1 /* Copyright (C) 2001-2006 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, modified 8 or distributed except as expressly authorized under the terms of that 9 license. Refer to licensing information at http://www.artifex.com/ 10 or contact Artifex Software, Inc., 7 Mt. Lassen Drive - Suite A-134, 11 San Rafael, CA 94903, U.S.A., +1(415)492-9861, for further information. 12 */ 13 14 /* $Id: gdevpdtt.h 8265 2007-10-02 07:31:58Z ken $ */ 15 /* Internal text processing interface for pdfwrite */ 16 17 #ifndef gdevpdtt_INCLUDED 18 # define gdevpdtt_INCLUDED 19 20 /* 21 * This file is only used internally to define the interface between 22 * gdevpdtt.c, gdevpdtc.c, and gdevpdte.c. 23 */ 24 25 /* ---------------- Coordinate systems ---------------- */ 26 27 /* 28 29 The current text code has to deal with 6 different coordinate systems. 30 This situation is complex, confusing, and fragile, but despite literally 31 years of struggle we have been unable to understand how to simplify it. 32 33 1) PostScript user space at the time a text operator is invoked. The 34 width values for ashow, xshow, etc. are specified in this space: these 35 values appear in the delta_all, delta_space, x_widths, and y_widths 36 members of the text member of the common part of the text enumerator 37 (pdf_text_enum_t). 38 39 2) Device space. For the pdfwrite device, this is a straightforward space 40 with (0,0) in the lower left corner. The usual resolution is 720 dpi, 41 but this is user-settable to provide a tradeoff between file size and 42 bitmap resolution (for rendered Patterns and fonts that have to be 43 converted to bitmaps). 44 45 3) PDF user space. During the processing of text operators, this is 46 always the same as the default PDF user space, in which 1 unit = 1/72", 47 because an Acrobat quirk apparently requires this. (See stream_to_text 48 in gdevpdfu.c.) 49 50 4) Font design space. This is the space in which the font's character 51 outlines and advance widths are specified, including the width values 52 returned by font->procs.glyph_info and by pdf_char_widths. 53 54 5) PDF unscaled text space. This space is used for the PDF width 55 adjustment values (Tc, Tw, and TL parameters, and " operator) and 56 positioning coordinates (Td and TD operators). 57 58 6) PDF text space. This space is used for the width adjustments for the 59 TJ operator. 60 61 The following convert between these spaces: 62 63 - The PostScript CTM (pte->pis->ctm) maps #1 to #2. 64 65 - The mapping from #3 to #2 is a scaling by pdev->HWResolution / 72. 66 67 - The mapping from #5 to #6 is a scaling by the size member of 68 pdf_text_state_values_t, which is the size value for the Tf operator. 69 70 - The matrix member of pdf_text_state_values_t maps #5 to #2, which is the 71 matrix for the Tm operator or its abbreviations. 72 73 - The FontMatrix of a font maps #4 to #1. 74 75 Note that the PDF text matrix (set by the Tm operator) maps #5 to #3. 76 However, this is not actually stored anywhere: it is computed by 77 multiplying the #5->#2 matrix by the #2->#3 scaling. 78 79 */ 80 81 /* ---------------- Types and structures ---------------- */ 82 83 #ifndef pdf_char_glyph_pair_DEFINED 84 # define pdf_char_glyph_pair_DEFINED 85 typedef struct pdf_char_glyph_pair_s pdf_char_glyph_pair_t; 86 #endif 87 88 #ifndef pdf_char_glyph_pairs_DEFINED 89 # define pdf_char_glyph_pairs_DEFINED 90 typedef struct pdf_char_glyph_pairs_s pdf_char_glyph_pairs_t; 91 #endif 92 93 /* Define a structure for a text characters list. */ 94 /* It must not contain pointers due to variable length. */ 95 struct pdf_char_glyph_pairs_s { 96 int num_all_chars; 97 int num_unused_chars; 98 int unused_offset; /* The origin of the unused character table.