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 |
5aaab254 | 18 | S_regcurly(pTHX_ 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 | } | |
04e98a4d AD |
33 | if (*s != '}') |
34 | return FALSE; | |
35 | return TRUE; | |
36 | } | |
db30362b | 37 | |
68b355dd KW |
38 | /* XXX Add documentation after final interface and behavior is decided */ |
39 | /* May want to show context for error, so would pass Perl_bslash_c(pTHX_ const char* current, const char* start, const bool output_warning) | |
40 | U8 source = *current; | |
68b355dd KW |
41 | */ |
42 | ||
43 | STATIC char | |
17a3df4c | 44 | S_grok_bslash_c(pTHX_ const char source, const bool utf8, const bool output_warning) |
68b355dd KW |
45 | { |
46 | ||
47 | U8 result; | |
48 | ||
17a3df4c KW |
49 | if (utf8) { |
50 | /* Trying to deprecate non-ASCII usages. This construct has never | |
51 | * worked for a utf8 variant. So, even though are accepting non-ASCII | |
52 | * Latin1 in 5.14, no need to make them work under utf8 */ | |
53 | if (! isASCII(source)) { | |
54 | Perl_croak(aTHX_ "Character following \"\\c\" must be ASCII"); | |
55 | } | |
68b355dd KW |
56 | } |
57 | ||
58 | result = toCTRL(source); | |
17a3df4c KW |
59 | if (! isASCII(source)) { |
60 | Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_SYNTAX), | |
61 | "Character following \"\\c\" must be ASCII"); | |
62 | } | |
63 | else if (! isCNTRL(result) && output_warning) { | |
68b355dd | 64 | if (source == '{') { |
17a3df4c KW |
65 | Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_SYNTAX), |
66 | "\"\\c{\" is deprecated and is more clearly written as \";\""); | |
68b355dd | 67 | } |
17a3df4c | 68 | else { |
68b355dd KW |
69 | U8 clearer[3]; |
70 | U8 i = 0; | |
0eb30aeb | 71 | if (! isWORDCHAR(result)) { |
68b355dd KW |
72 | clearer[i++] = '\\'; |
73 | } | |
74 | clearer[i++] = result; | |
75 | clearer[i++] = '\0'; | |
76 | ||
17a3df4c KW |
77 | Perl_ck_warner(aTHX_ packWARN(WARN_SYNTAX), |
78 | "\"\\c%c\" is more clearly written simply as \"%s\"", | |
68b355dd KW |
79 | source, |
80 | clearer); | |
81 | } | |
82 | } | |
83 | ||
84 | return result; | |
85 | } | |
86 | ||
db30362b KW |
87 | STATIC bool |
88 | S_grok_bslash_o(pTHX_ const char *s, | |
89 | UV *uv, | |
90 | STRLEN *len, | |
91 | const char** error_msg, | |
92 | const bool output_warning) | |
93 | { | |
94 | ||
95 | /* Documentation to be supplied when interface nailed down finally | |
96 | * This returns FALSE if there is an error which the caller need not recover | |
97 | * from; , otherwise TRUE. In either case the caller should look at *len | |
98 | * On input: | |
99 | * s points to a string that begins with 'o', and the previous character | |
100 | * was a backslash. | |
101 | * uv points to a UV that will hold the output value, valid only if the | |
102 | * return from the function is TRUE | |
103 | * len on success will point to the next character in the string past the | |
104 | * end of this construct. | |
105 | * on failure, it will point to the failure | |
106 | * error_msg is a pointer that will be set to an internal buffer giving an | |
107 | * error message upon failure (the return is FALSE). Untouched if | |
108 | * function succeeds | |
109 | * output_warning says whether to output any warning messages, or suppress | |
110 | * them | |
111 | */ | |
112 | const char* e; | |
113 | STRLEN numbers_len; | |
114 | I32 flags = PERL_SCAN_ALLOW_UNDERSCORES | |
115 | | PERL_SCAN_DISALLOW_PREFIX | |
116 | /* XXX Until the message is improved in grok_oct, handle errors | |
117 | * ourselves */ | |
118 | | PERL_SCAN_SILENT_ILLDIGIT; | |
119 | ||
120 | PERL_ARGS_ASSERT_GROK_BSLASH_O; | |
121 | ||
122 | ||
123 | assert(*s == 'o'); | |
124 | s++; | |
125 | ||
126 | if (*s != '{') { | |
127 | *len = 1; /* Move past the o */ | |
128 | *error_msg = "Missing braces on \\o{}"; | |
129 | return FALSE; | |
130 | } | |
131 | ||
132 | e = strchr(s, '}'); | |
133 | if (!e) { | |
