Coverage Report

Created: 2026-09-14 07:34

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/ghostpdl/devices/vector/gdevpdtb.c
Line
Count
Source
1
/* Copyright (C) 2001-2024 Artifex Software, Inc.
2
   All Rights Reserved.
3
4
   This software is provided AS-IS with no warranty, either express or
5
   implied.
6
7
   This software is distributed under license and may not be copied,
8
   modified or distributed except as expressly authorized under the terms
9
   of the license contained in the file LICENSE in this distribution.
10
11
   Refer to licensing information at http://www.artifex.com or contact
12
   Artifex Software, Inc.,  39 Mesa Street, Suite 108A, San Francisco,
13
   CA 94129, USA, for further information.
14
*/
15
16
17
/* BaseFont implementation for pdfwrite */
18
#include "memory_.h"
19
#include "ctype_.h"
20
#include "gx.h"
21
#include "gserrors.h"
22
#include "gsutil.h"   /* for bytes_compare */
23
#include "gxfcid.h"
24
#include "gxfcopy.h"
25
#include "gxfont.h"   /* for gxfont42.h */
26
#include "gxfont42.h"
27
#include "gdevpsf.h"
28
#include "gdevpdfx.h"
29
#include "gdevpdfo.h"
30
#include "gdevpdtb.h"
31
#include "gdevpdtf.h"
32
#include "gdevpdtd.h"
33
#include "smd5.h"
34
#include "gxfcache.h"   /* for gs_purge_font_from_char_caches_completely */
35
/*
36
 * Adobe's Distiller Parameters documentation for Acrobat Distiller 5
37
 * says that all fonts other than Type 1 are always subsetted; the
38
 * documentation for Distiller 4 says that fonts other than Type 1 and
39
 * TrueType are subsetted.  We do the latter, except that we always
40
 * subset large TrueType fonts.
41
 */
42
19.8k
#define MAX_NO_SUBSET_GLYPHS 4096  /* arbitrary */
43
44
/* ---------------- Private ---------------- */
45
46
private_st_pdf_base_font();
47
gs_private_st_basic(st_pdf_base_font, pdf_base_font_t, "pdf_base_font_t",\
48
                    pdf_base_font_ptrs, pdf_base_font_data);
49
50
typedef struct pdf_base14_font_info_s {
51
    const char *urwname;
52
    const char *stdname;
53
} pdf_base14_font_info_t;
54
55
static const pdf_base14_font_info_t base14_font_info[] = {
56
    /* Standard mapping of URW fonts */
57
    {"NimbusMonL-Regu",       "Courier"              },
58
    {"NimbusMonL-Bold",       "Courier-Bold"         },
59
    {"NimbusMonL-ReguObli",   "Courier-Oblique"      },
60
    {"NimbusMonL-BoldObli",   "Courier-BoldOblique"  },
61
    {"NimbusSanL-Regu",       "Helvetica"            },
62
    {"NimbusSanL-Bold",       "Helvetica-Bold"       },
63
    {"NimbusSanL-ReguItal",   "Helvetica-Oblique"    },
64
    {"NimbusSanL-BoldItal",   "Helvetica-BoldOblique"},
65
    {"StandardSymL",          "Symbol"               },
66
    {"NimbusRomNo9L-Regu",    "Times-Roman"          },
67
    {"NimbusRomNo9L-Medi",    "Times-Bold"           },
68
    {"NimbusRomNo9L-ReguItal","Times-Italic"         },
69
    {"NimbusRomNo9L-MediItal","Times-BoldItalic"     },
70
    {"Dingbats",              "ZapfDingbats"         },
71
    /* A few other mappings of URW fonts */
72
    {"NimbusMono-Reg",        "Courier"              },
73
    {"NimbusMono-Bol",        "Courier-Bold"         },
74
    {"NimbusMono-Ita",        "Courier-Oblique"      },
75
    {"NimbusMono-BolIta",     "Courier-BoldOblique"   },
76
    {"NimbusSan-Reg",         "Helvetica"            },
77
    {"NimbusSan-Bol",         "Helvetica-Bold"       },
78
    {"NimbusSan-Ita",         "Helvetica-Oblique"    },
79
    {"NimbusSan-BolIta",      "Helvetica-BoldOblique"},
80
    {"A030-Reg",              "Helvetica"            },
81
    {"A030-Bol",              "Helvetica-Bold"       },
82
    {"A030-Ita",              "Helvetica-Oblique"    },
83
    {"A030-BolIta",           "Helvetica-BoldOblique"},
84
    {"NimbusSanNo2-Reg",      "Helvetica"            },
85
    {"NimbusSanNo2-Bol",      "Helvetica-Bold"       },
86
    {"NimbusSanNo2-Ita",      "Helvetica-Oblique"    },
87
    {"NimbusSanNo2-BolIta",   "Helvetica-BoldOblique"},
88
    {"NimbusRomanNo4-Lig",    "Times-Roman"          },
89
    {"NimbusRomanNo4-Bol",    "Times-Bold"           },
90
    {"NimbusRomanNo4-LigIta", "Times-Italic"         },
91
    {"NimbusRomanNo4-BolIta", "Times-BoldItalic"     },
92
    {"NimbusRomanNo9-Reg",    "Times-Roman"          },
93
    {"NimbusRomanNo9-Med",    "Times-Bold"           },
94
    {"NimbusRomanNo9-Ita",    "Times-Italic"         },
95
    {"NimbusRomanNo9-MedIta", "Times-BoldItalic"     },
96
    {"NimbusRom-Reg",         "Times-Roman"          },
97
    {"NimbusRom-Med",         "Times-Bold"           },
98
    {"NimbusRom-Ita",         "Times-Italic"         },
99
    {"NimbusRom-MedIta",      "Times-BoldItalic"     },
100
    {"NimbusRomNo9-Reg",      "Times-Roman"          },
101
    {"NimbusRomNo9-Bol",      "Times-Bold"           },
102
    {"NimbusRomNo9-Ita",      "Times-Italic"         },
103
    {"NimbusRomNo9-MedIta",   "Times-BoldItalic"     },
104
    {0}
105
};
106
107
/* Given a pointer to a font name, return a pointer to an
108
 * equivalent base 14 font name, of NULL if there is no
109
 * equivalent.
