *                                                                          *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS  *
17  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF               *
18  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.   *
19  * IN NO EVENT SHALL THE ABOVE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,   *
20  * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR    *
21  * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR    *
22  * THE USE OR OTHER DEALINGS IN THE SOFTWARE.                               *
23  *                                                                          *
24  * Except as contained in this notice, the name(s) of the above copyright   *
25  * holders shall not be used in advertising or otherwise to promote the     *
26  * sale, use or other dealings in this Software without prior written       *
27  * authorization.                                                           *
28  ****************************************************************************/
29 
30 /***************************************************************************
31 *                                                                          *
32 *  Author : Juergen Pfeifer                                                *
33 *                                                                          *
34 ***************************************************************************/
35 
36 #include "form.priv.h"
37 
38 MODULE_ID("$Id: fty_regex.c,v 1.33 2021/08/14 15:01:52 tom Exp $")
39 
40 #if HAVE_REGEX_H_FUNCS || HAVE_LIB_PCRE2	/* We prefer POSIX regex */
41 
42 #if HAVE_PCRE2POSIX_H
43 #include <pcre2posix.h>
44 
45 /* pcre2 used to provide its "POSIX" entrypoints using the same names as the
46  * standard ones in the C runtime, but that never worked because the linker
47  * would use the C runtime.  Debian patched the library to fix this symbol
48  * conflict, but overlooked the header file, and Debian's patch was made
49  * obsolete when pcre2 was changed early in 2019 to provide different names.
50  *
51  * Here is a workaround to make the older version of Debian's package work.
52  */
53 #if !defined(PCRE2regcomp) && defined(HAVE_PCRE2REGCOMP)
54 
55 #undef regcomp
56 #undef regexec
57 #undef regfree
58 
59 #ifdef __cplusplus
60 extern "C"
61 {
62 #endif
63   PCRE2POSIX_EXP_DECL int PCRE2regcomp(regex_t *, const char *, int);
64   PCRE2POSIX_EXP_DECL int PCRE2regexec(const regex_t *, const char *, size_t,
65 				       regmatch_t *, int);
66   PCRE2POSIX_EXP_DECL void PCRE2regfree(regex_t *);
67 #ifdef __cplusplus
68 }				/* extern "C" */
69 #endif
70 #define regcomp(r,s,n)          PCRE2regcomp(r,s,n)
71 #define regexec(r,s,n,m,x)      PCRE2regexec(r,s,n,m,x)
72 #define regfree(r)              PCRE2regfree(r)
73 #endif
74 /* end workaround... */
75 #elif HAVE_PCREPOSIX_H
76 #include <pcreposix.h>
77 #else
78 #include <regex.h>
79 #endif
80 
81 typedef struct
82   {
83     regex_t *pRegExp;
84     unsigned long *refCount;
85   }
86 RegExp_Arg;
87 
88 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
89 #undef RETURN
90 static int reg_errno;
91 
92 static char *
93 RegEx_Init(char *instring)
94 {
95   reg_errno = 0;
96   return instring;
97 }
98 
99 static char *
100 RegEx_Error(int code)
101 {
102   reg_errno = code;
103   return 0;
104 }
105 
106 #define INIT 		register char *sp = RegEx_Init(instring);
107 #define GETC()		(*sp++)
108 #define PEEKC()		(*sp)
109 #define UNGETC(c)	(--sp)
110 #define RETURN(c)	return(c)
111 #define ERROR(c)	return RegEx_Error(c)
112 
113 #if HAVE_REGEXP_H_FUNCS
114 #include <regexp.h>
115 #else
116 #include <regexpr.h>
117 #endif
118 
119 typedef struct
120 {
121   char *compiled_expression;
122   unsigned long *refCount;
123 }
124 RegExp_Arg;
125 
126 /* Maximum Length we allow for a compiled regular expression */
127 #define MAX_RX_LEN   (2048)
128 #define RX_INCREMENT (256)
129 
130 #endif
131 
132 #if HAVE_REGEX_H_FUNCS | HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
133 # define MAYBE_UNUSED
134 #else
135 # define MAYBE_UNUSED GCC_UNUSED
136 #endif
137 
138 /*---------------------------------------------------------------------------
139 |   Facility      :  libnform
140 |   Function      :  static void *Generic_RegularExpression_Type(void * arg)
141 |
142 |   Description   :  Allocate structure for regex type argument.