*/ 99 pdf_char_glyph_pair_t s[1]; /* Variable length. */ 100 }; 101 102 /* Define the text enumerator. */ 103 typedef struct pdf_text_enum_s { 104 gs_text_enum_common; 105 gs_text_enum_t *pte_default; 106 gs_fixed_point origin; 107 bool charproc_accum; 108 bool cdevproc_callout; 109 double cdevproc_result[10]; 110 pdf_char_glyph_pairs_t *cgp; 111 gs_char output_char_code; 112 } pdf_text_enum_t; 113 #define private_st_pdf_text_enum()\ 114 extern_st(st_gs_text_enum);\ 115 gs_private_st_suffix_add2(st_pdf_text_enum, pdf_text_enum_t,\ 116 "pdf_text_enum_t", pdf_text_enum_enum_ptrs, pdf_text_enum_reloc_ptrs,\ 117 st_gs_text_enum, pte_default, cgp) 118 119 /* 120 * Define quantities derived from the current font and CTM, used within 121 * the text processing loop. NOTE: This structure has no GC descriptor 122 * and must therefore only be used locally (allocated on the C stack). 123 */ 124 typedef struct pdf_text_process_state_s { 125 pdf_text_state_values_t values; 126 gs_font *font; 127 } pdf_text_process_state_t; 128 129 /* 130 * Define the structure used to return glyph width information. Note that 131 * there are two different sets of width information: real-number (x,y) 132 * values, which give the true advance width, and an integer value, which 133 * gives an X advance width for WMode = 0 or a Y advance width for WMode = 1. 134 * The return value from pdf_glyph_width() indicates which of these is/are 135 * valid. 136 */ 137 typedef struct pdf_glyph_width_s { 138 double w; 139 gs_point xy; 140 gs_point v; /* glyph origin shift */ 141 } pdf_glyph_width_t; 142 typedef struct pdf_glyph_widths_s { 143 pdf_glyph_width_t Width; /* unmodified, for Widths */ 144 pdf_glyph_width_t real_width; /* possibly modified, for rendering */ 145 bool replaced_v; 146 } pdf_glyph_widths_t; 147 148 /* ---------------- Procedures ---------------- */ 149 150 /* 151 * Declare the procedures for processing different species of text. 152 * These procedures may, but need not, copy pte->text into buf 153 * (a caller-supplied buffer large enough to hold the string). 154 */ 155 #define PROCESS_TEXT_PROC(proc)\ 156 int proc(gs_text_enum_t *pte, void *vbuf, uint bsize) 157 158 /* ------ gdevpdtt.c ------ */ 159 160 /* 161 * Compute and return the orig_matrix of a font. 162 */ 163 int pdf_font_orig_matrix(const gs_font *font, gs_matrix *pmat); 164 165 /* 166 * Check the Encoding compatibility 167 */ 168 bool pdf_check_encoding_compatibility(const pdf_font_resource_t *pdfont, 169 const pdf_char_glyph_pair_t *pairs, int num_chars); 170 171 /* 172 * Create or find a font resource object for a text. 173 */ 174 int 175 pdf_obtain_font_resource(pdf_text_enum_t *penum, 176 const gs_string *pstr, pdf_font_resource_t **ppdfont); 177 178 /* 179 * Create or find a font resource object for a glyphshow text. 180 */ 181 int pdf_obtain_font_resource_unencoded(pdf_text_enum_t *penum, 182 const gs_string *pstr, pdf_font_resource_t **ppdfont, const gs_glyph *gdata); 183 184 /* 185 * Create or find a CID font resource object for a glyph set. 186 */ 187 int pdf_obtain_cidfont_resource(gx_device_pdf *pdev, gs_font *subfont, 188 pdf_font_resource_t **ppdsubf, 189 pdf_char_glyph_pairs_t *cgp); 190 191 /* 192 * Create or find a parent Type 0 font resource object for a CID font resource. 193 */ 194 int pdf_obtain_parent_type0_font_resource(gx_device_pdf *pdev, pdf_font_resource_t *pdsubf, 195 uint font_index, const gs_const_string *CMapName, pdf_font_resource_t **pdfont); 196 197 /* 198 * Retrive font resource attached to a font. 199 * allocating glyph_usage and real_widths on request. 200 */ 201 int pdf_attached_font_resource(gx_device_pdf *pdev, gs_font *font, 202 pdf_font_resource_t **pdfont, byte **glyph_usage, 203 double **real_widths, int *num_chars, int *num_widths); 204 205 /* 206 * Attach font resource to a font. 207 */ 208 int pdf_attach_font_resource(gx_device_pdf *pdev, gs_font *font, 209 pdf_font_resource_t *pdfont); 210 211 /* 212 * Locate a font cache element. 213 */ 214 pdf_font_cache_elem_t ** 215 pdf_locate_font_cache_elem(gx_device_pdf *pdev, gs_font *font); 216 217 /* 218 * Create a font resource object for a gs_font of Type 3. 