xref: /openbsd/lib/libform/fty_regex.c (revision c7ef0cfc)
1 /*	$OpenBSD: fty_regex.c,v 1.10 2023/10/17 09:52:10 nicm Exp $	*/
2 /****************************************************************************
3  * Copyright 2018-2020,2021 Thomas E. Dickey                                *
4  * Copyright 1998-2012,2015 Free Software Foundation, Inc.                  *
5  *                                                                          *
6  * Permission is hereby granted, free of charge, to any person obtaining a  *
7  * copy of this software and associated documentation files (the            *
8  * "Software"), to deal in the Software without restriction, including      *
9  * without limitation the rights to use, copy, modify, merge, publish,      *
10  * distribute, distribute with modifications, sublicense, and/or sell       *
11  * copies of the Software, and to permit persons to whom the Software is    *
12  * furnished to do so, subject to the following conditions:                 *
13  *                                                                          *
14  * The above copyright notice and this permission notice shall be included  *
15  * in all copies or substantial portions of the Software.                   *
16  *                                                                          *
17  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS  *
18  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF               *
19  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.   *
20  * IN NO EVENT SHALL THE ABOVE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,   *
21  * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR    *
22  * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR    *
23  * THE USE OR OTHER DEALINGS IN THE SOFTWARE.                               *
24  *                                                                          *
25  * Except as contained in this notice, the name(s) of the above copyright   *
26  * holders shall not be used in advertising or otherwise to promote the     *
27  * sale, use or other dealings in this Software without prior written       *
28  * authorization.                                                           *
29  ****************************************************************************/
30 
31 /***************************************************************************
32 *                                                                          *
33 *  Author : Juergen Pfeifer                                                *
34 *                                                                          *
35 ***************************************************************************/
36 
37 #include "form.priv.h"
38 
39 MODULE_ID("$Id: fty_regex.c,v 1.10 2023/10/17 09:52:10 nicm Exp $")
40 
41 #if HAVE_REGEX_H_FUNCS || HAVE_LIB_PCRE2	/* We prefer POSIX regex */
42 
43 #if HAVE_PCRE2POSIX_H
44 #include <pcre2posix.h>
45 
46 /* pcre2 used to provide its "POSIX" entrypoints using the same names as the
47  * standard ones in the C runtime, but that never worked because the linker
48  * would use the C runtime.  Debian patched the library to fix this symbol
49  * conflict, but overlooked the header file, and Debian's patch was made
50  * obsolete when pcre2 was changed early in 2019 to provide different names.
51  *
52  * Here is a workaround to make the older version of Debian's package work.
53  */
54 #if !defined(PCRE2regcomp) && defined(HAVE_PCRE2REGCOMP)
55 
56 #undef regcomp
57 #undef regexec
58 #undef regfree
59 
60 #ifdef __cplusplus
61 extern "C"
62 {
63 #endif
64   PCRE2POSIX_EXP_DECL int PCRE2regcomp(regex_t *, const char *, int);
65   PCRE2POSIX_EXP_DECL int PCRE2regexec(const regex_t *, const char *, size_t,
66 				       regmatch_t *, int);
67   PCRE2POSIX_EXP_DECL void PCRE2regfree(regex_t *);
68 #ifdef __cplusplus
69 }				/* extern "C" */
70 #endif
71 #define regcomp(r,s,n)          PCRE2regcomp(r,s,n)
72 #define regexec(r,s,n,m,x)      PCRE2regexec(r,s,n,m,x)
73 #define regfree(r)              PCRE2regfree(r)
74 #endif
75 /* end workaround... */
76 #elif HAVE_PCREPOSIX_H
77 #include <pcreposix.h>
78 #else
79 #include <regex.h>
80 #endif
81 
82 typedef struct
83   {
84     regex_t *pRegExp;
85     unsigned long *refCount;
86   }
87 RegExp_Arg;
88 
89 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
90 #undef RETURN
91 static int reg_errno;
92 
93 static char *
94 RegEx_Init(char *instring)
95 {
96   reg_errno = 0;
97   return instring;
98 }
99 
100 static char *
101 RegEx_Error(int code)
102 {
103   reg_errno = code;
104   return 0;
105 }
106 
107 #define INIT 		register char *sp = RegEx_Init(instring);
108 #define GETC()		(*sp++)
109 #define PEEKC()		(*sp)
110 #define UNGETC(c)	(--sp)
111 #define RETURN(c)	return(c)
112 #define ERROR(c)	return RegEx_Error(c)
113 
114 #if HAVE_REGEXP_H_FUNCS
115 #include <regexp.h>
116 #else
117 #include <regexpr.h>
118 #endif
119 
120 typedef struct
121 {
122   char *compiled_expression;
123   unsigned long *refCount;
124 }
125 RegExp_Arg;
126 
127 /* Maximum Length we allow for a compiled regular expression */
128 #define MAX_RX_LEN   (2048)
129 #define RX_INCREMENT (256)
130 
131 #endif
132 
133 #if HAVE_REGEX_H_FUNCS | HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
134 # define MAYBE_UNUSED
135 #else
136 # define MAYBE_UNUSED GCC_UNUSED
137 #endif
138 
139 /*---------------------------------------------------------------------------
140 |   Facility      :  libnform
141 |   Function      :  static void *Generic_RegularExpression_Type(void * arg)
142 |
143 |   Description   :  Allocate structure for regex type argument.