143 |
144 |   Return Values :  Pointer to argument structure or NULL on error
145 +--------------------------------------------------------------------------*/
146 static void *
Generic_RegularExpression_Type(void * arg MAYBE_UNUSED)147 Generic_RegularExpression_Type(void *arg MAYBE_UNUSED)
148 {
149 #if HAVE_REGEX_H_FUNCS
150   char *rx = (char *)arg;
151   RegExp_Arg *preg = (RegExp_Arg *)0;
152 
153   if (rx)
154     {
155       preg = typeCalloc(RegExp_Arg, 1);
156 
157       if (preg)
158 	{
159 	  T((T_CREATE("RegExp_Arg %p"), (void *)preg));
160 	  if (((preg->pRegExp = typeMalloc(regex_t, 1)) != 0)
161 	      && !regcomp(preg->pRegExp, rx,
162 			  (REG_EXTENDED | REG_NOSUB | REG_NEWLINE)))
163 	    {
164 	      T((T_CREATE("regex_t %p"), (void *)preg->pRegExp));
165 	      if ((preg->refCount = typeMalloc(unsigned long, 1)) != 0)
166 		 *(preg->refCount) = 1;
167 	    }
168 	  else
169 	    {
170 	      if (preg->pRegExp)
171 		free(preg->pRegExp);
172 	      free(preg);
173 	      preg = (RegExp_Arg *)0;
174 	    }
175 	}
176     }
177   return ((void *)preg);
178 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
179   char *rx = (char *)arg;
180   RegExp_Arg *pArg = (RegExp_Arg *)0;
181 
182   if (rx)
183     {
184       pArg = typeMalloc(RegExp_Arg, 1);
185 
186       if (pArg)
187 	{
188 	  int blen = RX_INCREMENT;
189 
190 	  T((T_CREATE("RegExp_Arg %p"), pArg));
191 	  pArg->compiled_expression = NULL;
192 	  if ((pArg->refCount = typeMalloc(unsigned long, 1)) != 0)
193 	     *(pArg->refCount) = 1;
194 
195 	  do
196 	    {
197 	      char *buf = typeMalloc(char, blen);
198 
199 	      if (buf)
200 		{
201 #if HAVE_REGEXP_H_FUNCS
202 		  char *last_pos = compile(rx, buf, &buf[blen], '\0');
203 
204 #else /* HAVE_REGEXPR_H_FUNCS */
205 		  char *last_pos = compile(rx, buf, &buf[blen]);
206 #endif
207 		  if (reg_errno)
208 		    {
209 		      free(buf);
210 		      if (reg_errno == 50)
211 			blen += RX_INCREMENT;
212 		      else
213 			{
214 			  free(pArg);
215 			  pArg = NULL;
216 			  break;
217 			}
218 		    }
219 		  else
220 		    {
221 		      pArg->compiled_expression = buf;
222 		      break;
223 		    }
224 		}
225 	    }
226 	  while (blen <= MAX_RX_LEN);
227 	}
228       if (pArg && !pArg->compiled_expression)
229 	{
230 	  free(pArg);
231 	  pArg = NULL;
232 	}
233     }
234   return (void *)pArg;
235 #else
236   return 0;
237 #endif
238 }
239 
240 /*---------------------------------------------------------------------------
241 |   Facility      :  libnform
242 |   Function      :  static void *Make_RegularExpression_Type(va_list * ap)
243 |
244 |   Description   :  Allocate structure for regex type argument.
245 |
246 |   Return Values :  Pointer to argument structure or NULL on error
247 +--------------------------------------------------------------------------*/
248 static void *
Make_RegularExpression_Type(va_list * ap)249 Make_RegularExpression_Type(va_list *ap)
250 {
251   char *rx = va_arg(*ap, char *);
252 
253   return Generic_RegularExpression_Type((void *)rx);
254 }
255 
256 /*---------------------------------------------------------------------------
257 |   Facility      :  libnform
258 |   Function      :  static void *Copy_RegularExpression_Type(
259 |                                      const void * argp)
260 |
261 |   Description   :  Copy structure for regex type argument.
262 |
263 |   Return Values :  Pointer to argument structure or NULL on error.
