1 | /* |
2 | * Copyright (C) 2009-2010, Pino Toscano <pino@kde.org> |
3 | * Copyright (C) 2010, Hib Eris <hib@hiberis.nl> |
4 | * Copyright (C) 2014, 2015 Hans-Peter Deifel <hpdeifel@gmx.de> |
5 | * Copyright (C) 2015, Tamas Szekeres <szekerest@gmail.com> |
6 | * Copyright (C) 2016 Jakub Alba <jakubalba@gmail.com> |
7 | * Copyright (C) 2018, 2020-2022, Albert Astals Cid <aacid@kde.org> |
8 | * Copyright (C) 2018 Suzuki Toshiya <mpsuzuki@hiroshima-u.ac.jp> |
9 | * Copyright (C) 2018, 2020, Adam Reichold <adam.reichold@t-online.de> |
10 | * Copyright (C) 2022, Oliver Sander <oliver.sander@tu-dresden.de> |
11 | * |
12 | * This program is free software; you can redistribute it and/or modify |
13 | * it under the terms of the GNU General Public License as published by |
14 | * the Free Software Foundation; either version 2, or (at your option) |
15 | * any later version. |
16 | * |
17 | * This program is distributed in the hope that it will be useful, |
18 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
19 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
20 | * GNU General Public License for more details. |
21 | * |
22 | * You should have received a copy of the GNU General Public License |
23 | * along with this program; if not, write to the Free Software |
24 | * Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA 02110-1301, USA. |
25 | */ |
26 | |
27 | /** |
28 | \file poppler-global.h |
29 | */ |
30 | #include "poppler-global.h" |
31 | |
32 | #include "poppler-private.h" |
33 | #include "poppler-document-private.h" |
34 | |
35 | #include "DateInfo.h" |
36 | |
37 | #include <algorithm> |
38 | |
39 | #include <cerrno> |
40 | #include <cstring> |
41 | #include <ios> |
42 | #include <iostream> |
43 | |
44 | #include <iconv.h> |
45 | |
46 | #include "config.h" |
47 | |
48 | namespace { |
49 | |
50 | struct MiniIconv |
51 | { |
52 | MiniIconv(const char *to_code, const char *from_code) : i_(iconv_open(tocode: to_code, fromcode: from_code)) { } |
53 | ~MiniIconv() |
54 | { |
55 | if (is_valid()) { |
56 | iconv_close(cd: i_); |
57 | } |
58 | } |
59 | MiniIconv(const MiniIconv &) = delete; |
60 | MiniIconv &operator=(const MiniIconv &) = delete; |
61 | bool is_valid() const { return i_ != (iconv_t)-1; } |
62 | explicit operator iconv_t() const { return i_; } |
63 | iconv_t i_; |
64 | }; |
65 | |
66 | } |
67 | |
68 | using namespace poppler; |
69 | |
70 | /** |
71 | \namespace poppler |
72 | |
73 | Single namespace containing all the classes and functions of poppler-cpp. |
74 | */ |
75 | |
76 | /** |
77 | \class poppler::noncopyable |
78 | |
79 | A class that cannot be copied. |
80 | */ |
81 | |
82 | /** |
83 | \enum poppler::rotation_enum |
84 | |
85 | The case sensitivity. |
86 | */ |
87 | /** |
88 | \var poppler::rotation_enum poppler::rotate_0 |
89 | |
90 | A rotation of 0 degrees clockwise. |
91 | */ |
92 | /** |
93 | \var poppler::rotation_enum poppler::rotate_90 |
94 | |
95 | A rotation of 90 degrees clockwise. |
96 | */ |
97 | /** |
98 | \var poppler::rotation_enum poppler::rotate_180 |
99 | |
100 | A rotation of 180 degrees clockwise. |
101 | */ |
102 | /** |
103 | \var poppler::rotation_enum poppler::rotate_270 |
104 | |
105 | A rotation of 270 degrees clockwise. |
106 | */ |
107 | |
108 | /** |
109 | \enum poppler::page_box_enum |
110 | |
111 | A possible box of a page in a PDF %document. |
112 | */ |
113 | /** |
114 | \var poppler::page_box_enum poppler::media_box |
115 | |
116 | The "media" box. |
117 | */ |
118 | /** |
119 | \var poppler::page_box_enum poppler::crop_box |
120 | |
121 | The "crop" box. |
122 | */ |
123 | /** |
124 | \var poppler::page_box_enum poppler::bleed_box |
125 | |
126 | The "bleed" box. |
127 | */ |
128 | /** |
129 | \var poppler::page_box_enum poppler::trim_box |
130 | |
131 | The "trim" box. |
132 | */ |
133 | /** |
134 | \var poppler::page_box_enum poppler::art_box |
135 | |
136 | The "art" box. |
137 | */ |
138 | |
139 | /** |
140 | \enum poppler::permission_enum |
141 | |
142 | A possible permission in a PDF %document. |
143 | */ |
144 | /** |
145 | \var poppler::permission_enum poppler::perm_print |
146 | |
147 | The permission to allow the print of a %document. |
148 | */ |
149 | /** |
150 | \var poppler::permission_enum poppler::perm_change |
151 | |
152 | The permission to change a %document. |
153 | |
