3 * Copyright (C) 2012 by Larry Wall and others
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
9 #ifndef PERL_INVLIST_INLINE_H_
10 #define PERL_INVLIST_INLINE_H_
12 #if defined(PERL_IN_UTF8_C) \
13 || defined(PERL_IN_REGCOMP_C) \
14 || defined(PERL_IN_REGEXEC_C) \
15 || defined(PERL_IN_TOKE_C) \
16 || defined(PERL_IN_PP_C) \
17 || defined(PERL_IN_OP_C) \
18 || defined(PERL_IN_DOOP_C)
20 /* An element is in an inversion list iff its index is even numbered: 0, 2, 4,
22 #define ELEMENT_RANGE_MATCHES_INVLIST(i) (! ((i) & 1))
23 #define PREV_RANGE_MATCHES_INVLIST(i) (! ELEMENT_RANGE_MATCHES_INVLIST(i))
25 /* This converts to/from our UVs to what the SV code is expecting: bytes. */
26 #define TO_INTERNAL_SIZE(x) ((x) * sizeof(UV))
27 #define FROM_INTERNAL_SIZE(x) ((x)/ sizeof(UV))
29 PERL_STATIC_INLINE bool
30 S_is_invlist(const SV* const invlist)
32 return invlist != NULL && SvTYPE(invlist) == SVt_INVLIST;
35 PERL_STATIC_INLINE bool*
36 S_get_invlist_offset_addr(SV* invlist)
38 /* Return the address of the field that says whether the inversion list is
39 * offset (it contains 1) or not (contains 0) */
40 PERL_ARGS_ASSERT_GET_INVLIST_OFFSET_ADDR;
42 assert(is_invlist(invlist));
44 return &(((XINVLIST*) SvANY(invlist))->is_offset);
48 S__invlist_len(SV* const invlist)
50 /* Returns the current number of elements stored in the inversion list's
53 PERL_ARGS_ASSERT__INVLIST_LEN;
55 assert(is_invlist(invlist));
57 return (SvCUR(invlist) == 0)
59 : FROM_INTERNAL_SIZE(SvCUR(invlist)) - *get_invlist_offset_addr(invlist);
62 PERL_STATIC_INLINE bool
63 S__invlist_contains_cp(SV* const invlist, const UV cp)
65 /* Does <invlist> contain code point <cp> as part of the set? */
67 IV index = _invlist_search(invlist, cp);
69 PERL_ARGS_ASSERT__INVLIST_CONTAINS_CP;
71 return index >= 0 && ELEMENT_RANGE_MATCHES_INVLIST(index);
74 PERL_STATIC_INLINE UV*
75 S_invlist_array(SV* const invlist)
77 /* Returns the pointer to the inversion list's array. Every time the
78 * length changes, this needs to be called in case malloc or realloc moved
81 PERL_ARGS_ASSERT_INVLIST_ARRAY;
83 /* Must not be empty. If these fail, you probably didn't check for <len>
84 * being non-zero before trying to get the array */
85 assert(_invlist_len(invlist));
87 /* The very first element always contains zero, The array begins either
88 * there, or if the inversion list is offset, at the element after it.
89 * The offset header field determines which; it contains 0 or 1 to indicate
90 * how much additionally to add */
91 assert(0 == *(SvPVX(invlist)));
92 return ((UV *) SvPVX(invlist) + *get_invlist_offset_addr(invlist));
96 #if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_OP_C) || defined(PERL_IN_DOOP_C)
98 PERL_STATIC_INLINE void
99 S_invlist_extend(pTHX_ SV* const invlist, const UV new_max)
101 /* Grow the maximum size of an inversion list */
103 PERL_ARGS_ASSERT_INVLIST_EXTEND;
105 assert(SvTYPE(invlist) == SVt_INVLIST);
107 /* Add one to account for the zero element at the beginning which may not
108 * be counted by the calling parameters */
109 SvGROW((SV *)invlist, TO_INTERNAL_SIZE(new_max + 1));
112 PERL_STATIC_INLINE void
113 S_invlist_set_len(pTHX_ SV* const invlist, const UV len, const bool offset)
115 /* Sets the current number of elements stored in the inversion list.
