Commit | Line | Data |
---|---|---|
04e98a4d AD |
1 | /* dquote_static.c |
2 | * | |
3efe3cb8 | 3 | * This file contains static functions that are related to |
04e98a4d AD |
4 | * parsing double-quotish expressions, but are used in more than |
5 | * one file. | |
6 | * | |
7 | * It is currently #included by regcomp.c and toke.c. | |
8 | */ | |
9 | ||
881ffab6 | 10 | #define PERL_IN_DQUOTE_STATIC_C |
881ffab6 KW |
11 | #include "embed.h" |
12 | ||
04e98a4d AD |
13 | /* |
14 | - regcurly - a little FSA that accepts {\d+,?\d*} | |
15 | Pulled from regcomp.c. | |
16 | */ | |
04e98a4d | 17 | PERL_STATIC_INLINE I32 |
ddeaf645 | 18 | S_regcurly(const char *s) |
04e98a4d | 19 | { |
881ffab6 | 20 | PERL_ARGS_ASSERT_REGCURLY; |
04e98a4d AD |
21 | |
22 | if (*s++ != '{') | |
23 | return FALSE; | |
24 | if (!isDIGIT(*s)) | |
25 | return FALSE; | |
26 | while (isDIGIT(*s)) | |
27 | s++; | |
2ec31dd9 | 28 | if (*s == ',') { |
04e98a4d | 29 | s++; |
2ec31dd9 KW |
30 | while (isDIGIT(*s)) |
31 | s++; | |
32 | } | |
4d68ffa0 | 33 | |
412f55bb | 34 | return *s == '}'; |
04e98a4d | 35 | } |
db30362b | 36 | |
68b355dd KW |
37 | /* XXX Add documentation after final interface and behavior is decided */ |
38 | /* May want to show context for error, so would pass Perl_bslash_c(pTHX_ const char* current, const char* start, const bool output_warning) | |
39 | U8 source = *current; | |
68b355dd KW |
40 | */ |
41 | ||
42 | STATIC char | |
421e43ba | 43 | S_grok_bslash_c(pTHX_ const char source, const bool output_warning) |
68b355dd KW |
44 | { |
45 | ||
46 | U8 result; | |
47 | ||
421e43ba | 48 | if (! isPRINT_A(source)) { |
7357bd17 KW |
49 | Perl_croak(aTHX_ "%s", |
50 | "Character following \"\\c\" must be printable ASCII"); | |
68b355dd | 51 | } |
421e43ba | 52 | else if (source == '{') { |
a27ed980 KW |
53 | const char control = toCTRL('{'); |
54 | if (isPRINT_A(control)) { | |
55 | /* diag_listed_as: Use "%s" instead of "%s" */ | |
56 | Perl_croak(aTHX_ "Use \"%c\" instead of \"\\c{\"", control); | |
57 | } | |
58 | else { | |
59 | Perl_croak(aTHX_ "Sequence \"\\c{\" invalid"); | |
60 | } | |
421e43ba | 61 | } |
68b355dd KW |
62 | |
63 | result = toCTRL(source); | |
5e784d58 | 64 | if (output_warning && isPRINT_A(result)) { |
4d8be631 KW |
65 | U8 clearer[3]; |
66 | U8 i = 0; | |
67 | if (! isWORDCHAR(result)) { | |
68 | clearer[i++] = '\\'; | |
69 | } | |
70 | clearer[i++] = result; | |
71 | clearer[i++] = '\0'; | |
68b355dd | 72 | |
4d8be631 KW |
73 | Perl_ck_warner(aTHX_ packWARN(WARN_SYNTAX), |
74 | "\"\\c%c\" is more clearly written simply as \"%s\"", | |
75 | source, | |
76 | clearer); | |
68b355dd KW |
77 | } |
78 | ||
79 | return result; | |
80 | } | |
81 | ||
db30362b | 82 | STATIC bool |
00ce5563 | 83 | S_grok_bslash_o(pTHX_ char **s, UV *uv, const char** error_msg, |
80f4111b | 84 | const bool output_warning, const bool strict, |
17896a85 | 85 | const bool silence_non_portable, |
80f4111b | 86 | const bool UTF) |
db30362b KW |
87 | { |
88 | ||
89 | /* Documentation to be supplied when interface nailed down finally | |
90 | * This returns FALSE if there is an error which the caller need not recover | |
61caa4da DM |
91 | * from; otherwise TRUE. In either case the caller should look at *len [???]. |
92 | * It guarantees that the returned codepoint, *uv, when expressed as | |
7a4ca5b4 | 93 | * utf8 bytes, would fit within the skipped "\o{...}" bytes. |
db30362b | 94 | * On input: |
00ce5563 KW |