110
 */
111
const char *pdf_find_base14_name(const byte *str, uint size)
112
28
{
113
28
    const pdf_base14_font_info_t *ppsf;
114
115
1.31k
    for (ppsf = base14_font_info; ppsf->urwname; ++ppsf) {
116
1.28k
        if (strlen(ppsf->urwname) == size) {
117
10
            if (!memcmp(ppsf->urwname, (const char *)str, size))
118
0
                return ppsf->stdname;
119
10
        }
120
1.28k
    }
121
28
    return NULL;
122
28
}
123
124
125
/*
126
 * Determine whether a font is a subset font by examining the name.
127
 */
128
bool
129
pdf_has_subset_prefix(const byte *str, uint size)
130
8.64M
{
131
8.64M
    int i;
132
133
8.64M
    if (size < SUBSET_PREFIX_SIZE || str[SUBSET_PREFIX_SIZE - 1] != '+')
134
5.15M
        return false;
135
24.4M
    for (i = 0; i < SUBSET_PREFIX_SIZE - 1; ++i)
136
20.9M
        if ((uint)(str[i] - 'A') >= 26)
137
292
            return false;
138
3.49M
    return true;
139
3.49M
}
140
141
static inline ulong
142
hash(ulong v, int index, ushort w)
143
1.88M
{
144
1.88M
    return v * 3141592653u + w;
145
1.88M
}
146
147
/*
148
 * Add the XXXXXX+ prefix for a subset font.
149
 */
150
int
151
pdf_add_subset_prefix(const gx_device_pdf *pdev, gs_string *pstr, byte *used, int count, char *md5_hash)
152
18.8k
{
153
18.8k
    uint size = pstr->size;
154
18.8k
    byte *data = gs_resize_string(pdev->pdf_memory, pstr->data, size,
155
18.8k
                                  size + SUBSET_PREFIX_SIZE,
156
18.8k
                                  "pdf_add_subset_prefix");
157
18.8k
    int len = (count + 7) / 8;
158
18.8k
    int len0 = len & ~(sizeof(ushort) - 1);
159
18.8k
    ulong v = 0;
160
18.8k
    int i;
161
162
18.8k
    if (data == 0)
163
0
        return_error(gs_error_VMerror);
164
165
18.8k
    if (md5_hash) {
166
94.1k
        for (i = 0; i < 8; i += sizeof(ushort)) {
167
75.2k
      v = hash(v, i, (*(md5_hash + i) | *(md5_hash + i + 1) << 8));
168
75.2k
        }
169
18.8k
    }
170
171
    /* Hash the 'used' array. */
172
1.83M
    for (i = 0; i < len0; i += sizeof(ushort))
173
1.81M
        v = hash(v, i, *(ushort *)(used + i));
174
19.4k
    for (; i < len; i++)
175
554
        v = hash(v, i, used[i]);
176
177
18.8k
    memmove(data + SUBSET_PREFIX_SIZE, data, size);
178
132k
    for (i = 0; i < SUBSET_PREFIX_SIZE - 1; ++i, v /= 26)
179
113k
        data[i] = 'A' + (v % 26);
180
18.8k
    data[SUBSET_PREFIX_SIZE - 1] = '+';
181
18.8k
    pstr->data = data;
182
18.8k
    pstr->size = size + SUBSET_PREFIX_SIZE;
183
18.8k
    return 0;
184
18.8k
}
185
186
/* Finish writing FontFile* data. */
187
static int
188
pdf_end_fontfile(gx_device_pdf *pdev, pdf_data_writer_t *pdw)
189
18.5k
{
190
    /* We would like to call pdf_end_data,
191
       but we don't want to write the object to the output file now. */
192
18.5k
    return pdf_close_aside(pdw->pdev);
193
18.5k
}
194
195
static int copied_font_notify(void *proc_data, void *event_data)
196
24.6k
{
197
24.6k
    return gs_purge_font_from_char_caches_completely((gs_font *)proc_data);
198
24.6k
}
199
200
/* ---------------- Public ---------------- */
201
/*
202
 * Allocate and initialize a base font structure, making the required
203
 * stable copy/ies of the gs_font.  Note that this removes any XXXXXX+
204
 * font name prefix from the copy.  If is_standard is true, the copy is
205
 * a complete one, and adding glyphs or Encoding entries is not allowed.