134 | *len = 2; /* Move past the o{ */ | |
135 | *error_msg = "Missing right brace on \\o{"; | |
136 | return FALSE; | |
137 | } | |
138 | ||
139 | /* Return past the '}' no matter what is inside the braces */ | |
7d2838d8 | 140 | *len = e - s + 2; /* 2 = 1 for the 'o' + 1 for the '}' */ |
db30362b KW |
141 | |
142 | s++; /* Point to first digit */ | |
143 | ||
144 | numbers_len = e - s; | |
145 | if (numbers_len == 0) { | |
146 | *error_msg = "Number with no digits"; | |
147 | return FALSE; | |
148 | } | |
149 | ||
8b8acc9e | 150 | *uv = grok_oct(s, &numbers_len, &flags, NULL); |
db30362b KW |
151 | /* Note that if has non-octal, will ignore everything starting with that up |
152 | * to the '}' */ | |
153 | ||
154 | if (output_warning && numbers_len != (STRLEN) (e - s)) { | |
155 | Perl_ck_warner(aTHX_ packWARN(WARN_DIGIT), | |
156 | /* diag_listed_as: Non-octal character '%c'. Resolved as "%s" */ | |
157 | "Non-octal character '%c'. Resolved as \"\\o{%.*s}\"", | |
158 | *(s + numbers_len), | |
159 | (int) numbers_len, | |
160 | s); | |
161 | } | |
162 | ||
163 | return TRUE; | |
164 | } | |
165 | ||
a0481293 KW |
166 | PERL_STATIC_INLINE bool |
167 | S_grok_bslash_x(pTHX_ const char *s, | |
168 | UV *uv, | |
169 | STRLEN *len, | |
170 | const char** error_msg, | |
171 | const bool output_warning) | |
172 | { | |
173 | ||
174 | /* Documentation to be supplied when interface nailed down finally | |
175 | * This returns FALSE if there is an error which the caller need not recover | |
176 | * from; , otherwise TRUE. In either case the caller should look at *len | |
177 | * On input: | |
178 | * s points to a string that begins with 'x', and the previous character | |
179 | * was a backslash. | |
180 | * uv points to a UV that will hold the output value, valid only if the | |
181 | * return from the function is TRUE | |
182 | * len on success will point to the next character in the string past the | |
183 | * end of this construct. | |
184 | * on failure, it will point to the failure | |
185 | * error_msg is a pointer that will be set to an internal buffer giving an | |
186 | * error message upon failure (the return is FALSE). Untouched if | |
187 | * function succeeds | |
188 | * output_warning says whether to output any warning messages, or suppress | |
189 | * them | |
190 | */ | |
191 | const char* e; | |
192 | STRLEN numbers_len; | |
193 | I32 flags = PERL_SCAN_ALLOW_UNDERSCORES | |
194 | | PERL_SCAN_DISALLOW_PREFIX; | |
195 | ||
196 | PERL_ARGS_ASSERT_GROK_BSLASH_X; | |
197 | ||
4f8dbb2d | 198 | PERL_UNUSED_ARG(output_warning); |
a0481293 KW |
199 | |
200 | assert(*s == 'x'); | |
201 | s++; | |
202 | ||
203 | if (*s != '{') { | |
204 | I32 flags = PERL_SCAN_DISALLOW_PREFIX; | |
205 | *len = 2; | |
206 | *uv = grok_hex(s, len, &flags, NULL); | |
207 | (*len)++; | |
208 | return TRUE; | |
209 | } | |
210 | ||
211 | e = strchr(s, '}'); | |
212 | if (!e) { | |
213 | *len = 2; /* Move past the 'x{' */ | |
214 | /* XXX The corresponding message above for \o is just '\\o{'; other | |
215 | * messages for other constructs include the '}', so are inconsistent. | |
216 | */ | |
217 | *error_msg = "Missing right brace on \\x{}"; | |
218 | return FALSE; | |
219 | } | |
220 | ||
221 | /* Return past the '}' no matter what is inside the braces */ | |
222 | *len = e - s + 2; /* 2 = 1 for the 'x' + 1 for the '}' */ | |
223 | ||
224 | s++; /* Point to first digit */ | |
225 | ||
226 | numbers_len = e - s; | |
227 | *uv = grok_hex(s, &numbers_len, &flags, NULL); | |
228 | /* Note that if has non-hex, will ignore everything starting with that up | |
229 | * to the '}' */ | |
230 | ||
231 | return TRUE; | |
232 | } | |
233 | ||
04e98a4d AD |
234 | /* |
235 | * Local variables: | |
236 | * c-indentation-style: bsd | |
237 | * c-basic-offset: 4 | |
14d04a33 | 238 | * indent-tabs-mode: nil |
04e98a4d AD |
239 | * End: |
240 | * | |
14d04a33 | 241 | * ex: set ts=8 sts=4 sw=4 et: |
04e98a4d | 242 | */ |