144 |
145 |   Return Values :  Pointer to argument structure or NULL on error
146 +--------------------------------------------------------------------------*/
147 static void *
Generic_RegularExpression_Type(void * arg MAYBE_UNUSED)148 Generic_RegularExpression_Type(void *arg MAYBE_UNUSED)
149 {
150 #if HAVE_REGEX_H_FUNCS
151   char *rx = (char *)arg;
152   RegExp_Arg *preg = (RegExp_Arg *)0;
153 
154   if (rx)
155     {
156       preg = typeCalloc(RegExp_Arg, 1);
157 
158       if (preg)
159 	{
160 	  T((T_CREATE("RegExp_Arg %p"), (void *)preg));
161 	  if (((preg->pRegExp = typeMalloc(regex_t, 1)) != 0)
162 	      && !regcomp(preg->pRegExp, rx,
163 			  (REG_EXTENDED | REG_NOSUB | REG_NEWLINE)))
164 	    {
165 	      T((T_CREATE("regex_t %p"), (void *)preg->pRegExp));
166 	      if ((preg->refCount = typeMalloc(unsigned long, 1)) != 0)
167 		 *(preg->refCount) = 1;
168 	    }
169 	  else
170 	    {
171 	      if (preg->pRegExp)
172 		free(preg->pRegExp);
173 	      free(preg);
174 	      preg = (RegExp_Arg *)0;
175 	    }
176 	}
177     }
178   return ((void *)preg);
179 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
180   char *rx = (char *)arg;
181   RegExp_Arg *pArg = (RegExp_Arg *)0;
182 
183   if (rx)
184     {
185       pArg = typeMalloc(RegExp_Arg, 1);
186 
187       if (pArg)
188 	{
189 	  int blen = RX_INCREMENT;
190 
191 	  T((T_CREATE("RegExp_Arg %p"), pArg));
192 	  pArg->compiled_expression = NULL;
193 	  if ((pArg->refCount = typeMalloc(unsigned long, 1)) != 0)
194 	     *(pArg->refCount) = 1;
195 
196 	  do
197 	    {
198 	      char *buf = typeMalloc(char, blen);
199 
200 	      if (buf)
201 		{
202 #if HAVE_REGEXP_H_FUNCS
203 		  char *last_pos = compile(rx, buf, &buf[blen], '\0');
204 
205 #else /* HAVE_REGEXPR_H_FUNCS */
206 		  char *last_pos = compile(rx, buf, &buf[blen]);
207 #endif
208 		  if (reg_errno)
209 		    {
210 		      free(buf);
211 		      if (reg_errno == 50)
212 			blen += RX_INCREMENT;
213 		      else
214 			{
215 			  free(pArg);
216 			  pArg = NULL;
217 			  break;
218 			}
219 		    }
220 		  else
221 		    {
222 		      pArg->compiled_expression = buf;
223 		      break;
224 		    }
225 		}
226 	    }
227 	  while (blen <= MAX_RX_LEN);
228 	}
229       if (pArg && !pArg->compiled_expression)
230 	{
231 	  free(pArg);
232 	  pArg = NULL;
233 	}
234     }
235   return (void *)pArg;
236 #else
237   return 0;
238 #endif
239 }
240 
241 /*---------------------------------------------------------------------------
242 |   Facility      :  libnform
243 |   Function      :  static void *Make_RegularExpression_Type(va_list * ap)
244 |
245 |   Description   :  Allocate structure for regex type argument.
246 |
247 |   Return Values :  Pointer to argument structure or NULL on error
248 +--------------------------------------------------------------------------*/
249 static void *
Make_RegularExpression_Type(va_list * ap)250 Make_RegularExpression_Type(va_list *ap)
251 {
252   char *rx = va_arg(*ap, char *);
253 
254   return Generic_RegularExpression_Type((void *)rx);
255 }
256 
257 /*---------------------------------------------------------------------------
258 |   Facility      :  libnform
259 |   Function      :  static void *Copy_RegularExpression_Type(
260 |                                      const void * argp)
261 |
262 |   Description   :  Copy structure for regex type argument.
263 |
264 |   Return Values :  Pointer to argument structure or NULL on error.