116 * Updates SvCUR correspondingly */
118 PERL_ARGS_ASSERT_INVLIST_SET_LEN;
120 assert(SvTYPE(invlist) == SVt_INVLIST);
125 : TO_INTERNAL_SIZE(len + offset));
126 assert(SvLEN(invlist) == 0 || SvCUR(invlist) <= SvLEN(invlist));
129 PERL_STATIC_INLINE SV*
130 S_add_cp_to_invlist(pTHX_ SV* invlist, const UV cp) {
131 return _add_range_to_invlist(invlist, cp, cp);
134 PERL_STATIC_INLINE UV
135 S_invlist_highest(SV* const invlist)
137 /* Returns the highest code point that matches an inversion list. This API
138 * has an ambiguity, as it returns 0 under either the highest is actually
139 * 0, or if the list is empty. If this distinction matters to you, check
140 * for emptiness before calling this function */
142 UV len = _invlist_len(invlist);
145 PERL_ARGS_ASSERT_INVLIST_HIGHEST;
151 array = invlist_array(invlist);
153 /* The last element in the array in the inversion list always starts a
154 * range that goes to infinity. That range may be for code points that are
155 * matched in the inversion list, or it may be for ones that aren't
156 * matched. In the latter case, the highest code point in the set is one
157 * less than the beginning of this range; otherwise it is the final element
158 * of this range: infinity */
159 return (ELEMENT_RANGE_MATCHES_INVLIST(len - 1))
161 : array[len - 1] - 1;
164 # if defined(PERL_IN_REGCOMP_C)
166 PERL_STATIC_INLINE UV
167 S_invlist_highest_range_start(SV* const invlist)
169 /* Returns the lowest code point of the highest range in the inversion
170 * list parameter. This API has an ambiguity: it returns 0 either when
171 * the lowest such point is actually 0 or when the list is empty. If this
172 * distinction matters to you, check for emptiness before calling this
175 UV len = _invlist_len(invlist);
178 PERL_ARGS_ASSERT_INVLIST_HIGHEST_RANGE_START;
184 array = invlist_array(invlist);
186 /* The last element in the array in the inversion list always starts a
187 * range that goes to infinity. That range may be for code points that are
188 * matched in the inversion list, or it may be for ones that aren't
189 * matched. In the first case, the lowest code point in the matching range
190 * is that the one that started the range. If the other case, the final
191 * matching range begins at the next element down (which may be 0 in the
193 return (ELEMENT_RANGE_MATCHES_INVLIST(len - 1))
202 #if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_OP_C)
204 PERL_STATIC_INLINE STRLEN*
205 S_get_invlist_iter_addr(SV* invlist)
207 /* Return the address of the UV that contains the current iteration
210 PERL_ARGS_ASSERT_GET_INVLIST_ITER_ADDR;
212 assert(is_invlist(invlist));
214 return &(((XINVLIST*) SvANY(invlist))->iterator);
217 PERL_STATIC_INLINE void
218 S_invlist_iterinit(SV* invlist) /* Initialize iterator for invlist */
220 PERL_ARGS_ASSERT_INVLIST_ITERINIT;
222 *get_invlist_iter_addr(invlist) = 0;
225 PERL_STATIC_INLINE void
226 S_invlist_iterfinish(SV* invlist)
228 /* Terminate iterator for invlist. This is to catch development errors.
229 * Any iteration that is interrupted before completed should call this
230 * function. Functions that add code points anywhere else but to the end
231 * of an inversion list assert that they are not in the middle of an
232 * iteration. If they were, the addition would make the iteration
233 * problematical: if the iteration hadn't reached the place where things
234 * were being added, it would be ok */
236 PERL_ARGS_ASSERT_INVLIST_ITERFINISH;
238 *get_invlist_iter_addr(invlist) = (STRLEN) UV_MAX;
242 S_invlist_iternext(SV* invlist, UV* start, UV* end)
244 /* An C<invlist_iterinit> call on <invlist> must be used to set this up.
245 * This call sets in <*start> and <*end>, the next range in <invlist>.
246 * Returns <TRUE> if successful and the next call will return the next
247 * range; <FALSE> if was already at the end of the list. If the latter,
248 * <*start> and <*end> are unchanged, and the next call to this function
249 * will start over at the beginning of the list */
251 STRLEN* pos = get_invlist_iter_addr(invlist);
252 UV len = _invlist_len(invlist);
255 PERL_ARGS_ASSERT_INVLIST_ITERNEXT;
258 *pos = (STRLEN) UV_MAX; /* Force iterinit() to be required next time */
262 array = invlist_array(invlist);
264 *start = array[(*pos)++];
270 *end = array[(*pos)++] - 1;
278 #ifndef PERL_IN_REGCOMP_C
280 /* These symbols are only needed later in regcomp.c */
281 # undef TO_INTERNAL_SIZE
282 # undef FROM_INTERNAL_SIZE
285 #endif /* PERL_INVLIST_INLINE_H_ */