static int find_destination_pages(fz_context *ctx, pdf_obj *current, int page_num, pdf_obj **dest_pages, int *index)
{
if(!strcmp(pdf_to_name(ctx, pdf_dict_gets(ctx, current, "Type")), "Page")) {
return(--page_num);
} if(!strcmp(pdf_to_name(ctx, pdf_dict_gets(ctx, current, "Type")), "Pages")) {
pdf_obj *kids = pdf_dict_gets(ctx, current, "Kids");
pdf_obj *count_obj = pdf_dict_gets(ctx, current, "Count");
if(!pdf_is_array(ctx, kids) || !pdf_is_int(ctx, count_obj))
return(-2);
int count = pdf_to_int(ctx, count_obj);
int i;
for(i = 0; i < count; i++) {
pdf_obj *current_kid = pdf_array_get(ctx, kids, i);
page_num = find_destination_pages(ctx, current_kid, page_num, dest_pages, index);
if(page_num == -1) {
*index = i;
*dest_pages = current;
return(-2);
} else if(page_num == -2) {
return(-2); // just return, preserve index and dest_pages
}
}
return(page_num);
}
return(page_num);
}
pdfout_data *
pdfout_data_scalar_from_pdf (fz_context *ctx, pdf_obj *obj)
{
const char *s;
if (pdf_is_null (ctx, obj))
return pdfout_data_scalar_new (ctx, "null", strlen ("null"));
else if (pdf_is_bool (ctx, obj))
{
if (pdf_to_bool (ctx, obj))
s = "true";
else
s = "false";
return pdfout_data_scalar_new (ctx, s, strlen (s));
}
else if (pdf_is_name (ctx, obj))
{
s = pdf_to_name (ctx, obj);
return pdfout_data_scalar_new (ctx, s, strlen (s));
}
else if (pdf_is_string (ctx, obj))
{
int len;
char *str = pdfout_str_obj_to_utf8 (ctx, obj, &len);
pdfout_data *result = pdfout_data_scalar_new (ctx, str, len);
free (str);
return result;
}
else if (pdf_is_int (ctx, obj))
{
int n = pdf_to_int (ctx, obj);
char buf[200];
int len = pdfout_snprintf (ctx, buf, "%d", n);
return pdfout_data_scalar_new (ctx, buf, len);
}
else if (pdf_is_real (ctx, obj))
{
float f = pdf_to_real (ctx, obj);
char buf[200];
int len = pdfout_snprintf (ctx, buf, "%g", f);
return pdfout_data_scalar_new (ctx, buf, len);
}
else
abort();
}
static void
gathershadings(int page, pdf_obj *pageref, pdf_obj *pageobj, pdf_obj *dict)
{
int i, n;
n = pdf_dict_len(dict);
for (i = 0; i < n; i++)
{
pdf_obj *shade;
pdf_obj *type;
int k;
shade = pdf_dict_get_val(dict, i);
if (!pdf_is_dict(shade))
{
fz_warn(ctx, "not a shading dict (%d %d R)", pdf_to_num(shade), pdf_to_gen(shade));
continue;
}
type = pdf_dict_gets(shade, "ShadingType");
if (!pdf_is_int(type) || pdf_to_int(type) < 1 || pdf_to_int(type) > 7)
{
fz_warn(ctx, "not a shading type (%d %d R)", pdf_to_num(shade), pdf_to_gen(shade));
type = NULL;
}
for (k = 0; k < shadings; k++)
if (!pdf_objcmp(shading[k].u.shading.obj, shade))
break;
if (k < shadings)
continue;
shading = fz_resize_array(ctx, shading, shadings+1, sizeof(struct info));
shadings++;
shading[shadings - 1].page = page;
shading[shadings - 1].pageref = pageref;
shading[shadings - 1].pageobj = pageobj;
shading[shadings - 1].u.shading.obj = shade;
shading[shadings - 1].u.shading.type = type;
}
}
pdfout_data *
pdfout_page_labels_get (fz_context *ctx, pdf_document *doc)
{
pdf_obj *trailer = pdf_trailer (ctx, doc);
pdf_obj *labels_obj = pdf_dict_getp (ctx, trailer, "Root/PageLabels");
pdf_obj *array = pdf_dict_gets (ctx, labels_obj, "Nums");
if (array && pdf_is_array (ctx, array) == false)
pdfout_throw (ctx, "Nums is not an array");
int length = pdf_array_len (ctx, array);
pdfout_data *labels = pdfout_data_array_new (ctx);
for (int i = 0; i < length / 2; ++i)
{
pdf_obj *object = pdf_array_get (ctx, array, 2 * i);
if (pdf_is_int (ctx, object) == false)
pdfout_throw (ctx, "key in number tree not an int");
pdfout_data *hash = pdfout_data_hash_new (ctx);
int page = pdf_to_int (ctx, object);
if (page < 0)
pdfout_throw (ctx, "key in number tree is < 0");
push_int_key (ctx, hash, "page", page);
pdf_obj *dict = pdf_array_get (ctx, array, 2 * i + 1);
if (pdf_is_dict (ctx, dict) == false)
pdfout_throw (ctx, "value in number tree not a dict");
parse_dict (ctx, dict, hash);
pdfout_data_array_push (ctx, labels, hash);
}
return labels;
}
pdf_font_desc *
pdf_load_type3_font(fz_context *ctx, pdf_document *doc, pdf_obj *rdb, pdf_obj *dict)
{
char buf[256];
char *estrings[256];
pdf_font_desc *fontdesc = NULL;
pdf_obj *encoding;
pdf_obj *widths;
pdf_obj *charprocs;
pdf_obj *obj;
int first, last;
int i, k, n;
fz_rect bbox;
fz_matrix matrix;
fz_font *font;
fz_var(fontdesc);
/* Make a new type3 font entry in the document */
if (doc->num_type3_fonts == doc->max_type3_fonts)
{
int new_max = doc->max_type3_fonts * 2;
if (new_max == 0)
new_max = 4;
doc->type3_fonts = fz_resize_array(ctx, doc->type3_fonts, new_max, sizeof(*doc->type3_fonts));
doc->max_type3_fonts = new_max;
}
fz_try(ctx)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Name);
if (pdf_is_name(ctx, obj))
fz_strlcpy(buf, pdf_to_name(ctx, obj), sizeof buf);
else
fz_strlcpy(buf, "Unnamed-T3", sizeof buf);
fontdesc = pdf_new_font_desc(ctx);
obj = pdf_dict_get(ctx, dict, PDF_NAME_FontMatrix);
pdf_to_matrix(ctx, obj, &matrix);
obj = pdf_dict_get(ctx, dict, PDF_NAME_FontBBox);
fz_transform_rect(pdf_to_rect(ctx, obj, &bbox), &matrix);
font = fz_new_type3_font(ctx, buf, &matrix);
fontdesc->font = font;