154 | This is a generic "change" permission, so other permissions could affect |
155 | some types of changes. |
156 | */ |
157 | /** |
158 | \var poppler::permission_enum poppler::perm_copy |
159 | |
160 | The permission to allow the copy or extraction of the text in a %document. |
161 | */ |
162 | /** |
163 | \var poppler::permission_enum poppler::perm_add_notes |
164 | |
165 | The permission to allow the addition or editing of annotations, |
166 | and the filling of interactive form fields (including signature fields). |
167 | */ |
168 | /** |
169 | \var poppler::permission_enum poppler::perm_fill_forms |
170 | |
171 | The permission to allow the filling of interactive form fields |
172 | (including signature fields). |
173 | |
174 | \note this permission can be set even when the \ref poppler::perm_add_notes "perm_add_notes" |
175 | is not: this means that only the filling of forms is allowed. |
176 | */ |
177 | /** |
178 | \var poppler::permission_enum poppler::perm_accessibility |
179 | |
180 | The permission to allow the extracting of content (for example, text) for |
181 | accessibility usage (e.g. for a screen reader). |
182 | */ |
183 | /** |
184 | \var poppler::permission_enum poppler::perm_assemble |
185 | |
186 | The permission to allow to "assemble" a %document. |
187 | |
188 | This implies operations such as the insertion, the rotation and the deletion |
189 | of pages; the creation of bookmarks and thumbnail images. |
190 | |
191 | \note this permission can be set even when the \ref poppler::perm_change "perm_change" |
192 | is not |
193 | */ |
194 | /** |
195 | \var poppler::permission_enum poppler::perm_print_high_resolution |
196 | |
197 | The permission to allow the high resolution print of a %document. |
198 | */ |
199 | |
200 | /** |
201 | \enum poppler::case_sensitivity_enum |
202 | |
203 | The case sensitivity. |
204 | */ |
205 | |
206 | noncopyable::noncopyable() { } |
207 | |
208 | noncopyable::~noncopyable() { } |
209 | |
210 | noncopyable &noncopyable::operator=(noncopyable &&other) noexcept = default; |
211 | |
212 | ustring::ustring() { } |
213 | |
214 | ustring::ustring(size_type len, value_type ch) : std::basic_string<value_type>(len, ch) { } |
215 | |
216 | ustring::~ustring() { } |
217 | |
218 | byte_array ustring::to_utf8() const |
219 | { |
220 | if (!size()) { |
221 | return byte_array(); |
222 | } |
223 | |
224 | #ifdef WORDS_BIGENDIAN |
225 | MiniIconv ic("UTF-8" , "UTF-16BE" ); |
226 | #else |
227 | MiniIconv ic("UTF-8" , "UTF-16LE" ); |
228 | #endif |
229 | if (!ic.is_valid()) { |
230 | return byte_array(); |
231 | } |
232 | const value_type *me_data = data(); |
233 | byte_array str(size() * sizeof(value_type)); |
234 | char *str_data = &str[0]; |
235 | size_t me_len_char = size() * sizeof(value_type); |
236 | size_t str_len_left = str.size(); |
237 | size_t ir = iconv(cd: static_cast<iconv_t>(ic), inbuf: (ICONV_CONST char **)&me_data, inbytesleft: &me_len_char, outbuf: &str_data, outbytesleft: &str_len_left); |
238 | if ((ir == (size_t)-1) && (errno == E2BIG)) { |
239 | const size_t delta = str_data - &str[0]; |
240 | str_len_left += str.size(); |
241 | str.resize(new_size: str.size() * 2); |
242 | str_data = &str[delta]; |
243 | ir = iconv(cd: static_cast<iconv_t>(ic), inbuf: (ICONV_CONST char **)&me_data, inbytesleft: &me_len_char, outbuf: &str_data, outbytesleft: &str_len_left); |
244 | if (ir == (size_t)-1) { |
245 | return byte_array(); |
246 | } |
247 | } |
248 | str.resize(new_size: str.size() - str_len_left); |
249 | return str; |
250 | } |
251 | |
252 | std::string ustring::to_latin1() const |
253 | { |
254 | if (!size()) { |
255 | return std::string(); |
256 | } |
257 | |
258 | const size_type mylength = size(); |
259 | std::string ret(mylength, '\0'); |
260 | const value_type *me = data(); |
261 | for (size_type i = 0; i < mylength; ++i) { |
262 | ret[i] = (char)*me++; |
263 | } |
264 | return ret; |
265 | } |
266 | |
267 | ustring ustring::from_utf8(const char *str, int len) |
268 | { |
269 | if (len <= 0) { |
270 | len = std::strlen(s: str); |
271 | if (len <= 0) { |
272 | return ustring(); |
273 | } |
274 | } |
275 | |
276 | #ifdef WORDS_BIGENDIAN |
277 | MiniIconv ic("UTF-16BE" , "UTF-8" ); |
278 | #else |
279 | MiniIconv ic("UTF-16LE" , "UTF-8" ); |
280 | #endif |
281 | if (!ic.is_valid()) { |
282 | return ustring(); |
283 | } |
284 | |