265 +--------------------------------------------------------------------------*/
266 static void *
Copy_RegularExpression_Type(const void * argp MAYBE_UNUSED)267 Copy_RegularExpression_Type(const void *argp MAYBE_UNUSED)
268 {
269 #if (HAVE_REGEX_H_FUNCS | HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS)
270   const RegExp_Arg *ap = (const RegExp_Arg *)argp;
271   const RegExp_Arg *result = (const RegExp_Arg *)0;
272 
273   if (ap)
274     {
275       *(ap->refCount) += 1;
276       result = ap;
277     }
278   return (void *)result;
279 #else
280   return 0;
281 #endif
282 }
283 
284 /*---------------------------------------------------------------------------
285 |   Facility      :  libnform
286 |   Function      :  static void Free_RegularExpression_Type(void * argp)
287 |
288 |   Description   :  Free structure for regex type argument.
289 |
290 |   Return Values :  -
291 +--------------------------------------------------------------------------*/
292 static void
Free_RegularExpression_Type(void * argp MAYBE_UNUSED)293 Free_RegularExpression_Type(void *argp MAYBE_UNUSED)
294 {
295 #if HAVE_REGEX_H_FUNCS | HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
296   RegExp_Arg *ap = (RegExp_Arg *)argp;
297 
298   if (ap)
299     {
300       if (--(*(ap->refCount)) == 0)
301 	{
302 #if HAVE_REGEX_H_FUNCS
303 	  if (ap->pRegExp)
304 	    {
305 	      free(ap->refCount);
306 	      regfree(ap->pRegExp);
307 	      free(ap->pRegExp);
308 	    }
309 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
310 	  if (ap->compiled_expression)
311 	    {
312 	      free(ap->refCount);
313 	      free(ap->compiled_expression);
314 	    }
315 #endif
316 	  free(ap);
317 	}
318     }
319 #endif
320 }
321 
322 /*---------------------------------------------------------------------------
323 |   Facility      :  libnform
324 |   Function      :  static bool Check_RegularExpression_Field(
325 |                                      FIELD * field,
326 |                                      const void  * argp)
327 |
328 |   Description   :  Validate buffer content to be a valid regular expression
329 |
330 |   Return Values :  TRUE  - field is valid
331 |                    FALSE - field is invalid
332 +--------------------------------------------------------------------------*/
333 static bool
Check_RegularExpression_Field(FIELD * field MAYBE_UNUSED,const void * argp MAYBE_UNUSED)334 Check_RegularExpression_Field(FIELD *field MAYBE_UNUSED,
335 			      const void *argp MAYBE_UNUSED)
336 {
337   bool match = FALSE;
338 
339 #if HAVE_REGEX_H_FUNCS
340   const RegExp_Arg *ap = (const RegExp_Arg *)argp;
341 
342   if (ap && ap->pRegExp)
343     match = (regexec(ap->pRegExp, field_buffer(field, 0), 0, NULL, 0)
344 	     ? FALSE
345 	     : TRUE);
346 #elif HAVE_REGEXP_H_FUNCS | HAVE_REGEXPR_H_FUNCS
347   RegExp_Arg *ap = (RegExp_Arg *)argp;
348 
349   if (ap && ap->compiled_expression)
350     match = (step(field_buffer(field, 0), ap->compiled_expression)
351 	     ? TRUE
352 	     : FALSE);
353 #endif
354   return match;
355 }
356 
357 static FIELDTYPE typeREGEXP =
358 {
359   _HAS_ARGS | _RESIDENT,
360   1,				/* this is mutable, so we can't be const */
361   (FIELDTYPE *)0,
362   (FIELDTYPE *)0,
363   Make_RegularExpression_Type,
364   Copy_RegularExpression_Type,
365   Free_RegularExpression_Type,
366   INIT_FT_FUNC(Check_RegularExpression_Field),
367   INIT_FT_FUNC(NULL),
368   INIT_FT_FUNC(NULL),
369   INIT_FT_FUNC(NULL),
370 #if NCURSES_INTEROP_FUNCS
371   Generic_RegularExpression_Type
372 #endif
373 };
374 
375 FORM_EXPORT_VAR(FIELDTYPE *) TYPE_REGEXP = &typeREGEXP;
376 
377 #if NCURSES_INTEROP_FUNCS
378 /* The next routines are to simplify the use of ncurses from
379    programming languages with restrictions on interop with C level
380    constructs (e.g. variable access or va_list + ellipsis constructs)
381 */
382 FORM_EXPORT(FIELDTYPE *)
_nc_TYPE_REGEXP(void)383 _nc_TYPE_REGEXP(void)
384 {
385   return TYPE_REGEXP;
386 }
387 #endif
388 
389 /* fty_regex.c ends here */
390