fontdesc->size += sizeof(fz_font) + 256 * (sizeof(fz_buffer*) + sizeof(float));
fz_set_font_bbox(ctx, font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/* Encoding */
for (i = 0; i < 256; i++)
estrings[i] = NULL;
encoding = pdf_dict_get(ctx, dict, PDF_NAME_Encoding);
if (!encoding)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: Type3 font missing Encoding");
}
if (pdf_is_name(ctx, encoding))
pdf_load_encoding(estrings, pdf_to_name(ctx, encoding));
if (pdf_is_dict(ctx, encoding))
{
pdf_obj *base, *diff, *item;
base = pdf_dict_get(ctx, encoding, PDF_NAME_BaseEncoding);
if (pdf_is_name(ctx, base))
pdf_load_encoding(estrings, pdf_to_name(ctx, base));
diff = pdf_dict_get(ctx, encoding, PDF_NAME_Differences);
if (pdf_is_array(ctx, diff))
{
n = pdf_array_len(ctx, diff);
k = 0;
for (i = 0; i < n; i++)
{
item = pdf_array_get(ctx, diff, i);
if (pdf_is_int(ctx, item))
k = pdf_to_int(ctx, item);
if (pdf_is_name(ctx, item) && k >= 0 && k < nelem(estrings))
estrings[k++] = pdf_to_name(ctx, item);
}
}
}
fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 1);
fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);
pdf_load_to_unicode(ctx, doc, fontdesc, estrings, NULL, pdf_dict_get(ctx, dict, PDF_NAME_ToUnicode));
/* Widths */
pdf_set_default_hmtx(ctx, fontdesc, 0);
first = pdf_to_int(ctx, pdf_dict_get(ctx, dict, PDF_NAME_FirstChar));
last = pdf_to_int(ctx, pdf_dict_get(ctx, dict, PDF_NAME_LastChar));
if (first < 0 || last > 255 || first > last)
first = last = 0;
widths = pdf_dict_get(ctx, dict, PDF_NAME_Widths);
if (!widths)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: Type3 font missing Widths");
}
for (i = first; i <= last; i++)
{
float w = pdf_to_real(ctx, pdf_array_get(ctx, widths, i - first));
w = font->t3matrix.a * w * 1000;
font->t3widths[i] = w * 0.001f;
pdf_add_hmtx(ctx, fontdesc, i, i, w);
}
pdf_end_hmtx(ctx, fontdesc);
/* Resources -- inherit page resources if the font doesn't have its own */
font->t3freeres = pdf_t3_free_resources;
font->t3resources = pdf_dict_get(ctx, dict, PDF_NAME_Resources);
if (!font->t3resources)
font->t3resources = rdb;
if (font->t3resources)
pdf_keep_obj(ctx, font->t3resources);
if (!font->t3resources)
fz_warn(ctx, "no resource dictionary for type 3 font!");
font->t3doc = doc;
font->t3run = pdf_run_glyph_func;
/* CharProcs */
charprocs = pdf_dict_get(ctx, dict, PDF_NAME_CharProcs);
if (!charprocs)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: Type3 font missing CharProcs");
}
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
obj = pdf_dict_gets(ctx, charprocs, estrings[i]);
if (pdf_is_stream(ctx, obj))
{
font->t3procs[i] = pdf_load_stream(ctx, obj);
fz_trim_buffer(ctx, font->t3procs[i]);
fontdesc->size += fz_buffer_storage(ctx, font->t3procs[i], NULL);
fontdesc->size += 0; // TODO: display list size calculation
}
}
}
}
fz_catch(ctx)
{
pdf_drop_font(ctx, fontdesc);
fz_rethrow(ctx);
}
doc->type3_fonts[doc->num_type3_fonts++] = fz_keep_font(ctx, font);
return fontdesc;
}
/*
* Load CMap stream in PDF file
*/
pdf_cmap *
pdf_load_embedded_cmap(pdf_document *doc, pdf_obj *stmobj)
{
fz_stream *file = NULL;
pdf_cmap *cmap = NULL;
pdf_cmap *usecmap;
pdf_obj *wmode;
pdf_obj *obj = NULL;
fz_context *ctx = doc->ctx;
int phase = 0;
fz_var(phase);
fz_var(obj);
fz_var(file);
fz_var(cmap);
if (pdf_obj_marked(stmobj))
fz_throw(ctx, FZ_ERROR_GENERIC, "Recursion in embedded cmap");
if ((cmap = pdf_find_item(ctx, pdf_free_cmap_imp, stmobj)) != NULL)
{
return cmap;
}
fz_try(ctx)
{
file = pdf_open_stream(doc, pdf_to_num(stmobj), pdf_to_gen(stmobj));
phase = 1;
cmap = pdf_load_cmap(ctx, file);
phase = 2;
fz_close(file);
file = NULL;
wmode = pdf_dict_gets(stmobj, "WMode");
if (pdf_is_int(wmode))
pdf_set_cmap_wmode(ctx, cmap, pdf_to_int(wmode));
obj = pdf_dict_gets(stmobj, "UseCMap");
if (pdf_is_name(obj))
{
usecmap = pdf_load_system_cmap(ctx, pdf_to_name(obj));
pdf_set_usecmap(ctx, cmap, usecmap);
pdf_drop_cmap(ctx, usecmap);
}
else if (pdf_is_indirect(obj))
{
phase = 3;
pdf_mark_obj(obj);
usecmap = pdf_load_embedded_cmap(doc, obj);
pdf_unmark_obj(obj);
phase = 4;
pdf_set_usecmap(ctx, cmap, usecmap);
pdf_drop_cmap(ctx, usecmap);
}
pdf_store_item(ctx, stmobj, cmap, pdf_cmap_size(ctx, cmap));
}
fz_catch(ctx)
{
if (file)
fz_close(file);
if (cmap)
pdf_drop_cmap(ctx, cmap);
if (phase < 1)
fz_rethrow_message(ctx, "cannot open cmap stream (%d %d R)", pdf_to_num(stmobj), pdf_to_gen(stmobj));
else if (phase < 2)
fz_rethrow_message(ctx, "cannot parse cmap stream (%d %d R)", pdf_to_num(stmobj), pdf_to_gen(stmobj));
else if (phase < 3)
fz_rethrow_message(ctx, "cannot load system usecmap '%s'", pdf_to_name(obj));
else
{
if (phase == 3)
pdf_unmark_obj(obj);
fz_rethrow_message(ctx, "cannot load embedded usecmap (%d %d R)", pdf_to_num(obj), pdf_to_gen(obj));
}
}
return cmap;
}
static void
pdf_process_extgstate(fz_context *ctx, pdf_processor *proc, pdf_csi *csi, pdf_obj *dict)
{
pdf_obj *obj;
obj = pdf_dict_get(ctx, dict, PDF_NAME_LW);
if (pdf_is_number(ctx, obj) && proc->op_w)