206
 */
207
int
208
pdf_base_font_alloc(gx_device_pdf *pdev, pdf_base_font_t **ppbfont,
209
                    gs_font_base *font, const gs_matrix *orig_matrix,
210
                    bool is_standard)
211
40.8k
{
212
40.8k
    gs_memory_t *mem = pdev->pdf_memory;
213
40.8k
    gs_font *copied;
214
40.8k
    gs_font *complete;
215
40.8k
    pdf_base_font_t *pbfont =
216
40.8k
        gs_alloc_struct(mem, pdf_base_font_t,
217
40.8k
                        &st_pdf_base_font, "pdf_base_font_alloc");
218
40.8k
    const gs_font_name *pfname = &font->font_name;
219
40.8k
    gs_const_string font_name;
220
40.8k
    char fnbuf[2*sizeof(long) + 3]; /* .F########\0 */
221
40.8k
    int code, reserve_glyphs = -1;
222
223
40.8k
    if (pbfont == 0)
224
0
        return_error(gs_error_VMerror);
225
40.8k
    memset(pbfont, 0, sizeof(*pbfont));
226
40.8k
    switch (font->FontType) {
227
20.3k
    case ft_encrypted:
228
20.6k
    case ft_encrypted2:
229
20.6k
        {
230
20.6k
            int index, count;
231
20.6k
            gs_glyph glyph;
232
233
20.6k
            for (index = 0, count = 0;
234
16.1M
                (font->procs.enumerate_glyph((gs_font *)font, &index,
235
16.1M
                                              GLYPH_SPACE_NAME, &glyph),
236
16.1M
                  index != 0);
237
20.6k
                 )
238
16.1M
                    ++count;
239
20.6k
            pbfont->num_glyphs = count;
240
20.6k
            pbfont->do_subset = (is_standard ? DO_SUBSET_NO : DO_SUBSET_UNKNOWN);
241
20.6k
        }
242
        /* If we find an excessively large type 1 font we won't be able to emit
243
         * a complete copy. Instead we will emit multiple subsets. Detect that here
244
         * and only reserve enough space in the font copy for the maximum subset
245
         * glyphs, 257.
246
         * This also prevents us making a 'complete' copy of the font below. NB the
247
         * value 2048 is merely a guess, intended to prevent copying very large fonts.
248
         */
249
20.6k
        if(pbfont->num_glyphs > 2048 && !is_standard) {
250
0
            reserve_glyphs = 257;
251
0
            if(pbfont->do_subset != DO_SUBSET_NO){
252
0
                char buf[gs_font_name_max + 1];
253
0
                int l = min(font->font_name.size, sizeof(buf) - 1);
254
255
0
                memcpy(buf, font->font_name.chars, l);
256
0
                buf[l] = 0;
257
0
                emprintf1(pdev->memory,
258
0
                          "Can't embed the complete font %s as it is too large, embedding a subset.\n",
259
0
                          buf);
260
0
            }
261
0
        }
262
20.6k
        break;
263
19.8k
    case ft_TrueType:
264
19.8k
        pbfont->num_glyphs = ((gs_font_type42 *)font)->data.trueNumGlyphs;
265
19.8k
        pbfont->do_subset =
266
19.8k
            (pbfont->num_glyphs <= MAX_NO_SUBSET_GLYPHS ?
267
15.3k
             DO_SUBSET_UNKNOWN : DO_SUBSET_YES);
268
19.8k
        break;
269
26
    case ft_CID_encrypted:
270
26
        pbfont->num_glyphs = ((gs_font_cid0 *)font)->cidata.common.CIDCount;
271
26
        goto cid;
272
338
    case ft_CID_TrueType:
273
338
        pbfont->num_glyphs = ((gs_font_cid2 *)font)->cidata.common.CIDCount;
274
364
    cid:
275
364
        pbfont->do_subset = DO_SUBSET_YES;
276
364
        pbfont->CIDSet =
277
364
            gs_alloc_bytes(mem, (pbfont->num_glyphs + 7) / 8,
278
364
                           "pdf_base_font_alloc(CIDSet)");
279
364
        if (pbfont->CIDSet == 0) {
280
0
            code = gs_note_error(gs_error_VMerror);
281
0
            goto fail;
282
0
        }
283
364
        pbfont->CIDSetLength = (pbfont->num_glyphs + 7) / 8;
284
364
        memset(pbfont->CIDSet, 0, (pbfont->num_glyphs + 7) / 8);
285
364
        break;
286
0
    default:
287
0
        code = gs_note_error(gs_error_rangecheck);
288
0
        goto fail;
289
40.8k
    }
290
291
40.8k
    code = gs_copy_font((gs_font *)font, orig_matrix, mem, &copied, reserve_glyphs);
292
40.8k
    if (code < 0)
293
16.2k
        goto fail;
294
24.6k
    gs_notify_register(&copied->notify_list, copied_font_notify, copied);
295
24.6k
    {
296
        /*
297
         * Adobe Technical Note # 5012 "The Type 42 Font Format Specification" says :
298
         *
299
         * There is a known bug in the TrueType rasterizer included in versions of the
300
         * PostScript interpreter previous to version 2013. The problem is that the
301
         * translation components of the FontMatrix, as used as an argument to the
302
         * definefont or makefont operators, are ignored. Translation of user space is
303
         * not affected by this bug.
304
         *
305
         * Besides that, we found that Adobe Acrobat Reader 4 and 5 ignore
306
         * FontMatrix.ty .