95 | * s is the address of a pointer to a NULL terminated string that begins |
96 | * with 'o', and the previous character was a backslash. At exit, *s | |
97 | * will be advanced to the byte just after those absorbed by this | |
98 | * function. Hence the caller can continue parsing from there. In | |
99 | * the case of an error, this routine has generally positioned *s to | |
100 | * point just to the right of the first bad spot, so that a message | |
101 | * that has a "<--" to mark the spot will be correctly positioned. | |
db30362b KW |
102 | * uv points to a UV that will hold the output value, valid only if the |
103 | * return from the function is TRUE | |
db30362b KW |
104 | * error_msg is a pointer that will be set to an internal buffer giving an |
105 | * error message upon failure (the return is FALSE). Untouched if | |
106 | * function succeeds | |
107 | * output_warning says whether to output any warning messages, or suppress | |
108 | * them | |
80f4111b KW |
109 | * strict is true if this should fail instead of warn if there are |
110 | * non-octal digits within the braces | |
17896a85 KW |
111 | * silence_non_portable is true if to suppress warnings about the code |
112 | * point returned being too large to fit on all platforms. | |
80f4111b | 113 | * UTF is true iff the string *s is encoded in UTF-8. |
db30362b | 114 | */ |
00ce5563 | 115 | char* e; |
db30362b KW |
116 | STRLEN numbers_len; |
117 | I32 flags = PERL_SCAN_ALLOW_UNDERSCORES | |
118 | | PERL_SCAN_DISALLOW_PREFIX | |
119 | /* XXX Until the message is improved in grok_oct, handle errors | |
120 | * ourselves */ | |
121 | | PERL_SCAN_SILENT_ILLDIGIT; | |
122 | ||
61caa4da DM |
123 | #ifdef DEBUGGING |
124 | char *start = *s - 1; | |
125 | assert(*start == '\\'); | |
126 | #endif | |
127 | ||
db30362b KW |
128 | PERL_ARGS_ASSERT_GROK_BSLASH_O; |
129 | ||
130 | ||
00ce5563 KW |
131 | assert(**s == 'o'); |
132 | (*s)++; | |
db30362b | 133 | |
00ce5563 | 134 | if (**s != '{') { |
db30362b KW |
135 | *error_msg = "Missing braces on \\o{}"; |
136 | return FALSE; | |
137 | } | |
138 | ||
00ce5563 | 139 | e = strchr(*s, '}'); |
db30362b | 140 | if (!e) { |
00ce5563 | 141 | (*s)++; /* Move past the '{' */ |
b8de99ca KW |
142 | while (isOCTAL(**s)) { /* Position beyond the legal digits */ |
143 | (*s)++; | |
144 | } | |
00ce5563 | 145 | *error_msg = "Missing right brace on \\o{"; |
db30362b KW |
146 | return FALSE; |
147 | } | |
148 | ||
00ce5563 KW |
149 | (*s)++; /* Point to expected first digit (could be first byte of utf8 |
150 | sequence if not a digit) */ | |
151 | numbers_len = e - *s; | |
db30362b | 152 | if (numbers_len == 0) { |
00ce5563 | 153 | (*s)++; /* Move past the } */ |
db30362b KW |
154 | *error_msg = "Number with no digits"; |
155 | return FALSE; | |
156 | } | |
157 | ||
17896a85 KW |
158 | if (silence_non_portable) { |
159 | flags |= PERL_SCAN_SILENT_NON_PORTABLE; | |
160 | } | |
161 | ||
00ce5563 | 162 | *uv = grok_oct(*s, &numbers_len, &flags, NULL); |
db30362b KW |
163 | /* Note that if has non-octal, will ignore everything starting with that up |
164 | * to the '}' */ | |
165 | ||
80f4111b KW |
166 | if (numbers_len != (STRLEN) (e - *s)) { |
167 | if (strict) { | |
168 | *s += numbers_len; | |
169 | *s += (UTF) ? UTF8SKIP(*s) : (STRLEN) 1; | |
170 | *error_msg = "Non-octal character"; | |
171 | return FALSE; | |
172 | } | |
173 | else if (output_warning) { | |
b67d718a KW |
174 | Perl_ck_warner(aTHX_ packWARN(WARN_DIGIT), |
175 | /* diag_listed_as: Non-octal character '%c'. Resolved as "%s" */ | |