proc->op_w(ctx, proc, pdf_to_real(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_LC);
if (pdf_is_int(ctx, obj) && proc->op_J)
proc->op_J(ctx, proc, fz_clampi(pdf_to_int(ctx, obj), 0, 2));
obj = pdf_dict_get(ctx, dict, PDF_NAME_LJ);
if (pdf_is_int(ctx, obj) && proc->op_j)
proc->op_j(ctx, proc, fz_clampi(pdf_to_int(ctx, obj), 0, 2));
obj = pdf_dict_get(ctx, dict, PDF_NAME_ML);
if (pdf_is_number(ctx, obj) && proc->op_M)
proc->op_M(ctx, proc, pdf_to_real(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_D);
if (pdf_is_array(ctx, obj) && proc->op_d)
{
pdf_obj *dash_array = pdf_array_get(ctx, obj, 0);
pdf_obj *dash_phase = pdf_array_get(ctx, obj, 1);
proc->op_d(ctx, proc, dash_array, pdf_to_real(ctx, dash_phase));
}
obj = pdf_dict_get(ctx, dict, PDF_NAME_RI);
if (pdf_is_name(ctx, obj) && proc->op_ri)
proc->op_ri(ctx, proc, pdf_to_name(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_FL);
if (pdf_is_number(ctx, obj) && proc->op_i)
proc->op_i(ctx, proc, pdf_to_real(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_Font);
if (pdf_is_array(ctx, obj) && proc->op_Tf)
{
pdf_obj *font_ref = pdf_array_get(ctx, obj, 0);
pdf_obj *font_size = pdf_array_get(ctx, obj, 1);
pdf_font_desc *font = load_font_or_hail_mary(ctx, csi->doc, csi->rdb, font_ref, 0, csi->cookie);
fz_try(ctx)
proc->op_Tf(ctx, proc, "ExtGState", font, pdf_to_real(ctx, font_size));
fz_always(ctx)
pdf_drop_font(ctx, font);
fz_catch(ctx)
fz_rethrow(ctx);
}
/* transfer functions */
obj = pdf_dict_get(ctx, dict, PDF_NAME_TR2);
if (pdf_is_name(ctx, obj))
if (!pdf_name_eq(ctx, obj, PDF_NAME_Identity) && !pdf_name_eq(ctx, obj, PDF_NAME_Default))
fz_warn(ctx, "ignoring transfer function");
if (!obj) /* TR is ignored in the presence of TR2 */
{
pdf_obj *tr = pdf_dict_get(ctx, dict, PDF_NAME_TR);
if (pdf_is_name(ctx, tr))
if (!pdf_name_eq(ctx, tr, PDF_NAME_Identity))
fz_warn(ctx, "ignoring transfer function");
}
/* transparency state */
obj = pdf_dict_get(ctx, dict, PDF_NAME_CA);
if (pdf_is_number(ctx, obj) && proc->op_gs_CA)
proc->op_gs_CA(ctx, proc, pdf_to_real(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_ca);
if (pdf_is_number(ctx, obj) && proc->op_gs_ca)
proc->op_gs_ca(ctx, proc, pdf_to_real(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_BM);
if (pdf_is_array(ctx, obj))
obj = pdf_array_get(ctx, obj, 0);
if (pdf_is_name(ctx, obj) && proc->op_gs_BM)
proc->op_gs_BM(ctx, proc, pdf_to_name(ctx, obj));
obj = pdf_dict_get(ctx, dict, PDF_NAME_SMask);
if (proc->op_gs_SMask)
{
if (pdf_is_dict(ctx, obj))
{
pdf_xobject *xobj;
pdf_obj *group, *s, *bc, *tr;
float softmask_bc[FZ_MAX_COLORS];
fz_colorspace *colorspace;
int colorspace_n = 1;
int k, luminosity;
fz_var(xobj);
group = pdf_dict_get(ctx, obj, PDF_NAME_G);
if (!group)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot load softmask xobject (%d 0 R)", pdf_to_num(ctx, obj));
xobj = pdf_load_xobject(ctx, csi->doc, group);
fz_try(ctx)
{
colorspace = pdf_xobject_colorspace(ctx, xobj);
if (colorspace)
{
colorspace_n = fz_colorspace_n(ctx, colorspace);
fz_drop_colorspace(ctx, colorspace);
}
/* Default background color is black. */
for (k = 0; k < colorspace_n; k++)
softmask_bc[k] = 0;
/* Which in CMYK means not all zeros! This should really be
* a test for subtractive color spaces, but this will have
* to do for now. */
if (colorspace == fz_device_cmyk(ctx))
softmask_bc[3] = 1.0;
bc = pdf_dict_get(ctx, obj, PDF_NAME_BC);
if (pdf_is_array(ctx, bc))
{
for (k = 0; k < colorspace_n; k++)
softmask_bc[k] = pdf_to_real(ctx, pdf_array_get(ctx, bc, k));
}
s = pdf_dict_get(ctx, obj, PDF_NAME_S);
if (pdf_name_eq(ctx, s, PDF_NAME_Luminosity))
luminosity = 1;
else
luminosity = 0;
tr = pdf_dict_get(ctx, obj, PDF_NAME_TR);
if (tr && !pdf_name_eq(ctx, tr, PDF_NAME_Identity))
fz_warn(ctx, "ignoring transfer function");
proc->op_gs_SMask(ctx, proc, xobj, csi->rdb, softmask_bc, luminosity);
}
fz_always(ctx)
{
pdf_drop_xobject(ctx, xobj);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
else if (pdf_is_name(ctx, obj) && pdf_name_eq(ctx, obj, PDF_NAME_None))
{
proc->op_gs_SMask(ctx, proc, NULL, NULL, NULL, 0);
}
}
fz_link_dest
pdf_parse_link_dest(pdf_document *xref, pdf_obj *dest)
{
fz_link_dest ld;
pdf_obj *obj;
int l_from_2 = 0;
int b_from_3 = 0;
int r_from_4 = 0;
int t_from_5 = 0;
int t_from_3 = 0;
int t_from_2 = 0;
int z_from_4 = 0;
dest = resolve_dest(xref, dest);
if (dest == NULL || !pdf_is_array(dest))
{
ld.kind = FZ_LINK_NONE;
return ld;
}
obj = pdf_array_get(dest, 0);
if (pdf_is_int(obj))
ld.ld.gotor.page = pdf_to_int(obj);
else
ld.ld.gotor.page = pdf_lookup_page_number(xref, obj);
ld.kind = FZ_LINK_GOTO;
ld.ld.gotor.flags = 0;
ld.ld.gotor.lt.x = 0;
ld.ld.gotor.lt.y = 0;
ld.ld.gotor.rb.x = 0;
ld.ld.gotor.rb.y = 0;
ld.ld.gotor.file_spec = NULL;
ld.ld.gotor.new_window = 0;
obj = pdf_array_get(dest, 1);
if (!pdf_is_name(obj))
return ld;
if (!strcmp("XYZ", pdf_to_name(obj)))
{
l_from_2 = t_from_3 = z_from_4 = 1;
ld.ld.gotor.flags |= fz_link_flag_r_is_zoom;
}
else if ((!strcmp("Fit", pdf_to_name(obj))) || (!strcmp("FitB", pdf_to_name(obj))))
{
ld.ld.gotor.flags |= fz_link_flag_fit_h;