307
         */
308
24.6k
        copied->FontMatrix.tx = copied->FontMatrix.ty = 0;
309
24.6k
    }
310
311
24.6k
    if (pbfont->do_subset != DO_SUBSET_YES && reserve_glyphs == -1) {
312
        /* The only possibly non-subsetted fonts are Type 1/2 and Type 42. */
313
24.1k
        if (is_standard)
314
3.16k
            complete = copied;
315
21.0k
        else {
316
21.0k
            code = gs_copy_font((gs_font *)font, &font->FontMatrix, mem, &complete, -1);
317
21.0k
            if (code < 0) {
318
0
                gs_free_copied_font(copied);
319
0
                goto fail;
320
0
            }
321
21.0k
        }
322
24.1k
        code = gs_copy_font_complete((gs_font *)font, complete);
323
24.1k
        if (code < 0 && pbfont->do_subset == DO_SUBSET_NO) {
324
0
            char buf[gs_font_name_max + 1];
325
0
            int l = min(copied->font_name.size, sizeof(buf) - 1);
326
327
0
            memcpy(buf, copied->font_name.chars, l);
328
0
            buf[l] = 0;
329
0
            emprintf1(pdev->memory,
330
0
                      "Can't embed the complete font %s due to font error.\n",
331
0
                      buf);
332
0
            gs_free_copied_font(copied);
333
0
            copied = NULL;
334
0
            goto fail;
335
0
        }
336
24.1k
        if (code < 0) {
337
            /* A font error happened, but it may be caused by a glyph,
338
               which is not used in the document. Continue with subsetting the font.
339
               If the failed glyph will be used in the document,
340
               another error will hgappen when the glyph is used.
341
             */
342
2.31k
            gs_free_copied_font(complete);
343
2.31k
            complete = copied;
344
2.31k
        }
345
24.1k
    } else
346
499
        complete = copied;
347
24.6k
    pbfont->copied = (gs_font_base *)copied;
348
24.6k
    pbfont->complete = (gs_font_base *)complete;
349
350
    /* replace the font cache of the copied fonts with our own font cache
351
     * this is required for PCL, see 'pdf_free_pdf_font_cache' in gdevpdf.c
352
     * for further details.
353
     */
354
24.6k
    pdev->pdf_font_dir->global_glyph_code = font->dir->global_glyph_code;
355
356
24.6k
    pbfont->copied->dir = pbfont->complete->dir = pdev->pdf_font_dir;
357
358
24.6k
    if (pbfont->copied->FontType == ft_CID_encrypted) {
359
26
        gs_font_cid0 *copied0 = (gs_font_cid0 *)pbfont->copied;
360
26
        int i;
361
87
        for (i = 0; i < copied0->cidata.FDArray_size; ++i) {
362
61
            ((gs_font *)copied0->cidata.FDArray[i])->dir = pdev->pdf_font_dir;
363
61
        }
364
26
    }
365
366
24.6k
    pbfont->is_standard = is_standard;
367
24.6k
    if (pfname->size > 0) {
368
23.8k
        font_name.data = pfname->chars;
369
23.8k
        font_name.size = pfname->size;
370
32.7k
        while (pdf_has_subset_prefix(font_name.data, font_name.size)) {
371
            /* Strip off an existing subset prefix. */
372
8.94k
            font_name.data += SUBSET_PREFIX_SIZE;
373
8.94k
            font_name.size -= SUBSET_PREFIX_SIZE;
374
8.94k
        }
375
23.8k
    } else {
376
826
        gs_snprintf(fnbuf, sizeof(fnbuf), "Anonymous");
377
826
        font_name.data = (byte *)fnbuf;
378
826
        font_name.size = strlen(fnbuf);
379
826
    }
380
24.6k
    pbfont->font_name.data =
381
24.6k
        gs_alloc_string(mem, font_name.size, "pdf_base_font_alloc(font_name)");
382
24.6k
    if (pbfont->font_name.data == 0)
383
0
        goto fail;
384
24.6k
    memcpy(pbfont->font_name.data, font_name.data, font_name.size);
385
24.6k
    pbfont->font_name.size = font_name.size;
386
24.6k
    *ppbfont = pbfont;
387
24.6k
    return 0;
388
16.2k
 fail:
389
16.2k
    pdf_base_font_free(pdev, pbfont);
390
16.2k
    return code;
391
24.6k
}
392
393
/*
394
 * Return a reference to the name of a base font.  This name is guaranteed
395
 * not to have a XXXXXX+ prefix.  The client may change the name at will,
396
 * but must not add a XXXXXX+ prefix.
397
 */
398
gs_string *
399
pdf_base_font_name(pdf_base_font_t *pbfont)
400
40.7k
{
401
40.7k
    return &pbfont->font_name;
402
40.7k
}
403
404
/*
405
 * Return the (copied, subset) font associated with a base font.
406
 * This procedure probably shouldn't exist....
407
 */
408
gs_font_base *
409
pdf_base_font_font(const pdf_base_font_t *pbfont, bool complete)
410
36.9M
{
411
36.9M
    return (complete ? pbfont->complete : pbfont->copied);
412
36.9M
}
413
414
/*
415
 * Check for subset font.
416
 */
417
bool
418
pdf_base_font_is_subset(const pdf_base_font_t *pbfont)
419
18.8k
{
420
18.8k
    return pbfont->do_subset == DO_SUBSET_YES;
421
18.8k
}
422
423
/*
424
 * Drop the copied complete font associated with a base font.
