Commit 137dc7acb9f46dfe40b73dd0079bf130eb6981e0
1 parent
6c820d6c
Refactor: move resolution of literal to its own method
Showing
2 changed files
with
83 additions
and
76 deletions
include/qpdf/QPDFTokenizer.hh
| @@ -133,6 +133,7 @@ class QPDFTokenizer | @@ -133,6 +133,7 @@ class QPDFTokenizer | ||
| 133 | 133 | ||
| 134 | private: | 134 | private: |
| 135 | void reset(); | 135 | void reset(); |
| 136 | + void resolveLiteral(); | ||
| 136 | 137 | ||
| 137 | // Lexer state | 138 | // Lexer state |
| 138 | enum { st_top, st_in_comment, st_in_string, st_lt, st_gt, | 139 | enum { st_top, st_in_comment, st_in_string, st_lt, st_gt, |
libqpdf/QPDFTokenizer.cc
| @@ -50,10 +50,90 @@ QPDFTokenizer::reset() | @@ -50,10 +50,90 @@ QPDFTokenizer::reset() | ||
| 50 | } | 50 | } |
| 51 | 51 | ||
| 52 | void | 52 | void |
| 53 | -QPDFTokenizer::presentCharacter(char ch) | 53 | +QPDFTokenizer::resolveLiteral() |
| 54 | { | 54 | { |
| 55 | PCRE num_re("^[\\+\\-]?(?:\\.\\d+|\\d+(?:\\.\\d+)?)$"); | 55 | PCRE num_re("^[\\+\\-]?(?:\\.\\d+|\\d+(?:\\.\\d+)?)$"); |
| 56 | 56 | ||
| 57 | + if ((val.length() > 0) && (val[0] == '/')) | ||
| 58 | + { | ||
| 59 | + type = tt_name; | ||
| 60 | + // Deal with # in name token. Note: '/' by itself is a | ||
| 61 | + // valid name, so don't strip leading /. That way we | ||
| 62 | + // don't have to deal with the empty string as a name. | ||
| 63 | + std::string nval = "/"; | ||
| 64 | + char const* valstr = val.c_str() + 1; | ||
| 65 | + for (char const* p = valstr; *p; ++p) | ||
| 66 | + { | ||
| 67 | + if ((*p == '#') && this->pound_special_in_name) | ||
| 68 | + { | ||
| 69 | + if (p[1] && p[2] && | ||
| 70 | + is_hex_digit(p[1]) && is_hex_digit(p[2])) | ||
| 71 | + { | ||
| 72 | + char num[3]; | ||
| 73 | + num[0] = p[1]; | ||
| 74 | + num[1] = p[2]; | ||
| 75 | + num[2] = '\0'; | ||
| 76 | + char ch = (char)(strtol(num, 0, 16)); | ||
| 77 | + if (ch == '\0') | ||
| 78 | + { | ||
| 79 | + type = tt_bad; | ||
| 80 | + QTC::TC("qpdf", "QPDF_Tokenizer null in name"); | ||
| 81 | + error_message = | ||
| 82 | + "null character not allowed in name token"; | ||
| 83 | + nval += "#00"; | ||
| 84 | + } | ||
| 85 | + else | ||
| 86 | + { | ||
| 87 | + nval += ch; | ||
| 88 | + } | ||
| 89 | + p += 2; | ||
| 90 | + } | ||
| 91 | + else | ||
| 92 | + { | ||
| 93 | + QTC::TC("qpdf", "QPDF_Tokenizer bad name"); | ||
| 94 | + type = tt_bad; | ||
| 95 | + error_message = "invalid name token"; | ||
| 96 | + nval += *p; | ||
| 97 | + } | ||
| 98 | + } | ||
| 99 | + else | ||
| 100 | + { | ||
| 101 | + nval += *p; | ||
| 102 | + } | ||
| 103 | + } | ||
| 104 | + val = nval; | ||
| 105 | + } | ||
| 106 | + else if (num_re.match(val.c_str())) | ||
| 107 | + { | ||
| 108 | + if (val.find('.') != std::string::npos) | ||
| 109 | + { | ||
| 110 | + type = tt_real; | ||
| 111 | + } | ||
| 112 | + else | ||
| 113 | + { | ||
| 114 | + type = tt_integer; | ||
| 115 | + } | ||
| 116 | + } | ||
| 117 | + else if ((val == "true") || (val == "false")) | ||
| 118 | + { | ||
| 119 | + type = tt_bool; | ||
| 120 | + } | ||
| 121 | + else if (val == "null") | ||
| 122 | + { | ||
| 123 | + type = tt_null; | ||
| 124 | + } | ||
| 125 | + else | ||
| 126 | + { | ||
| 127 | + // I don't really know what it is, so leave it as tt_word. | ||
| 128 | + // Lots of cases ($, #, etc.) other than actual words fall | ||