ld.ld.gotor.flags |= fz_link_flag_fit_v;
}
else if ((!strcmp("FitH", pdf_to_name(obj))) || (!strcmp("FitBH", pdf_to_name(obj))))
{
t_from_2 = 1;
ld.ld.gotor.flags |= fz_link_flag_fit_h;
}
else if ((!strcmp("FitV", pdf_to_name(obj))) || (!strcmp("FitBV", pdf_to_name(obj))))
{
l_from_2 = 1;
ld.ld.gotor.flags |= fz_link_flag_fit_v;
}
else if (!strcmp("FitR", pdf_to_name(obj)))
{
l_from_2 = b_from_3 = r_from_4 = t_from_5 = 1;
ld.ld.gotor.flags |= fz_link_flag_fit_h;
ld.ld.gotor.flags |= fz_link_flag_fit_v;
}
if (l_from_2)
{
obj = pdf_array_get(dest, 2);
if (pdf_is_int(obj))
{
ld.ld.gotor.flags |= fz_link_flag_l_valid;
ld.ld.gotor.lt.x = pdf_to_int(obj);
}
else if (pdf_is_real(obj))
{
ld.ld.gotor.flags |= fz_link_flag_l_valid;
ld.ld.gotor.lt.x = pdf_to_real(obj);
}
}
if (b_from_3)
{
obj = pdf_array_get(dest, 3);
if (pdf_is_int(obj))
{
ld.ld.gotor.flags |= fz_link_flag_b_valid;
ld.ld.gotor.rb.y = pdf_to_int(obj);
}
else if (pdf_is_real(obj))
{
ld.ld.gotor.flags |= fz_link_flag_b_valid;
ld.ld.gotor.rb.y = pdf_to_real(obj);
}
}
if (r_from_4)
{
obj = pdf_array_get(dest, 4);
if (pdf_is_int(obj))
{
ld.ld.gotor.flags |= fz_link_flag_r_valid;
ld.ld.gotor.rb.x = pdf_to_int(obj);
}
else if (pdf_is_real(obj))
{
ld.ld.gotor.flags |= fz_link_flag_r_valid;
ld.ld.gotor.rb.x = pdf_to_real(obj);
}
}
if (t_from_5 || t_from_3 || t_from_2)
{
if (t_from_5)
obj = pdf_array_get(dest, 5);
else if (t_from_3)
obj = pdf_array_get(dest, 3);
else
obj = pdf_array_get(dest, 2);
if (pdf_is_int(obj))
{
ld.ld.gotor.flags |= fz_link_flag_t_valid;
ld.ld.gotor.lt.y = pdf_to_int(obj);
}
else if (pdf_is_real(obj))
{
ld.ld.gotor.flags |= fz_link_flag_t_valid;
ld.ld.gotor.lt.y = pdf_to_real(obj);
}
}
if (z_from_4)
{
obj = pdf_array_get(dest, 4);
if (pdf_is_int(obj))
{
ld.ld.gotor.flags |= fz_link_flag_r_valid;
ld.ld.gotor.rb.x = pdf_to_int(obj);
}
else if (pdf_is_real(obj))
{
ld.ld.gotor.flags |= fz_link_flag_r_valid;
ld.ld.gotor.rb.x = pdf_to_real(obj);
}
}
/* Duplicate the values out for the sake of stupid clients */
if ((ld.ld.gotor.flags & (fz_link_flag_l_valid | fz_link_flag_r_valid)) == fz_link_flag_l_valid)
ld.ld.gotor.rb.x = ld.ld.gotor.lt.x;
if ((ld.ld.gotor.flags & (fz_link_flag_l_valid | fz_link_flag_r_valid | fz_link_flag_r_is_zoom)) == fz_link_flag_r_valid)
ld.ld.gotor.lt.x = ld.ld.gotor.rb.x;
if ((ld.ld.gotor.flags & (fz_link_flag_t_valid | fz_link_flag_b_valid)) == fz_link_flag_t_valid)
ld.ld.gotor.rb.y = ld.ld.gotor.lt.y;
if ((ld.ld.gotor.flags & (fz_link_flag_t_valid | fz_link_flag_b_valid)) == fz_link_flag_b_valid)
ld.ld.gotor.lt.y = ld.ld.gotor.rb.y;
return ld;
}
int
pdf_repair_obj(fz_context *ctx, pdf_document *doc, pdf_lexbuf *buf, fz_off_t *stmofsp, int *stmlenp, pdf_obj **encrypt, pdf_obj **id, pdf_obj **page, fz_off_t *tmpofs, pdf_obj **root)
{
fz_stream *file = doc->file;
pdf_token tok;
int stm_len;
*stmofsp = 0;
if (stmlenp)
*stmlenp = -1;
stm_len = 0;
/* On entry to this function, we know that we've just seen
* '<int> <int> obj'. We expect the next thing we see to be a
* pdf object. Regardless of the type of thing we meet next
* we only need to fully parse it if it is a dictionary. */
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_OPEN_DICT)
{
pdf_obj *dict, *obj;
fz_try(ctx)
{
dict = pdf_parse_dict(ctx, doc, file, buf);
}
fz_catch(ctx)
{
fz_rethrow_if(ctx, FZ_ERROR_TRYLATER);
/* Don't let a broken object at EOF overwrite a good one */
if (file->eof)
fz_rethrow(ctx);
/* Silently swallow the error */
dict = pdf_new_dict(ctx, NULL, 2);
}
/* We must be careful not to try to resolve any indirections
* here. We have just read dict, so we know it to be a non
* indirected dictionary. Before we look at any values that
* we get back from looking up in it, we need to check they
* aren't indirected. */
if (encrypt || id || root)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Type);
if (!pdf_is_indirect(ctx, obj) && pdf_name_eq(ctx, obj, PDF_NAME_XRef))
{
if (encrypt)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Encrypt);
if (obj)
{
pdf_drop_obj(ctx, *encrypt);
*encrypt = pdf_keep_obj(ctx, obj);
}
}
if (id)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_ID);
if (obj)
{
pdf_drop_obj(ctx, *id);
*id = pdf_keep_obj(ctx, obj);
}
}
if (root)
*root = pdf_keep_obj(ctx, pdf_dict_get(ctx, dict, PDF_NAME_Root));
}
}
obj = pdf_dict_get(ctx, dict, PDF_NAME_Length);
if (!pdf_is_indirect(ctx, obj) && pdf_is_int(ctx, obj))
stm_len = pdf_to_int(ctx, obj);
if (doc->file_reading_linearly && page)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Type);
if (!pdf_is_indirect(ctx, obj) && pdf_name_eq(ctx, obj, PDF_NAME_Page))
{
pdf_drop_obj(ctx, *page);
*page = pdf_keep_obj(ctx, dict);
}
}
pdf_drop_obj(ctx, dict);
}
static fz_image *
pdf_load_image_imp(fz_context *ctx, pdf_document *doc, pdf_obj *rdb, pdf_obj *dict, fz_stream *cstm, int forcemask)
{
fz_image *image = NULL;
pdf_obj *obj, *res;
int w, h, bpc, n;
int imagemask;
int interpolate;
int indexed;
fz_image *mask = NULL; /* explicit mask/soft mask image */
int use_colorkey = 0;
fz_colorspace *colorspace = NULL;