425
 */
426
void
427
pdf_base_font_drop_complete(pdf_base_font_t *pbfont)
428
673
{
429
    /* gs_font is a subset of gs_font_base and we only want to
430
     * free the members which are common to both, so this cast is
431
     * (at the time of writing) safe.
432
     */
433
673
    if (pbfont->complete != pbfont->copied)
434
673
        gs_free_copied_font((gs_font *)pbfont->complete);
435
673
    pbfont->complete = NULL;
436
673
}
437
438
/*
439
 * Copy a glyph (presumably one that was just used) into a saved base
440
 * font.  Note that it is the client's responsibility to determine that
441
 * the source font is compatible with the target font.  (Normally they
442
 * will be the same.)
443
 */
444
int
445
pdf_base_font_copy_glyph(pdf_base_font_t *pbfont, gs_glyph glyph,
446
                         gs_font_base *font)
447
610k
{
448
610k
    int code;
449
450
    /* If we're a TrueType CIDFont, check teh GSUB table for replacement glyphs.
451
     * Bug #691574
452
     */
453
610k
    if (font->FontType == ft_CID_TrueType) {
454
71.1k
        code =
455
71.1k
            gs_copy_glyph_options((gs_font *)font, glyph,
456
71.1k
                              (gs_font *)pbfont->copied,
457
71.1k
                              (pbfont->is_standard ? COPY_GLYPH_NO_NEW : COPY_GLYPH_USE_GSUB));
458
539k
    } else {
459
539k
        code =
460
539k
            gs_copy_glyph_options((gs_font *)font, glyph,
461
539k
                              (gs_font *)pbfont->copied,
462
539k
                              (pbfont->is_standard ? COPY_GLYPH_NO_NEW : 0));
463
539k
    }
464
610k
    if (code < 0)
465
17.8k
        return code;
466
593k
    if (pbfont->CIDSet != 0 &&
467
70.2k
        (uint)(glyph - GS_MIN_CID_GLYPH) < pbfont->num_glyphs
468
593k
        ) {
469
70.2k
        uint cid = glyph - GS_MIN_CID_GLYPH;
470
471
70.2k
        pbfont->CIDSet[cid >> 3] |= 0x80 >> (cid & 7);
472
70.2k
    }
473
593k
    return 0;
474
610k
}
475
476
/*
477
 * Determine whether a font should be subsetted.
478
 */
479
bool
480
pdf_do_subset_font(gx_device_pdf *pdev, pdf_base_font_t *pbfont, gs_id rid)
481
58.6k
{
482
58.6k
    gs_font_base *copied = pbfont->copied;
483
484
    /*
485
     * If the decision has not been made already, determine whether or not
486
     * to subset the font.
487
     */
488
58.6k
    if (pbfont->do_subset == DO_SUBSET_UNKNOWN) {
489
21.0k
        int max_pct = pdev->params.MaxSubsetPct;
490
21.0k
        bool do_subset = pdev->params.SubsetFonts && max_pct > 0;
491
492
21.0k
        if (do_subset && max_pct < 100) {
493
            /* We want to subset iff used <= total * MaxSubsetPct / 100. */
494
0
            do_subset = false;
495
0
            if (max_pct > 0) {
496
0
                int max_subset_used = pbfont->num_glyphs * max_pct / 100;
497
0
                int used, index;
498
0
                gs_glyph ignore_glyph;
499
500
0
                do_subset = true;
501
0
                for (index = 0, used = 0;
502
0
                     (copied->procs.enumerate_glyph((gs_font *)copied,
503
0
                                                &index, GLYPH_SPACE_INDEX,
504
0
                                                &ignore_glyph), index != 0);
505
0
                     )
506
0
                    if (++used > max_subset_used) {
507
0
                        do_subset = false;
508
0
                        break;
509
0
                    }
510
0
            }
511
0
        }
512
21.0k
        pbfont->do_subset = (do_subset ? DO_SUBSET_YES : DO_SUBSET_NO);
513
21.0k
    }
514
58.6k
    return (pbfont->do_subset == DO_SUBSET_YES);
515
58.6k
}
516
517
/*
518
 * Write the FontFile entry for an embedded font, /FontFile<n> # # R.
519
 */
520
int
521
pdf_write_FontFile_entry(gx_device_pdf *pdev, pdf_base_font_t *pbfont)
522
18.5k
{
523
18.5k
    stream *s = pdev->strm;
524
18.5k
    const char *FontFile_key;
525
526
18.5k
    switch (pbfont->copied->FontType) {
527
2.41k
    case ft_TrueType:
528
2.71k
    case ft_CID_TrueType:
529
2.71k
        FontFile_key = "/FontFile2";
530
2.71k
        break;
531
15.7k
    default:      /* Type 1/2, CIDFontType 0 */
532
15.7k
        if (!pdev->HaveCFF)
533
9.74k
            FontFile_key = "/FontFile";
534
6.03k
        else
535
6.03k
            FontFile_key = "/FontFile3";
536
18.5k
    }
537
18.5k
    stream_puts(s, FontFile_key);
538
18.5k
    pprinti64d1(s, " %"PRId64" 0 R", pbfont->FontFile->id);
539
18.5k
    return 0;
540
18.5k
}
541
542
/*
543
 * Adjust font name for Acrobat Reader 3.