176 | "Non-octal character '%c'. Resolved as \"\\o{%.*s}\"", | |
177 | *(*s + numbers_len), | |
178 | (int) numbers_len, | |
179 | *s); | |
80f4111b | 180 | } |
db30362b KW |
181 | } |
182 | ||
00ce5563 KW |
183 | /* Return past the '}' */ |
184 | *s = e + 1; | |
185 | ||
7a4ca5b4 | 186 | /* guarantee replacing "\o{...}" with utf8 bytes fits within |
61caa4da DM |
187 | * existing space */ |
188 | assert(OFFUNISKIP(*uv) < *s - start); | |
189 | ||
db30362b KW |
190 | return TRUE; |
191 | } | |
192 | ||
a0481293 | 193 | PERL_STATIC_INLINE bool |
00ce5563 | 194 | S_grok_bslash_x(pTHX_ char **s, UV *uv, const char** error_msg, |
80f4111b | 195 | const bool output_warning, const bool strict, |
17896a85 | 196 | const bool silence_non_portable, |
80f4111b | 197 | const bool UTF) |
a0481293 KW |
198 | { |
199 | ||
200 | /* Documentation to be supplied when interface nailed down finally | |
201 | * This returns FALSE if there is an error which the caller need not recover | |
61caa4da DM |
202 | * from; otherwise TRUE. |
203 | * It guarantees that the returned codepoint, *uv, when expressed as | |
204 | * utf8 bytes, would fit within the skipped "\x{...}" bytes. | |
205 | * | |
a0481293 | 206 | * On input: |
00ce5563 KW |
207 | * s is the address of a pointer to a NULL terminated string that begins |
208 | * with 'x', and the previous character was a backslash. At exit, *s | |
209 | * will be advanced to the byte just after those absorbed by this | |
210 | * function. Hence the caller can continue parsing from there. In | |
211 | * the case of an error, this routine has generally positioned *s to | |
212 | * point just to the right of the first bad spot, so that a message | |
213 | * that has a "<--" to mark the spot will be correctly positioned. | |
a0481293 KW |
214 | * uv points to a UV that will hold the output value, valid only if the |
215 | * return from the function is TRUE | |
a0481293 KW |
216 | * error_msg is a pointer that will be set to an internal buffer giving an |
217 | * error message upon failure (the return is FALSE). Untouched if | |
218 | * function succeeds | |
219 | * output_warning says whether to output any warning messages, or suppress | |
220 | * them | |
80f4111b KW |
221 | * strict is true if anything out of the ordinary should cause this to |
222 | * fail instead of warn or be silent. For example, it requires | |
223 | * exactly 2 digits following the \x (when there are no braces). | |
224 | * 3 digits could be a mistake, so is forbidden in this mode. | |
344451ec KW |
225 | * silence_non_portable is true if to suppress warnings about the code |
226 | * point returned being too large to fit on all platforms. | |
80f4111b | 227 | * UTF is true iff the string *s is encoded in UTF-8. |
a0481293 | 228 | */ |
00ce5563 | 229 | char* e; |
a0481293 | 230 | STRLEN numbers_len; |
344451ec | 231 | I32 flags = PERL_SCAN_DISALLOW_PREFIX; |
61caa4da DM |
232 | #ifdef DEBUGGING |
233 | char *start = *s - 1; | |
234 | assert(*start == '\\'); | |
235 | #endif | |
a0481293 KW |
236 | |
237 | PERL_ARGS_ASSERT_GROK_BSLASH_X; | |
238 | ||
00ce5563 KW |
239 | assert(**s == 'x'); |
240 | (*s)++; | |
a0481293 | 241 | |
879eb604 | 242 | if (strict || ! output_warning) { |
80f4111b KW |
243 | flags |= PERL_SCAN_SILENT_ILLDIGIT; |
244 | } | |
245 | ||
00ce5563 | 246 | if (**s != '{') { |
80f4111b KW |
247 | STRLEN len = (strict) ? 3 : 2; |
248 | ||
00ce5563 KW |
249 | *uv = grok_hex(*s, &len, &flags, NULL); |