float decode[FZ_MAX_COLORS * 2];
int colorkey[FZ_MAX_COLORS * 2];
int stride;
int i;
fz_compressed_buffer *buffer;
/* special case for JPEG2000 images */
if (pdf_is_jpx_image(ctx, dict))
return pdf_load_jpx_imp(ctx, doc, rdb, dict, cstm, forcemask);
w = pdf_to_int(ctx, pdf_dict_geta(ctx, dict, PDF_NAME(Width), PDF_NAME(W)));
h = pdf_to_int(ctx, pdf_dict_geta(ctx, dict, PDF_NAME(Height), PDF_NAME(H)));
bpc = pdf_to_int(ctx, pdf_dict_geta(ctx, dict, PDF_NAME(BitsPerComponent), PDF_NAME(BPC)));
if (bpc == 0)
bpc = 8;
imagemask = pdf_to_bool(ctx, pdf_dict_geta(ctx, dict, PDF_NAME(ImageMask), PDF_NAME(IM)));
interpolate = pdf_to_bool(ctx, pdf_dict_geta(ctx, dict, PDF_NAME(Interpolate), PDF_NAME(I)));
indexed = 0;
use_colorkey = 0;
if (imagemask)
bpc = 1;
if (w <= 0)
fz_throw(ctx, FZ_ERROR_GENERIC, "image width is zero (or less)");
if (h <= 0)
fz_throw(ctx, FZ_ERROR_GENERIC, "image height is zero (or less)");
if (bpc <= 0)
fz_throw(ctx, FZ_ERROR_GENERIC, "image depth is zero (or less)");
if (bpc > 16)
fz_throw(ctx, FZ_ERROR_GENERIC, "image depth is too large: %d", bpc);
if (w > (1 << 16))
fz_throw(ctx, FZ_ERROR_GENERIC, "image is too wide");
if (h > (1 << 16))
fz_throw(ctx, FZ_ERROR_GENERIC, "image is too high");
fz_var(mask);
fz_var(image);
fz_var(colorspace);
fz_try(ctx)
{
obj = pdf_dict_geta(ctx, dict, PDF_NAME(ColorSpace), PDF_NAME(CS));
if (obj && !imagemask && !forcemask)
{
/* colorspace resource lookup is only done for inline images */
if (pdf_is_name(ctx, obj))
{
res = pdf_dict_get(ctx, pdf_dict_get(ctx, rdb, PDF_NAME(ColorSpace)), obj);
if (res)
obj = res;
}
colorspace = pdf_load_colorspace(ctx, obj);
indexed = fz_colorspace_is_indexed(ctx, colorspace);
n = fz_colorspace_n(ctx, colorspace);
}
else
{
n = 1;
}
obj = pdf_dict_geta(ctx, dict, PDF_NAME(Decode), PDF_NAME(D));
if (obj)
{
for (i = 0; i < n * 2; i++)
decode[i] = pdf_array_get_real(ctx, obj, i);
}
else if (fz_colorspace_is_lab(ctx, colorspace) || fz_colorspace_is_lab_icc(ctx, colorspace))
{
decode[0] = 0;
decode[1] = 100;
decode[2] = -128;
decode[3] = 127;
decode[4] = -128;
decode[5] = 127;
}
else
{
float maxval = indexed ? (1 << bpc) - 1 : 1;
for (i = 0; i < n * 2; i++)
decode[i] = i & 1 ? maxval : 0;
}
obj = pdf_dict_geta(ctx, dict, PDF_NAME(SMask), PDF_NAME(Mask));
if (pdf_is_dict(ctx, obj))
{
/* Not allowed for inline images or soft masks */
if (cstm)
fz_warn(ctx, "Ignoring invalid inline image soft mask");
else if (forcemask)
fz_warn(ctx, "Ignoring recursive image soft mask");
else
{
mask = pdf_load_image_imp(ctx, doc, rdb, obj, NULL, 1);
obj = pdf_dict_get(ctx, obj, PDF_NAME(Matte));
if (pdf_is_array(ctx, obj))
{
use_colorkey = 1;
for (i = 0; i < n; i++)
colorkey[i] = pdf_array_get_real(ctx, obj, i) * 255;
}
}
}
else if (pdf_is_array(ctx, obj))
{
use_colorkey = 1;
for (i = 0; i < n * 2; i++)
{
if (!pdf_is_int(ctx, pdf_array_get(ctx, obj, i)))
{
fz_warn(ctx, "invalid value in color key mask");
use_colorkey = 0;
}
colorkey[i] = pdf_array_get_int(ctx, obj, i);
}
}
/* Do we load from a ref, or do we load an inline stream? */
if (cstm == NULL)
{
/* Just load the compressed image data now and we can decode it on demand. */
buffer = pdf_load_compressed_stream(ctx, doc, pdf_to_num(ctx, dict));
image = fz_new_image_from_compressed_buffer(ctx, w, h, bpc, colorspace, 96, 96, interpolate, imagemask, decode, use_colorkey ? colorkey : NULL, buffer, mask);
image->invert_cmyk_jpeg = 0;
}
else
{
/* Inline stream */
stride = (w * n * bpc + 7) / 8;
image = fz_new_image_from_compressed_buffer(ctx, w, h, bpc, colorspace, 96, 96, interpolate, imagemask, decode, use_colorkey ? colorkey : NULL, NULL, mask);
image->invert_cmyk_jpeg = 0;
pdf_load_compressed_inline_image(ctx, doc, dict, stride * h, cstm, indexed, (fz_compressed_image *)image);
}
}
fz_always(ctx)
{
fz_drop_colorspace(ctx, colorspace);
fz_drop_image(ctx, mask);
}
fz_catch(ctx)
{
fz_drop_image(ctx, image);
fz_rethrow(ctx);
}
return image;
}
static pdf_image *
pdf_load_image_imp(pdf_document *xref, pdf_obj *rdb, pdf_obj *dict, fz_stream *cstm, int forcemask)
{
fz_stream *stm = NULL;
pdf_image *image = NULL;
pdf_obj *obj, *res;
int w, h, bpc, n;
int imagemask;
int interpolate;
int indexed;
fz_image *mask = NULL; /* explicit mask/soft mask image */
int usecolorkey;
int i;
fz_context *ctx = xref->ctx;
fz_var(stm);
fz_var(mask);
image = fz_malloc_struct(ctx, pdf_image);
fz_try(ctx)
{
/* special case for JPEG2000 images */
if (pdf_is_jpx_image(ctx, dict))
{
pdf_load_jpx(xref, dict, image, forcemask);
if (forcemask)
{
fz_pixmap *mask_pixmap;
if (image->n != 2)
fz_throw(ctx, "soft mask must be grayscale");
mask_pixmap = fz_alpha_from_gray(ctx, image->tile, 1);
fz_drop_pixmap(ctx, image->tile);
image->tile = mask_pixmap;
}
break; /* Out of fz_try */
}
w = pdf_to_int(pdf_dict_getsa(dict, "Width", "W"));
h = pdf_to_int(pdf_dict_getsa(dict, "Height", "H"));
bpc = pdf_to_int(pdf_dict_getsa(dict, "BitsPerComponent", "BPC"));
if (bpc == 0)
bpc = 8;
imagemask = pdf_to_bool(pdf_dict_getsa(dict, "ImageMask", "IM"));
interpolate = pdf_to_bool(pdf_dict_getsa(dict, "Interpolate", "I"));