544
 */
545
static int
546
pdf_adjust_font_name(gx_device_pdf *pdev, int64_t id, pdf_base_font_t *pbfont)
547
0
{
548
    /*
549
     * In contradiction with previous version of pdfwrite,
550
     * this always adds a suffix. We don't check the original name
551
     * for uniquity bacause the layered architecture
552
     * (see gdevpdtx.h) doesn't provide an easy access for
553
     * related information.
554
     */
555
0
    int i;
556
0
    byte *chars = (byte *)pbfont->font_name.data; /* break 'const' */
557
0
    byte *data;
558
0
    uint size = pbfont->font_name.size;
559
0
    char suffix[sizeof(int64_t) * 2 + 2];
560
0
    uint suffix_size;
561
562
0
#define SUFFIX_CHAR '~'
563
    /*
564
     * If the name looks as though it has one of our unique suffixes,
565
     * remove the suffix.
566
     */
567
0
    for (i = size;
568
0
         i > 0 && isxdigit(chars[i - 1]);
569
0
         --i)
570
0
        DO_NOTHING;
571
0
    if (i < size && i > 0 && chars[i - 1] == SUFFIX_CHAR) {
572
0
        do {
573
0
            --i;
574
0
        } while (i > 0 && chars[i - 1] == SUFFIX_CHAR);
575
0
        size = i + 1;
576
0
    }
577
    /* Create a unique name. */
578
0
    gs_snprintf(suffix, sizeof(suffix), "%c%lx", SUFFIX_CHAR, id);
579
0
    suffix_size = strlen(suffix);
580
0
    data = gs_resize_string(pdev->pdf_memory, chars, size,
581
0
                                  size + suffix_size,
582
0
                                  "pdf_adjust_font_name");
583
0
    if (data == 0)
584
0
        return_error(gs_error_VMerror);
585
0
    memcpy(data + size, (const byte *)suffix, suffix_size);
586
0
    pbfont->font_name.data = data;
587
0
    pbfont->font_name.size = size + suffix_size;
588
0
#undef SUFFIX_CHAR
589
0
    return 0;
590
0
}
591
592
/*
593
 * Write an embedded font.
594
 */
595
int
596
pdf_write_embedded_font(gx_device_pdf *pdev, pdf_base_font_t *pbfont, font_type FontType,
597
                        gs_int_rect *FontBBox, gs_id rid, cos_dict_t **ppcd)
598
18.5k
{
599
18.5k
    bool do_subset = pdf_do_subset_font(pdev, pbfont, rid);
600
18.5k
    gs_font_base *out_font =
601
18.5k
        (do_subset || pbfont->complete == NULL ? pbfont->copied : pbfont->complete);
602
18.5k
    gs_const_string fnstr;
603
18.5k
    pdf_data_writer_t writer;
604
18.5k
    byte digest[6] = {0,0,0,0,0,0};
605
18.5k
    int code = 0;
606
18.5k
    int options=0;
607
608
18.5k
    if (pbfont->written)
609
0
        return 0;   /* already written */
610
18.5k
    code = copied_order_font((gs_font *)out_font);
611
18.5k
    if (code < 0)
612
0
        return code;
613
    /* Since we now always ASCIIHex encode the eexec encrypted portion of a
614
     * Type 1 font, such a font cannot contain any binary data, if its not being
615
     * compressed then there is no reason to ASCII encode it (which will happen
616
     * we set DATA_STREAM_BINARY and the device does not permit binary output)
617
     * NB if HaveCFF is true then we convert type 1 to CFF which is a binary
618
     * format, so we still need to set DATA_STREAM_BINARY.
619
     */
620
18.5k
    if (pdev->CompressFonts)
621
18.5k
        options = DATA_STREAM_BINARY | DATA_STREAM_COMPRESS;
622
0
    else
623
0
         if (FontType != ft_encrypted || pdev->HaveCFF)
624
0
            options = DATA_STREAM_BINARY;
625
    /* Don't set DATA_STREAM_ENCRYPT since we write to a temporary file.
626
     * See comment in pdf_begin_encrypt.
627
     */
628
18.5k
    code = pdf_begin_data_stream(pdev, &writer, options, 0);
629
18.5k
    if (code < 0)
630
0
        return code;
631
18.5k
    if (pdev->PDFA != 0) {
632
0
        stream *s = s_MD5C_make_stream(pdev->pdf_memory, writer.binary.strm);
633
634
0
        if (s == NULL)
635
0
            return_error(gs_error_VMerror);
636
0
        writer.binary.strm = s;
637
0
    }
638
18.5k
    if (pdev->CompatibilityLevel == 1.2 &&
639
11.3k
        !do_subset && !pbfont->is_standard ) {
640
        /*
641
         * Due to a bug in Acrobat Reader 3, we need to generate
642
         * unique font names, except base 14 fonts being not embedded.
643
         * To recognize base 14 fonts here we used the knowledge
644
         * that pbfont->is_standard is true for base 14 fonts only.
645
         * Note that subsetted fonts already have an unique name
646
         * due to subset prefix.