285 | // +1, because iconv inserts byte order marks |
286 | ustring ret(len + 1, 0); |
287 | char *ret_data = reinterpret_cast<char *>(&ret[0]); |
288 | char *str_data = const_cast<char *>(str); |
289 | size_t str_len_char = len; |
290 | size_t ret_len_left = ret.size() * sizeof(ustring::value_type); |
291 | size_t ir = iconv(cd: static_cast<iconv_t>(ic), inbuf: (ICONV_CONST char **)&str_data, inbytesleft: &str_len_char, outbuf: &ret_data, outbytesleft: &ret_len_left); |
292 | if ((ir == (size_t)-1) && (errno == E2BIG)) { |
293 | const size_t delta = ret_data - reinterpret_cast<char *>(&ret[0]); |
294 | ret_len_left += ret.size() * sizeof(ustring::value_type); |
295 | ret.resize(n: ret.size() * 2); |
296 | ret_data = reinterpret_cast<char *>(&ret[0]) + delta; |
297 | ir = iconv(cd: static_cast<iconv_t>(ic), inbuf: (ICONV_CONST char **)&str_data, inbytesleft: &str_len_char, outbuf: &ret_data, outbytesleft: &ret_len_left); |
298 | if (ir == (size_t)-1) { |
299 | return ustring(); |
300 | } |
301 | } |
302 | ret.resize(n: ret.size() - ret_len_left / sizeof(ustring::value_type)); |
303 | |
304 | return ret; |
305 | } |
306 | |
307 | ustring ustring::from_latin1(const std::string &str) |
308 | { |
309 | const size_type l = str.size(); |
310 | if (!l) { |
311 | return ustring(); |
312 | } |
313 | const char *c = str.data(); |
314 | ustring ret(l, 0); |
315 | for (size_type i = 0; i < l; ++i) { |
316 | ret[i] = static_cast<unsigned char>(*c); |
317 | c++; |
318 | } |
319 | return ret; |
320 | } |
321 | |
322 | /** |
323 | Converts a string representing a PDF date to a value compatible with time_type. |
324 | */ |
325 | time_type poppler::convert_date(const std::string &date) |
326 | { |
327 | GooString gooDateStr(date.c_str()); |
328 | return static_cast<time_type>(dateStringToTime(dateString: &gooDateStr)); |
329 | } |
330 | |
331 | /** |
332 | Converts a string representing a PDF date to a value compatible with time_t. |
333 | */ |
334 | time_t poppler::convert_date_t(const std::string &date) |
335 | { |
336 | GooString gooDateStr(date.c_str()); |
337 | return dateStringToTime(dateString: &gooDateStr); |
338 | } |
339 | |
340 | std::ostream &poppler::operator<<(std::ostream &stream, const byte_array &array) |
341 | { |
342 | stream << "[" ; |
343 | const std::ios_base::fmtflags f = stream.flags(); |
344 | std::hex(base&: stream); |
345 | const char *data = &array[0]; |
346 | const byte_array::size_type out_len = std::min<byte_array::size_type>(a: array.size(), b: 50); |
347 | for (byte_array::size_type i = 0; i < out_len; ++i) { |
348 | if (i != 0) { |
349 | stream << " " ; |
350 | } |
351 | stream << ((data[i] & 0xf0) >> 4) << (data[i] & 0xf); |
352 | } |
353 | stream.flags(fmtfl: f); |
354 | if (out_len < array.size()) { |
355 | stream << " ..." ; |
356 | } |
357 | stream << "]" ; |
358 | return stream; |
359 | } |
360 | |
361 | /** |
362 | * Sets a custom data directory for initialization of global parameters |
363 | * |
364 | * If no instances of \see document currently exist, this will save the |
365 | * given path as a custom data directory to be used when the first instance |
366 | * of the \see document is constructed. |
367 | * |
368 | * \returns true on success, false on failure |
369 | * |
370 | * \since 0.73.0 |
371 | */ |
372 | bool poppler::set_data_dir(const std::string &new_data_dir) |
373 | { |
374 | return GlobalParamsIniter::setCustomDataDir(new_data_dir); |
375 | } |
376 | |
377 | /** |
378 | \typedef poppler::debug_func |
379 | |
380 | Debug/error function. |
381 | |
382 | This function type is used for debugging & error output; |
383 | the first parameter is the actual message, the second is the unaltered |
384 | closure argument which was passed to the set_debug_error_function() call. |
385 | |
386 | \since 0.30.0 |
387 | */ |
388 | |
389 | /** |
390 | Set a new debug/error output function. |
391 | |
392 | If not set, by default error and debug messages will be sent to stderr. |
393 | |
394 | \param debug_function the new debug function |
395 | \param closure user data which will be passed as-is to the debug function |
396 | |
397 | \since 0.30.0 |
398 | */ |
399 | void poppler::set_debug_error_function(debug_func debug_function, void *closure) |
400 | { |
401 | poppler::detail::user_debug_function = debug_function; |
402 | poppler::detail::debug_closure = closure; |
403 | } |
404 | |