219 */ 220 int pdf_make_font3_resource(gx_device_pdf *pdev, gs_font *font, 221 pdf_font_resource_t **ppdfont); 222 223 /* 224 * Compute the cached values in the text processing state from the text 225 * parameters, pdfont, and pis->ctm. Return either an error code (< 0) or a 226 * mask of operation attributes that the caller must emulate. Currently the 227 * only such attributes are TEXT_ADD_TO_ALL_WIDTHS and 228 * TEXT_ADD_TO_SPACE_WIDTH. 229 */ 230 int pdf_update_text_state(pdf_text_process_state_t *ppts, 231 const pdf_text_enum_t *penum, 232 pdf_font_resource_t *pdfont, 233 const gs_matrix *pfmat); 234 235 /* 236 * Set up commands to make the output state match the processing state. 237 * General graphics state commands are written now; text state commands 238 * are written later. 239 */ 240 int pdf_set_text_process_state(gx_device_pdf *pdev, 241 const gs_text_enum_t *pte, /*for pdcolor, pis*/ 242 pdf_text_process_state_t *ppts); 243 244 /* 245 * Get the widths (unmodified and possibly modified) of a glyph in a (base) 246 * font. If the width is cachable (implying that the w values area valid), 247 * return 0; if only the xy values are valid, or the width is not cachable 248 * for some other reason, return 1. 249 * Return TEXT_PROCESS_CDEVPROC if a CDevProc callout is needed. 250 * cdevproc_result != NULL if we restart after a CDevProc callout. 251 */ 252 int pdf_glyph_widths(pdf_font_resource_t *pdfont, int wmode, gs_glyph glyph, 253 gs_font *font, pdf_glyph_widths_t *pwidths, 254 const double cdevproc_result[10]); 255 256 /* 257 * Fall back to the default text processing code when needed. 258 */ 259 int pdf_default_text_begin(gs_text_enum_t *pte, const gs_text_params_t *text, 260 gs_text_enum_t **ppte); 261 262 /* 263 * Check for simple font. 264 */ 265 bool pdf_is_simple_font(gs_font *font); 266 267 /* 268 * Check for CID font. 269 */ 270 bool pdf_is_CID_font(gs_font *font); 271 272 /* Release a text characters colloction. */ 273 void pdf_text_release_cgp(pdf_text_enum_t *penum); 274 275 /* Return char code by glyph. */ 276 gs_char pdf_find_glyph(pdf_font_resource_t *pdfont, gs_glyph glyph); 277 278 /* ------ gdevpdtc.c ------ */ 279 280 PROCESS_TEXT_PROC(process_composite_text); 281 PROCESS_TEXT_PROC(process_cmap_text); 282 PROCESS_TEXT_PROC(process_cid_text); 283 284 /* ------ gdevpdte.c ------ */ 285 286 PROCESS_TEXT_PROC(process_plain_text); 287 288 /* 289 * Process a string with a simple gs_font. 290 */ 291 int pdf_process_string_aux(pdf_text_enum_t *penum, gs_string *pstr, 292 const gs_glyph *gdata, const gs_matrix *pfmat, 293 pdf_text_process_state_t *ppts); 294 295 /* 296 * Emulate TEXT_ADD_TO_ALL_WIDTHS and/or TEXT_ADD_TO_SPACE_WIDTH, 297 * and implement TEXT_REPLACE_WIDTHS if requested. 298 * Uses and updates ppts->values.matrix; uses ppts->values.pdfont. 299 */ 300 int process_text_modify_width(pdf_text_enum_t *pte, gs_font *font, 301 pdf_text_process_state_t *ppts, 302 const gs_const_string *pstr, 303 gs_point *pdpt, const gs_glyph *gdata, bool composite); 304 305 /* 306 * Add char code pair to ToUnicode CMap, 307 * creating the CMap on neccessity. 308 */ 309 int 310 pdf_add_ToUnicode(gx_device_pdf *pdev, gs_font *font, pdf_font_resource_t *pdfont, 311 gs_glyph glyph, gs_char ch, const gs_const_string *gnstr); 312 313 /* 314 * Get character code from a glyph code. 315 * An usage of this function is very undesirable, 316 * because a glyph may be unlisted in Encoding. 317 */ 318 int pdf_encode_glyph(gs_font_base *bfont, gs_glyph glyph0, 319 byte *buf, int buf_size, int *char_code_length); 320 321 int pdf_shift_text_currentpoint(pdf_text_enum_t *penum, gs_point *wpt); 322 323 void adjust_first_last_char(pdf_font_resource_t *pdfont, byte *str, int size); 324 325 float pdf_calculate_text_size(gs_imager_state *pis, pdf_font_resource_t *pdfont, 326 const gs_matrix *pfmat, gs_matrix *smat, gs_matrix *tmat, 327 gs_font *font, gx_device_pdf *pdev); 328 #endif /* gdevpdtt_INCLUDED */ 329