647
         */
648
0
         int code = pdf_adjust_font_name(pdev, writer.pres->object->id, pbfont);
649
0
         if (code < 0)
650
0
            return code;
651
0
    }
652
18.5k
    fnstr.data = pbfont->font_name.data;
653
18.5k
    fnstr.size = pbfont->font_name.size;
654
    /* Now write the font (or subset). */
655
18.5k
    switch (FontType) {
656
657
0
    case ft_composite:
658
        /* Nothing to embed -- the descendant fonts do it all. */
659
0
        code = 0;
660
0
        break;
661
662
280
    case ft_encrypted2:
663
280
        if (!pdev->HaveCFF) {
664
            /* Must convert to Type 1 charstrings. */
665
0
            return_error(gs_error_unregistered); /* Not implemented yet. */
666
0
        }
667
15.8k
    case ft_encrypted:
668
15.8k
        if (pdev->HavePDFWidths) {
669
6.06k
            code = copied_drop_extension_glyphs((gs_font *)out_font);
670
6.06k
            if (code < 0)
671
0
                return code;
672
6.06k
        }
673
15.8k
        if (!pdev->HaveCFF) {
674
            /* Write the type 1 font with no converting to CFF. */
675
9.74k
            int lengths[3];
676
677
9.74k
            if (pbfont->do_subset != DO_SUBSET_NO)
678
9.74k
                code = psf_write_type1_font(writer.binary.strm,
679
9.74k
                                (gs_font_type1 *)out_font,
680
9.74k
                                WRITE_TYPE1_WITH_LENIV | WRITE_TYPE1_EEXEC |
681
9.74k
                                WRITE_TYPE1_EEXEC_PAD | WRITE_TYPE1_ASCIIHEX,
682
9.74k
                                NULL, 0, &fnstr, lengths);
683
0
            else
684
0
                code = psf_write_type1_font(writer.binary.strm,
685
0
                                (gs_font_type1 *)out_font,
686
0
                                WRITE_TYPE1_WITH_LENIV | WRITE_TYPE1_EEXEC |
687
0
                                WRITE_TYPE1_EEXEC_PAD | WRITE_TYPE1_ASCIIHEX | WRITE_TYPE1_XUID,
688
0
                                NULL, 0, &fnstr, lengths);
689
9.74k
            if (lengths[0] > 0) {
690
9.74k
                if (code < 0)
691
0
                    goto finish;
692
9.74k
                code = cos_dict_put_c_key_int((cos_dict_t *)writer.pres->object,
693
9.74k
                            "/Length1", lengths[0]);
694
9.74k
            }
695
9.74k
            if (lengths[1] > 0) {
696
9.74k
                if (code < 0)
697
0
                    goto finish;
698
9.74k
                code = cos_dict_put_c_key_int((cos_dict_t *)writer.pres->object,
699
9.74k
                            "/Length2", lengths[1]);
700
9.74k
                if (code < 0)
701
0
                    return code;
702
9.74k
                code = cos_dict_put_c_key_int((cos_dict_t *)writer.pres->object,
703
9.74k
                            "/Length3", lengths[2]);
704
9.74k
            }
705
9.74k
        } else {
706
            /*
707
             * Since we only support PDF 1.2 and later, always write Type 1
708
             * fonts as Type1C (Type 2).  Acrobat Reader apparently doesn't
709
             * accept CFF fonts with Type 1 CharStrings, so we need to convert
710
             * them.  Also remove lenIV, so Type 2 fonts will compress better.
711
             */
712
6.08k
#define TYPE2_OPTIONS (WRITE_TYPE2_NO_LENIV | WRITE_TYPE2_CHARSTRINGS)
713
6.06k
            code = cos_dict_put_string_copy((cos_dict_t *)writer.pres->object, "/Subtype", "/Type1C");
714
6.06k
            if (code < 0)
715
0
                return code;
716
6.06k
            code = psf_write_type2_font(writer.binary.strm,
717
6.06k
                                        (gs_font_type1 *)out_font,
718
6.06k
                                        TYPE2_OPTIONS |
719
6.06k
                            (pdev->CompatibilityLevel < 1.3 ? WRITE_TYPE2_AR3 : 0) |
720
6.06k
                            (pbfont->do_subset == DO_SUBSET_NO ? WRITE_TYPE2_XUID : 0),
721
6.06k
                                        NULL, 0, &fnstr, FontBBox);
722
6.06k
        }
723
15.8k
        goto finish;
724
725
15.8k
    case ft_TrueType: {
726
2.41k
        gs_font_type42 *const pfont = (gs_font_type42 *)out_font;
727
2.41k
#define TRUETYPE_OPTIONS (WRITE_TRUETYPE_NAME | WRITE_TRUETYPE_HVMTX)
728
        /* Acrobat Reader 3 doesn't handle cmap format 6 correctly. */
729
2.41k
        const int options = TRUETYPE_OPTIONS |
730
2.41k
            (pdev->PDFA != 0 ? WRITE_TRUETYPE_UNICODE_CMAP : 0) |
731
2.41k
            (pdev->CompatibilityLevel <= 1.2 ?
732
1.62k
             WRITE_TRUETYPE_NO_TRIMMED_TABLE : 0) |
733
            /* Generate a cmap only for incrementally downloaded fonts
734
               and for subsetted fonts. */
735
2.41k
            (pfont->data.numGlyphs != pfont->data.trueNumGlyphs ||
736
2.20k
             pbfont->do_subset == DO_SUBSET_YES ?