| 129 | + // into this category, but that's okay at least for now. | ||
| 130 | + type = tt_word; | ||
| 131 | + } | ||
| 132 | +} | ||
| 133 | + | ||
| 134 | +void | ||
| 135 | +QPDFTokenizer::presentCharacter(char ch) | ||
| 136 | +{ | ||
| 57 | if (state == st_token_ready) | 137 | if (state == st_token_ready) |
| 58 | { | 138 | { |
| 59 | throw std::logic_error( | 139 | throw std::logic_error( |
| @@ -342,81 +422,7 @@ QPDFTokenizer::presentCharacter(char ch) | @@ -342,81 +422,7 @@ QPDFTokenizer::presentCharacter(char ch) | ||
| 342 | 422 | ||
| 343 | if ((state == st_token_ready) && (type == tt_word)) | 423 | if ((state == st_token_ready) && (type == tt_word)) |
| 344 | { | 424 | { |
| 345 | - if ((val.length() > 0) && (val[0] == '/')) | ||
| 346 | - { | ||
| 347 | - type = tt_name; | ||
| 348 | - // Deal with # in name token. Note: '/' by itself is a | ||
| 349 | - // valid name, so don't strip leading /. That way we | ||
| 350 | - // don't have to deal with the empty string as a name. | ||
| 351 | - std::string nval = "/"; | ||
| 352 | - char const* valstr = val.c_str() + 1; | ||
| 353 | - for (char const* p = valstr; *p; ++p) | ||
| 354 | - { | ||
| 355 | - if ((*p == '#') && this->pound_special_in_name) | ||
| 356 | - { | ||
| 357 | - if (p[1] && p[2] && | ||
| 358 | - is_hex_digit(p[1]) && is_hex_digit(p[2])) | ||
| 359 | - { | ||
| 360 | - char num[3]; | ||
| 361 | - num[0] = p[1]; | ||
| 362 | - num[1] = p[2]; | ||
| 363 | - num[2] = '\0'; | ||
| 364 | - char ch = (char)(strtol(num, 0, 16)); | ||
| 365 | - if (ch == '\0') | ||
| 366 | - { | ||
| 367 | - type = tt_bad; | ||
| 368 | - QTC::TC("qpdf", "QPDF_Tokenizer null in name"); | ||
| 369 | - error_message = | ||
| 370 | - "null character not allowed in name token"; | ||
| 371 | - nval += "#00"; | ||
| 372 | - } | ||
| 373 | - else | ||
| 374 | - { | ||
| 375 | - nval += ch; | ||
| 376 | - } | ||
| 377 | - p += 2; | ||
| 378 | - } | ||
| 379 | - else | ||
| 380 | - { | ||
| 381 | - QTC::TC("qpdf", "QPDF_Tokenizer bad name"); | ||
| 382 | - type = tt_bad; | ||
| 383 | - error_message = "invalid name token"; | ||
| 384 | - nval += *p; | ||
| 385 | - } | ||
| 386 | - } | ||
| 387 | - else | ||
| 388 | - { | ||
| 389 | - nval += *p; | ||
| 390 | - } | ||
| 391 | - } | ||
| 392 | - val = nval; | ||
| 393 | - } | ||
| 394 | - else if (num_re.match(val.c_str())) | ||
| 395 | - { | ||
| 396 | - if (val.find('.') != std::string::npos) | ||
| 397 | - { | ||
| 398 | - type = tt_real; | ||
| 399 | - } | ||
| 400 | - else | ||
| 401 | - { | ||
| 402 | - type = tt_integer; | ||
| 403 | - } | ||
| 404 | - } | ||
| 405 | - else if ((val == "true") || (val == "false")) | ||
| 406 | - { | ||
| 407 | - type = tt_bool; | ||
| 408 | - } | ||
| 409 | - else if (val == "null") | ||
| 410 | - { | ||
| 411 | - type = tt_null; | ||
| 412 | - } | ||
| 413 | - else | ||
| 414 | - { | ||
| 415 | - // I don't really know what it is, so leave it as tt_word. | ||
| 416 | - // Lots of cases ($, #, etc.) other than actual words fall | ||
| 417 | - // into this category, but that's okay at least for now. | ||
| 418 | - type = tt_word; | ||
| 419 | - } | 425 | + resolveLiteral(); |
| 420 | } | 426 | } |
| 421 | 427 | ||
| 422 | if (! (betweenTokens() || ((state == st_token_ready) && unread_char))) | 428 | if (! (betweenTokens() || ((state == st_token_ready) && unread_char))) |