indexed = 0;
usecolorkey = 0;
mask = NULL;
if (imagemask)
bpc = 1;
if (w <= 0)
fz_throw(ctx, "image width is zero (or less)");
if (h <= 0)
fz_throw(ctx, "image height is zero (or less)");
if (bpc <= 0)
fz_throw(ctx, "image depth is zero (or less)");
if (bpc > 16)
fz_throw(ctx, "image depth is too large: %d", bpc);
if (w > (1 << 16))
fz_throw(ctx, "image is too wide");
if (h > (1 << 16))
fz_throw(ctx, "image is too high");
obj = pdf_dict_getsa(dict, "ColorSpace", "CS");
if (obj && !imagemask && !forcemask)
{
/* colorspace resource lookup is only done for inline images */
if (pdf_is_name(obj))
{
res = pdf_dict_get(pdf_dict_gets(rdb, "ColorSpace"), obj);
if (res)
obj = res;
}
image->base.colorspace = pdf_load_colorspace(xref, obj);
if (!strcmp(image->base.colorspace->name, "Indexed"))
indexed = 1;
n = image->base.colorspace->n;
}
else
{
n = 1;
}
obj = pdf_dict_getsa(dict, "Decode", "D");
if (obj)
{
for (i = 0; i < n * 2; i++)
image->decode[i] = pdf_to_real(pdf_array_get(obj, i));
}
else
{
float maxval = indexed ? (1 << bpc) - 1 : 1;
for (i = 0; i < n * 2; i++)
image->decode[i] = i & 1 ? maxval : 0;
}
obj = pdf_dict_getsa(dict, "SMask", "Mask");
if (pdf_is_dict(obj))
{
/* Not allowed for inline images or soft masks */
if (cstm)
fz_warn(ctx, "Ignoring invalid inline image soft mask");
else if (forcemask)
fz_warn(ctx, "Ignoring recursive image soft mask");
else
mask = (fz_image *)pdf_load_image_imp(xref, rdb, obj, NULL, 1);
}
else if (pdf_is_array(obj))
{
usecolorkey = 1;
for (i = 0; i < n * 2; i++)
{
if (!pdf_is_int(pdf_array_get(obj, i)))
{
fz_warn(ctx, "invalid value in color key mask");
usecolorkey = 0;
}
image->colorkey[i] = pdf_to_int(pdf_array_get(obj, i));
}
}
/* Now, do we load a ref, or do we load the actual thing? */
FZ_INIT_STORABLE(&image->base, 1, pdf_free_image);
image->base.get_pixmap = pdf_image_get_pixmap;
image->base.w = w;
image->base.h = h;
image->base.bpc = bpc;
image->n = n;
image->interpolate = interpolate;
image->imagemask = imagemask;
image->usecolorkey = usecolorkey;
image->base.mask = mask;
if (!cstm)
{
/* Just load the compressed image data now and we can
* decode it on demand. */
int num = pdf_to_num(dict);
int gen = pdf_to_gen(dict);
image->buffer = pdf_load_compressed_stream(xref, num, gen);
break; /* Out of fz_try */
}
/* We need to decompress the image now */
if (cstm)
{
int stride = (w * image->n * image->base.bpc + 7) / 8;
stm = pdf_open_inline_stream(xref, dict, stride * h, cstm, NULL);
}
else
{
stm = pdf_open_stream(xref, pdf_to_num(dict), pdf_to_gen(dict));
}
image->tile = decomp_image_from_stream(ctx, stm, image, cstm != NULL, indexed, 0, 0, 0);
}
fz_catch(ctx)
{
pdf_free_image(ctx, (fz_storable *) image);
fz_rethrow(ctx);
}
return image;
}
char *
pdf_parse_link_dest(fz_context *ctx, pdf_document *doc, pdf_obj *dest)
{
pdf_obj *obj;
char buf[256];
const char *ld;
int page;
int x, y;
dest = resolve_dest(ctx, doc, dest);
if (dest == NULL)
{
fz_warn(ctx, "undefined link destination");
return NULL;
}
if (pdf_is_name(ctx, dest))
{
ld = pdf_to_name(ctx, dest);
return fz_strdup(ctx, ld);
}
else if (pdf_is_string(ctx, dest))
{
ld = pdf_to_str_buf(ctx, dest);
return fz_strdup(ctx, ld);
}
obj = pdf_array_get(ctx, dest, 0);
if (pdf_is_int(ctx, obj))
page = pdf_to_int(ctx, obj);
else
{
fz_try(ctx)
page = pdf_lookup_page_number(ctx, doc, obj);
fz_catch(ctx)
page = -1;
}
x = y = 0;
obj = pdf_array_get(ctx, dest, 1);
if (pdf_name_eq(ctx, obj, PDF_NAME_XYZ))
{
x = pdf_to_int(ctx, pdf_array_get(ctx, dest, 2));
y = pdf_to_int(ctx, pdf_array_get(ctx, dest, 3));
}
else if (pdf_name_eq(ctx, obj, PDF_NAME_FitR))
{
x = pdf_to_int(ctx, pdf_array_get(ctx, dest, 2));
y = pdf_to_int(ctx, pdf_array_get(ctx, dest, 5));
}
else if (pdf_name_eq(ctx, obj, PDF_NAME_FitH) || pdf_name_eq(ctx, obj, PDF_NAME_FitBH))
y = pdf_to_int(ctx, pdf_array_get(ctx, dest, 2));
else if (pdf_name_eq(ctx, obj, PDF_NAME_FitV) || pdf_name_eq(ctx, obj, PDF_NAME_FitBV))
x = pdf_to_int(ctx, pdf_array_get(ctx, dest, 2));
if (page >= 0)
{
if (x != 0 || y != 0)
fz_snprintf(buf, sizeof buf, "#%d,%d,%d", page + 1, x, y);
else
fz_snprintf(buf, sizeof buf, "#%d", page + 1);
return fz_strdup(ctx, buf);
}
return NULL;
}
/*
* Load CMap stream in PDF file
*/
pdf_cmap *
pdf_load_embedded_cmap(fz_context *ctx, pdf_document *doc, pdf_obj *stmobj)
{
fz_stream *file = NULL;
pdf_cmap *cmap = NULL;
pdf_cmap *usecmap = NULL;
pdf_obj *obj;
fz_var(file);
fz_var(cmap);
fz_var(usecmap);
if (pdf_obj_marked(ctx, stmobj))
fz_throw(ctx, FZ_ERROR_GENERIC, "Recursion in embedded cmap");
if ((cmap = pdf_find_item(ctx, pdf_drop_cmap_imp, stmobj)) != NULL)
return cmap;
fz_try(ctx)
{
file = pdf_open_stream(ctx, stmobj);
cmap = pdf_load_cmap(ctx, file);
obj = pdf_dict_get(ctx, stmobj, PDF_NAME_WMode);
if (pdf_is_int(ctx, obj))
pdf_set_cmap_wmode(ctx, cmap, pdf_to_int(ctx, obj));
obj = pdf_dict_get(ctx, stmobj, PDF_NAME_UseCMap);
if (pdf_is_name(ctx, obj))
{
usecmap = pdf_load_system_cmap(ctx, pdf_to_name(ctx, obj));
pdf_set_usecmap(ctx, cmap, usecmap);
}
else if (pdf_is_indirect(ctx, obj))