264 +--------------------------------------------------------------------------*/
265 static void *
Copy_RegularExpression_Type(const void * argp MAYBE_UNUSED)266 Copy_RegularExpression_Type(const void *argp MAYBE_UNUSED)
267 {
268 #if (HAVE_REGEX_H_FUNCS | HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS)
269   const RegExp_Arg *ap = (const RegExp_Arg *)argp;
270   const RegExp_Arg *result = (const RegExp_Arg *)0;
271 
272   if (ap)
273     {
274       *(ap->refCount) += 1;
275       result = ap;
276     }
277   return (void *)result;
278 #else
279   return 0;
280 #endif
281 }
282 
283 /*---------------------------------------------------------------------------
284 |   Facility      :  libnform
285 |   Function      :  static void Free_RegularExpression_Type(void * argp)
286 |
287 |   Description   :  Free structure for regex type argument.
288 |
289 |   Return Values :  -
290 +--------------------------------------------------------------------------*/
291 static void
Free_RegularExpression_Type(void * argp MAYBE_UNUSED)292 Free_RegularExpression_Type(void *argp MAYBE_UNUSED)
293 {
294 #if HAVE_REGEX_H_FUNCS | HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
295   RegExp_Arg *ap = (RegExp_Arg *)argp;
296 
297   if (ap)
298     {
299       if (--(*(ap->refCount)) == 0)
300 	{
301 #if HAVE_REGEX_H_FUNCS
302 	  if (ap->pRegExp)
303 	    {
304 	      free(ap->refCount);
305 	      regfree(ap->pRegExp);
306 	      free(ap->pRegExp);
307 	    }
308 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
309 	  if (ap->compiled_expression)
310 	    {
311 	      free(ap->refCount);
312 	      free(ap->compiled_expression);
313 	    }
314 #endif
315 	  free(ap);
316 	}
317     }
318 #endif
319 }
320 
321 /*---------------------------------------------------------------------------
322 |   Facility      :  libnform
323 |   Function      :  static bool Check_RegularExpression_Field(
324 |                                      FIELD * field,
325 |                                      const void  * argp)
326 |
327 |   Description   :  Validate buffer content to be a valid regular expression
328 |
329 |   Return Values :  TRUE  - field is valid
330 |                    FALSE - field is invalid
331 +--------------------------------------------------------------------------*/
332 static bool
Check_RegularExpression_Field(FIELD * field MAYBE_UNUSED,const void * argp MAYBE_UNUSED)333 Check_RegularExpression_Field(FIELD *field MAYBE_UNUSED,
334 			      const void *argp MAYBE_UNUSED)
335 {
336   bool match = FALSE;
337 
338 #if HAVE_REGEX_H_FUNCS
339   const RegExp_Arg *ap = (const RegExp_Arg *)argp;
340 
341   if (ap && ap->pRegExp)
342     match = (regexec(ap->pRegExp, field_buffer(field, 0), 0, NULL, 0)
343 	     ? FALSE
344 	     : TRUE);
345 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
346   RegExp_Arg *ap = (RegExp_Arg *)argp;
347 
348   if (ap && ap->compiled_expression)
349     match = (step(field_buffer(field, 0), ap->compiled_expression)
350 	     ? TRUE
351 	     : FALSE);
352 #endif
353   return match;
354 }
355 
356 static FIELDTYPE typeREGEXP =
357 {
358   _HAS_ARGS | _RESIDENT,
359   1,				/* this is mutable, so we can't be const */
360   (FIELDTYPE *)0,
361   (FIELDTYPE *)0,
362   Make_RegularExpression_Type,
363   Copy_RegularExpression_Type,
364   Free_RegularExpression_Type,
365   INIT_FT_FUNC(Check_RegularExpression_Field),
366   INIT_FT_FUNC(NULL),
367   INIT_FT_FUNC(NULL),
368   INIT_FT_FUNC(NULL),
369 #if NCURSES_INTEROP_FUNCS
370   Generic_RegularExpression_Type
371 #endif
372 };
373 
374 FORM_EXPORT_VAR(FIELDTYPE *) TYPE_REGEXP = &typeREGEXP;
375 
376 #if NCURSES_INTEROP_FUNCS
377 /* The next routines are to simplify the use of ncurses from
378    programming languages with restrictions on interop with C level
379    constructs (e.g. variable access or va_list + ellipsis constructs)
380 */
381 FORM_EXPORT(FIELDTYPE *)
_nc_TYPE_REGEXP(void)382 _nc_TYPE_REGEXP(void)
383 {
384   return TYPE_REGEXP;
385 }
386 #endif
387 
388 /* fty_regex.c ends here */
389