| 1 | #include "mupdf/fitz.h" |
| 2 | #include "mupdf/pdf.h" |
| 3 | |
| 4 | #include <string.h> |
| 5 | |
| 6 | /* Load or synthesize ToUnicode map for fonts */ |
| 7 | |
| 8 | static void |
| 9 | pdf_remap_cmap_range(fz_context *ctx, pdf_cmap *ucs_from_gid, |
| 10 | unsigned int cpt, unsigned int gid, unsigned int n, pdf_cmap *ucs_from_cpt) |
| 11 | { |
| 12 | unsigned int k; |
| 13 | int ucsbuf[8]; |
| 14 | int ucslen; |
| 15 | |
| 16 | for (k = 0; k <= n; ++k) |
| 17 | { |
| 18 | ucslen = pdf_lookup_cmap_full(ucs_from_cpt, cpt + k, ucsbuf); |
| 19 | if (ucslen == 1) |
| 20 | pdf_map_range_to_range(ctx, ucs_from_gid, gid + k, gid + k, ucsbuf[0]); |
| 21 | else if (ucslen > 1) |
| 22 | pdf_map_one_to_many(ctx, ucs_from_gid, gid + k, ucsbuf, ucslen); |
| 23 | } |
| 24 | } |
| 25 | |
| 26 | static pdf_cmap * |
| 27 | pdf_remap_cmap(fz_context *ctx, pdf_cmap *gid_from_cpt, pdf_cmap *ucs_from_cpt) |
| 28 | { |
| 29 | pdf_cmap *ucs_from_gid; |
| 30 | unsigned int a, b, x; |
| 31 | int i; |
| 32 | |
| 33 | ucs_from_gid = pdf_new_cmap(ctx); |
| 34 | |
| 35 | fz_try(ctx) |
| 36 | { |
| 37 | if (gid_from_cpt->usecmap) |
| 38 | ucs_from_gid->usecmap = pdf_remap_cmap(ctx, gid_from_cpt->usecmap, ucs_from_cpt); |
| 39 | |
| 40 | for (i = 0; i < gid_from_cpt->codespace_len; i++) |
| 41 | { |
| 42 | pdf_add_codespace(ctx, ucs_from_gid, |
| 43 | gid_from_cpt->codespace[i].low, |
| 44 | gid_from_cpt->codespace[i].high, |
| 45 | gid_from_cpt->codespace[i].n); |
| 46 | } |
| 47 | |
| 48 | for (i = 0; i < gid_from_cpt->rlen; ++i) |
| 49 | { |
| 50 | a = gid_from_cpt->ranges[i].low; |
| 51 | b = gid_from_cpt->ranges[i].high; |
| 52 | x = gid_from_cpt->ranges[i].out; |
| 53 | pdf_remap_cmap_range(ctx, ucs_from_gid, a, x, b - a, ucs_from_cpt); |
| 54 | } |
| 55 | |
| 56 | for (i = 0; i < gid_from_cpt->xlen; ++i) |
| 57 | { |
| 58 | a = gid_from_cpt->xranges[i].low; |
| 59 | b = gid_from_cpt->xranges[i].high; |
| 60 | x = gid_from_cpt->xranges[i].out; |
| 61 | pdf_remap_cmap_range(ctx, ucs_from_gid, a, x, b - a, ucs_from_cpt); |
| 62 | } |
| 63 | |
| 64 | /* Font encoding CMaps don't have one-to-many mappings, so we can ignore the mranges. */ |
| 65 | |
| 66 | pdf_sort_cmap(ctx, ucs_from_gid); |
| 67 | } |
| 68 | fz_catch(ctx) |
| 69 | { |
| 70 | pdf_drop_cmap(ctx, ucs_from_gid); |
| 71 | fz_rethrow(ctx); |
| 72 | } |
| 73 | |
| 74 | return ucs_from_gid; |
| 75 | } |
| 76 | |
| 77 | void |
| 78 | pdf_load_to_unicode(fz_context *ctx, pdf_document *doc, pdf_font_desc *font, |
| 79 | const char **strings, char *collection, pdf_obj *cmapstm) |
| 80 | { |
| 81 | unsigned int cpt; |
| 82 | |
| 83 | if (pdf_is_stream(ctx, cmapstm)) |
| 84 | { |
| 85 | pdf_cmap *ucs_from_cpt = pdf_load_embedded_cmap(ctx, doc, cmapstm); |
| 86 | fz_try(ctx) |
| 87 | font->to_unicode = pdf_remap_cmap(ctx, font->encoding, ucs_from_cpt); |
| 88 | fz_always(ctx) |
| 89 | pdf_drop_cmap(ctx, ucs_from_cpt); |
| 90 | fz_catch(ctx) |
| 91 | fz_rethrow(ctx); |
| 92 | font->size += pdf_cmap_size(ctx, font->to_unicode); |
| 93 | } |
| 94 | |
| 95 | else if (collection) |
| 96 | { |
| 97 | if (!strcmp(collection, "Adobe-CNS1" )) |
| 98 | font->to_unicode = pdf_load_system_cmap(ctx, "Adobe-CNS1-UCS2" ); |
| 99 | else if (!strcmp(collection, "Adobe-GB1" )) |
| 100 | font->to_unicode = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2" ); |
| 101 | else if (!strcmp(collection, "Adobe-Japan1" )) |
| 102 | font->to_unicode = pdf_load_system_cmap(ctx, "Adobe-Japan1-UCS2" ); |
| 103 | else if (!strcmp(collection, "Adobe-Korea1" )) |
| 104 | font->to_unicode = pdf_load_system_cmap(ctx, "Adobe-Korea1-UCS2" ); |
| 105 | |
| 106 | return; |
| 107 | } |
| 108 | |
| 109 | if (strings) |
| 110 | { |
| 111 | /* TODO one-to-many mappings */ |
| 112 | |
| 113 | font->cid_to_ucs = fz_malloc_array(ctx, 256, unsigned short); |
| 114 | font->cid_to_ucs_len = 256; |
| 115 | font->size += 256 * sizeof *font->cid_to_ucs; |
| 116 | |
| 117 | for (cpt = 0; cpt < 256; cpt++) |
| 118 | { |
| 119 | if (strings[cpt]) |
| 120 | font->cid_to_ucs[cpt] = fz_unicode_from_glyph_name(strings[cpt]); |
| 121 | else |
| 122 | font->cid_to_ucs[cpt] = FZ_REPLACEMENT_CHARACTER; |
| 123 | } |
| 124 | } |
| 125 | |
| 126 | if (!font->to_unicode && !font->cid_to_ucs) |
| 127 | { |
| 128 | /* TODO: synthesize a ToUnicode if it's a freetype font with |
| 129 | * cmap and/or post tables or if it has glyph names. */ |
| 130 | } |
| 131 | } |
| 132 | |