737
2.41k
             WRITE_TRUETYPE_CMAP : 0);
738
2.41k
        stream poss;
739
740
2.41k
        if (pdev->HavePDFWidths) {
741
786
            code = copied_drop_extension_glyphs((gs_font *)out_font);
742
786
            if (code < 0)
743
0
                return code;
744
786
        }
745
2.41k
        s_init(&poss, pdev->memory);
746
2.41k
        swrite_position_only(&poss);
747
2.41k
        code = psf_write_truetype_font(&poss, pfont, options, NULL, 0, &fnstr);
748
2.41k
        if (code < 0)
749
0
            goto finish;
750
2.41k
        code = cos_dict_put_c_key_int((cos_dict_t *)writer.pres->object, "/Length1", stell(&poss));
751
2.41k
        if (code < 0)
752
0
            goto finish;
753
2.41k
        code = psf_write_truetype_font(writer.binary.strm, pfont,
754
2.41k
                                       options, NULL, 0, &fnstr);
755
2.41k
        goto finish;
756
2.41k
    }
757
758
15
    case ft_CID_encrypted:
759
15
        code = cos_dict_put_string_copy((cos_dict_t *)writer.pres->object, "/Subtype", "/CIDFontType0C");
760
15
        if (code < 0)
761
0
            return code;
762
15
        code = psf_write_cid0_font(writer.binary.strm,
763
15
                                   (gs_font_cid0 *)out_font, TYPE2_OPTIONS,
764
15
                                   NULL, 0, &fnstr);
765
15
        goto finish;
766
767
305
    case ft_CID_TrueType:
768
        /* CIDFontType 2 fonts don't use cmap, name, OS/2, or post. */
769
305
#define CID2_OPTIONS WRITE_TRUETYPE_HVMTX
770
305
        code = psf_write_cid2_font(writer.binary.strm,
771
305
                                   (gs_font_cid2 *)out_font,
772
305
                                   CID2_OPTIONS, NULL, 0, &fnstr);
773
18.5k
    finish:
774
18.5k
        if (pdev->PDFA != 0) {
775
0
            sflush(writer.binary.strm);
776
0
            s_MD5C_get_digest(writer.binary.strm, digest, sizeof(digest));
777
0
        }
778
18.5k
        *ppcd = (cos_dict_t *)writer.pres->object;
779
18.5k
        if (code < 0) {
780
41
            pdf_end_fontfile(pdev, &writer);
781
41
            pdf_obj_mark_unused(pdev, writer.pres->object->id);
782
41
            return code;
783
41
        }
784
18.5k
        code = pdf_end_fontfile(pdev, &writer);
785
18.5k
        break;
786
787
0
    default:
788
0
        code = gs_note_error(gs_error_rangecheck);
789
18.5k
    }
790
791
18.5k
    pbfont->written = true;
792
18.5k
    return code;
793
18.5k
}
794
795
/*
796
 * Write the CharSet for a subsetted font, as a PDF string.
797
 */
798
int
799
pdf_write_CharSet(gx_device_pdf *pdev, pdf_base_font_t *pbfont)
800
15.8k
{
801
15.8k
    stream *s = pdev->strm;
802
15.8k
    gs_font_base *font = pbfont->copied;
803
15.8k
    int index;
804
15.8k
    gs_glyph glyph;
805
806
15.8k
    stream_puts(s, "(");
807
15.8k
    for (index = 0;
808
315k
         (font->procs.enumerate_glyph((gs_font *)font, &index,
809
315k
                                      GLYPH_SPACE_NAME, &glyph),
810
315k
          index != 0);
811
299k
         ) {
812
299k
        gs_const_string gstr;
813
299k
        int code = font->procs.glyph_name((gs_font *)font, glyph, &gstr);
814
815
        /* Don't include .notdef. */
816
299k
        if (code >= 0 &&
817
299k
            bytes_compare(gstr.data, gstr.size, (const byte *)".notdef", 7)
818
299k
            )
819
283k
            pdf_put_name(pdev, gstr.data, gstr.size);
820
299k
    }
821
15.8k
    stream_puts(s, ")");
822
15.8k
    return 0;
823
15.8k
}
824
825
/*
826
 * Write the CIDSet object for a subsetted CIDFont.
827
 */
828
int
829
pdf_write_CIDSet(gx_device_pdf *pdev, pdf_base_font_t *pbfont,
830
                 int64_t *pcidset_id)
831
320
{
832
320
    pdf_data_writer_t writer;
833
320
    int code;
834
835
320
    code = pdf_begin_data_stream(pdev, &writer,
836
320
                      DATA_STREAM_BINARY |
837
320
                      (pdev->CompressFonts ? DATA_STREAM_COMPRESS : 0),
838
320
                      gs_no_id);
839
320
    if (code < 0)
840
0
        return code;
841
320
    stream_write(writer.binary.strm, pbfont->CIDSet,
842
320
                 pbfont->CIDSetLength);
843
320
    code = pdf_end_data(&writer);
844
320
    if (code < 0)
845
0
        return code;
846
320
    *pcidset_id = pdf_resource_id(writer.pres);
847
320
    return 0;
848
320
}
849
/*
850
 * Check whether a base font is standard.
851
 */
852
bool
853
pdf_is_standard_font(pdf_base_font_t *bfont)
854
0
{   return bfont->is_standard;
855
0
}
856
857
void
858
pdf_set_FontFile_object(pdf_base_font_t *bfont, cos_dict_t *pcd)
859
18.5k
{
860
18.5k
    bfont->FontFile = pcd;
861
18.5k
}
862
const cos_dict_t *
863
pdf_get_FontFile_object(pdf_base_font_t *bfont)
864
18.5k
{
865
18.5k
    return bfont->FontFile;
866
18.5k
}