250 | *s += len; | |
80f4111b KW |
251 | if (strict && len != 2) { |
252 | if (len < 2) { | |
253 | *s += (UTF) ? UTF8SKIP(*s) : 1; | |
254 | *error_msg = "Non-hex character"; | |
255 | } | |
256 | else { | |
257 | *error_msg = "Use \\x{...} for more than two hex characters"; | |
258 | } | |
259 | return FALSE; | |
260 | } | |
61caa4da | 261 | goto ok; |
a0481293 KW |
262 | } |
263 | ||
00ce5563 | 264 | e = strchr(*s, '}'); |
a0481293 | 265 | if (!e) { |
00ce5563 | 266 | (*s)++; /* Move past the '{' */ |
b8de99ca KW |
267 | while (isXDIGIT(**s)) { /* Position beyond the legal digits */ |
268 | (*s)++; | |
269 | } | |
a0481293 KW |
270 | /* XXX The corresponding message above for \o is just '\\o{'; other |
271 | * messages for other constructs include the '}', so are inconsistent. | |
272 | */ | |
273 | *error_msg = "Missing right brace on \\x{}"; | |
274 | return FALSE; | |
275 | } | |
276 | ||
00ce5563 KW |
277 | (*s)++; /* Point to expected first digit (could be first byte of utf8 |
278 | sequence if not a digit) */ | |
279 | numbers_len = e - *s; | |
80f4111b KW |
280 | if (numbers_len == 0) { |
281 | if (strict) { | |
282 | (*s)++; /* Move past the } */ | |
283 | *error_msg = "Number with no digits"; | |
284 | return FALSE; | |
285 | } | |
eb9a585f HS |
286 | *s = e + 1; |
287 | *uv = 0; | |
61caa4da | 288 | goto ok; |
80f4111b KW |
289 | } |
290 | ||
344451ec | 291 | flags |= PERL_SCAN_ALLOW_UNDERSCORES; |
17896a85 KW |
292 | if (silence_non_portable) { |
293 | flags |= PERL_SCAN_SILENT_NON_PORTABLE; | |
294 | } | |
80f4111b | 295 | |
00ce5563 | 296 | *uv = grok_hex(*s, &numbers_len, &flags, NULL); |
a0481293 KW |
297 | /* Note that if has non-hex, will ignore everything starting with that up |
298 | * to the '}' */ | |
299 | ||
80f4111b KW |
300 | if (strict && numbers_len != (STRLEN) (e - *s)) { |
301 | *s += numbers_len; | |
302 | *s += (UTF) ? UTF8SKIP(*s) : 1; | |
303 | *error_msg = "Non-hex character"; | |
304 | return FALSE; | |
305 | } | |
306 | ||
00ce5563 KW |
307 | /* Return past the '}' */ |
308 | *s = e + 1; | |
309 | ||
61caa4da DM |
310 | ok: |
311 | /* guarantee replacing "\x{...}" with utf8 bytes fits within | |
312 | * existing space */ | |
313 | assert(OFFUNISKIP(*uv) < *s - start); | |
a0481293 KW |
314 | return TRUE; |
315 | } | |
316 | ||
5e0a247b KW |
317 | STATIC char* |
318 | S_form_short_octal_warning(pTHX_ | |
319 | const char * const s, /* Points to first non-octal */ | |
320 | const STRLEN len /* Length of octals string, so | |
321 | (s-len) points to first | |
322 | octal */ | |
323 | ) { | |
324 | /* Return a character string consisting of a warning message for when a | |
325 | * string constant in octal is weird, like "\078". */ | |
326 | ||
327 | const char * sans_leading_zeros = s - len; | |
328 | ||
329 | PERL_ARGS_ASSERT_FORM_SHORT_OCTAL_WARNING; | |
330 | ||
331 | assert(*s == '8' || *s == '9'); | |
332 | ||
333 | /* Remove the leading zeros, retaining one zero so won't be zero length */ | |
334 | while (*sans_leading_zeros == '0') sans_leading_zeros++; | |
335 | if (sans_leading_zeros == s) { | |
336 | sans_leading_zeros--; | |
337 | } | |
338 | ||
339 | return Perl_form(aTHX_ | |
340 | "'%.*s' resolved to '\\o{%.*s}%c'", | |
341 | (int) (len + 2), s - len - 1, | |
342 | (int) (s - sans_leading_zeros), sans_leading_zeros, | |
343 | *s); | |
344 | } | |
345 | ||
04e98a4d | 346 | /* |
14d04a33 | 347 | * ex: set ts=8 sts=4 sw=4 et: |
04e98a4d | 348 | */ |