{
if (pdf_mark_obj(ctx, obj))
fz_throw(ctx, FZ_ERROR_GENERIC, "recursive CMap");
fz_try(ctx)
usecmap = pdf_load_embedded_cmap(ctx, doc, obj);
fz_always(ctx)
pdf_unmark_obj(ctx, obj);
fz_catch(ctx)
fz_rethrow(ctx);
pdf_set_usecmap(ctx, cmap, usecmap);
}
pdf_store_item(ctx, stmobj, cmap, pdf_cmap_size(ctx, cmap));
}
fz_always(ctx)
{
fz_drop_stream(ctx, file);
pdf_drop_cmap(ctx, usecmap);
}
fz_catch(ctx)
{
pdf_drop_cmap(ctx, cmap);
fz_rethrow(ctx);
}
return cmap;
}
static fz_image *
pdf_load_image_imp(pdf_document *doc, pdf_obj *rdb, pdf_obj *dict, fz_stream *cstm, int forcemask)
{
fz_stream *stm = NULL;
fz_image *image = NULL;
pdf_obj *obj, *res;
int w, h, bpc, n;
int imagemask;
int interpolate;
int indexed;
fz_image *mask = NULL; /* explicit mask/soft mask image */
int usecolorkey = 0;
fz_colorspace *colorspace = NULL;
float decode[FZ_MAX_COLORS * 2];
int colorkey[FZ_MAX_COLORS * 2];
int stride;
int i;
fz_context *ctx = doc->ctx;
fz_compressed_buffer *buffer;
fz_var(stm);
fz_var(mask);
fz_var(image);
fz_var(colorspace);
fz_try(ctx)
{
/* special case for JPEG2000 images */
if (pdf_is_jpx_image(ctx, dict))
{
// image = pdf_load_jpx(doc, dict, forcemask);
//
// if (forcemask)
// {
// fz_pixmap *mask_pixmap;
// if (image->n != 2)
// {
// fz_pixmap *gray;
// fz_irect bbox;
// fz_warn(ctx, "soft mask should be grayscale");
// gray = fz_new_pixmap_with_bbox(ctx, fz_device_gray(ctx), fz_pixmap_bbox(ctx, image->tile, &bbox));
// fz_convert_pixmap(ctx, gray, image->tile);
// fz_drop_pixmap(ctx, image->tile);
// image->tile = gray;
// }
// mask_pixmap = fz_alpha_from_gray(ctx, image->tile, 1);
// fz_drop_pixmap(ctx, image->tile);
// image->tile = mask_pixmap;
// }
break; /* Out of fz_try */
}
w = pdf_to_int(pdf_dict_getsa(dict, "Width", "W"));
h = pdf_to_int(pdf_dict_getsa(dict, "Height", "H"));
bpc = pdf_to_int(pdf_dict_getsa(dict, "BitsPerComponent", "BPC"));
if (bpc == 0)
bpc = 8;
imagemask = pdf_to_bool(pdf_dict_getsa(dict, "ImageMask", "IM"));
interpolate = pdf_to_bool(pdf_dict_getsa(dict, "Interpolate", "I"));
indexed = 0;
usecolorkey = 0;
if (imagemask)
bpc = 1;
if (w <= 0)
fz_throw(ctx, FZ_ERROR_GENERIC, "image width is zero (or less)");
if (h <= 0)
fz_throw(ctx, FZ_ERROR_GENERIC, "image height is zero (or less)");
if (bpc <= 0)
fz_throw(ctx, FZ_ERROR_GENERIC, "image depth is zero (or less)");
if (bpc > 16)
fz_throw(ctx, FZ_ERROR_GENERIC, "image depth is too large: %d", bpc);
if (w > (1 << 16))
fz_throw(ctx, FZ_ERROR_GENERIC, "image is too wide");
if (h > (1 << 16))
fz_throw(ctx, FZ_ERROR_GENERIC, "image is too high");
obj = pdf_dict_getsa(dict, "ColorSpace", "CS");
if (obj && !imagemask && !forcemask)
{
/* colorspace resource lookup is only done for inline images */
if (pdf_is_name(obj))
{
res = pdf_dict_get(pdf_dict_gets(rdb, "ColorSpace"), obj);
if (res)
obj = res;
}
colorspace = pdf_load_colorspace(doc, obj);
indexed = fz_colorspace_is_indexed(colorspace);
n = colorspace->n;
}
else
{
n = 1;
}
obj = pdf_dict_getsa(dict, "Decode", "D");
if (obj)
{
for (i = 0; i < n * 2; i++)
decode[i] = pdf_to_real(pdf_array_get(obj, i));
}
else
{
float maxval = indexed ? (1 << bpc) - 1 : 1;
for (i = 0; i < n * 2; i++)
decode[i] = i & 1 ? maxval : 0;
}
obj = pdf_dict_getsa(dict, "SMask", "Mask");
if (pdf_is_dict(obj))
{
/* Not allowed for inline images or soft masks */
if (cstm)
fz_warn(ctx, "Ignoring invalid inline image soft mask");
else if (forcemask)
fz_warn(ctx, "Ignoring recursive image soft mask");
else
{
mask = pdf_load_image_imp(doc, rdb, obj, NULL, 1);
obj = pdf_dict_gets(obj, "Matte");
if (pdf_is_array(obj))
{
usecolorkey = 1;
for (i = 0; i < n; i++)
colorkey[i] = pdf_to_real(pdf_array_get(obj, i)) * 255;
}
}
}
else if (pdf_is_array(obj))
{
usecolorkey = 1;
for (i = 0; i < n * 2; i++)
{
if (!pdf_is_int(pdf_array_get(obj, i)))
{
fz_warn(ctx, "invalid value in color key mask");
usecolorkey = 0;
}
colorkey[i] = pdf_to_int(pdf_array_get(obj, i));
}
}
/* Do we load from a ref, or do we load an inline stream? */
if (cstm == NULL)
{
/* Just load the compressed image data now and we can
* decode it on demand. */
int num = pdf_to_num(dict);
int gen = pdf_to_gen(dict);
buffer = pdf_load_compressed_stream(doc, num, gen);
image = fz_new_image(ctx, w, h, bpc, colorspace, 96, 96, interpolate, imagemask, decode, usecolorkey ? colorkey : NULL, buffer, mask);
}
else
{
/* Inline stream */
stride = (w * n * bpc + 7) / 8;
image = fz_new_image(ctx, w, h, bpc, colorspace, 96, 96, interpolate, imagemask, decode, usecolorkey ? colorkey : NULL, NULL, mask);
pdf_load_compressed_inline_image(doc, dict, stride * h, cstm, indexed, image);
}
}
fz_catch(ctx)
{
fz_drop_colorspace(ctx, colorspace);
fz_drop_image(ctx, mask);
fz_drop_image(ctx, image);
fz_rethrow(ctx);
}
return image;
}
static fz_image *
pdf_load_image_imp(pdf_document *xref, pdf_obj *rdb, pdf_obj *dict, fz_stream *cstm, int forcemask)
{
fz_stream *stm = NULL;
fz_image *image = NULL;
pdf_obj *obj, *res;
int w, h, bpc, n;
int imagemask;
int interpolate;
int indexed;
fz_image *mask = NULL; /* explicit mask/soft mask image */
int usecolorkey = 0;
fz_colorspace *colorspace = NULL;
float decode[FZ_MAX_COLORS * 2];
int colorkey[FZ_MAX_COLORS * 2];
int i;
fz_context *ctx = xref->ctx;
fz_var(stm);
fz_var(mask);
fz_var(image);
fz_var(colorspace);
fz_try(ctx)
{
/* special case for JPEG2000 images */
if (pdf_is_jpx_image(ctx, dict))
{
image = pdf_load_jpx(xref, dict, forcemask);
if (forcemask)
{
fz_pixmap *mask_pixmap;
if (image->n != 2)
{
/* SumatraPDF: ignore invalid JPX softmasks */
fz_warn(ctx, "soft mask must be grayscale");
mask_pixmap = fz_new_pixmap(ctx, NULL, image->tile->w, image->tile->h);
fz_clear_pixmap_with_value(ctx, mask_pixmap, 255);
}
else
mask_pixmap = fz_alpha_from_gray(ctx, image->tile, 1);
fz_drop_pixmap(ctx, image->tile);
image->tile = mask_pixmap;
}
break; /* Out of fz_try */
}
w = pdf_to_int(pdf_dict_getsa(dict, "Width", "W"));
h = pdf_to_int(pdf_dict_getsa(dict, "Height", "H"));
bpc = pdf_to_int(pdf_dict_getsa(dict, "BitsPerComponent", "BPC"));
if (bpc == 0)
bpc = 8;
imagemask = pdf_to_bool(pdf_dict_getsa(dict, "ImageMask", "IM"));
interpolate = pdf_to_bool(pdf_dict_getsa(dict, "Interpolate", "I"));
indexed = 0;
usecolorkey = 0;
mask = NULL;
if (imagemask)
bpc = 1;
if (w <= 0)
fz_throw(ctx, "image width is zero (or less)");
if (h <= 0)
fz_throw(ctx, "image height is zero (or less)");
if (bpc <= 0)
fz_throw(ctx, "image depth is zero (or less)");
if (bpc > 16)
fz_throw(ctx, "image depth is too large: %d", bpc);
if (w > (1 << 16))
fz_throw(ctx, "image is too wide");
if (h > (1 << 16))
fz_throw(ctx, "image is too high");
obj = pdf_dict_getsa(dict, "ColorSpace", "CS");
if (obj && !imagemask && !forcemask)
{
/* colorspace resource lookup is only done for inline images */
if (pdf_is_name(obj))
{
res = pdf_dict_get(pdf_dict_gets(rdb, "ColorSpace"), obj);
if (res)
obj = res;
}
colorspace = pdf_load_colorspace(xref, obj);
if (!strcmp(colorspace->name, "Indexed"))
indexed = 1;
n = colorspace->n;
}
else
{
n = 1;
}
obj = pdf_dict_getsa(dict, "Decode", "D");
if (obj)
{
for (i = 0; i < n * 2; i++)
decode[i] = pdf_to_real(pdf_array_get(obj, i));
}
else
{
float maxval = indexed ? (1 << bpc) - 1 : 1;
for (i = 0; i < n * 2; i++)
decode[i] = i & 1 ? maxval : 0;
}
obj = pdf_dict_getsa(dict, "SMask", "Mask");
if (pdf_is_dict(obj))
{
/* Not allowed for inline images or soft masks */
if (cstm)
fz_warn(ctx, "Ignoring invalid inline image soft mask");
else if (forcemask)
fz_warn(ctx, "Ignoring recursive image soft mask");
else
mask = pdf_load_image_imp(xref, rdb, obj, NULL, 1);
}
else if (pdf_is_array(obj))
{
usecolorkey = 1;
for (i = 0; i < n * 2; i++)
{
if (!pdf_is_int(pdf_array_get(obj, i)))
{
fz_warn(ctx, "invalid value in color key mask");
usecolorkey = 0;
}
colorkey[i] = pdf_to_int(pdf_array_get(obj, i));
}
}
/* Now, do we load a ref, or do we load the actual thing? */
if (!cstm)
{
/* Just load the compressed image data now and we can
* decode it on demand. */
int num = pdf_to_num(dict);
int gen = pdf_to_gen(dict);
fz_compressed_buffer *buffer = pdf_load_compressed_stream(xref, num, gen);
image = fz_new_image(ctx, w, h, bpc, colorspace, 96, 96, interpolate, imagemask, decode, usecolorkey ? colorkey : NULL, buffer, mask);
break; /* Out of fz_try */
}
/* We need to decompress the image now */
if (cstm)
{
int stride = (w * n * bpc + 7) / 8;
stm = pdf_open_inline_stream(xref, dict, stride * h, cstm, NULL);
}
else
{
stm = pdf_open_stream(xref, pdf_to_num(dict), pdf_to_gen(dict));
}
image = fz_new_image(ctx, w, h, bpc, colorspace, 96, 96, interpolate, imagemask, decode, usecolorkey ? colorkey : NULL, NULL, mask);
image->tile = fz_decomp_image_from_stream(ctx, stm, image, cstm != NULL, indexed, 0, 0);
}
fz_catch(ctx)
{
/* SumatraPDF: fix memory leak */
if (!image)
fz_drop_colorspace(ctx, colorspace);
else
fz_drop_image(ctx, image);
fz_rethrow(ctx);
}
/* cf. http://bugs.ghostscript.com/show_bug.cgi?id=693517 */
fz_try(ctx)
{
obj = pdf_dict_getp(dict, "SMask/Matte");
if (pdf_is_array(obj) && image->mask)
{
assert(!image->usecolorkey);
image->usecolorkey = 2;
for (i = 0; i < n; i++)
image->colorkey[i] = pdf_to_int(pdf_array_get(obj, i));
}
}
fz_catch(ctx)
{
fz_drop_image(ctx, image);
fz_rethrow(ctx);
}
return image;
}
pdf_font_desc *
pdf_load_type3_font(pdf_document *xref, pdf_obj *rdb, pdf_obj *dict)
{
char buf[256];
char *estrings[256];
pdf_font_desc *fontdesc = NULL;
pdf_obj *encoding;
pdf_obj *widths;
pdf_obj *charprocs;
pdf_obj *obj;
int first, last;
int i, k, n;
fz_rect bbox;
fz_matrix matrix;
fz_context *ctx = xref->ctx;
fz_var(fontdesc);
fz_try(ctx)
{
obj = pdf_dict_gets(dict, "Name");
if (pdf_is_name(obj))
fz_strlcpy(buf, pdf_to_name(obj), sizeof buf);
else
sprintf(buf, "Unnamed-T3");
fontdesc = pdf_new_font_desc(ctx);
obj = pdf_dict_gets(dict, "FontMatrix");
pdf_to_matrix(ctx, obj, &matrix);
obj = pdf_dict_gets(dict, "FontBBox");
fz_transform_rect(pdf_to_rect(ctx, obj, &bbox), &matrix);
fontdesc->font = fz_new_type3_font(ctx, buf, &matrix);
fontdesc->size += sizeof(fz_font) + 256 * (sizeof(fz_buffer*) + sizeof(float));
fz_set_font_bbox(ctx, fontdesc->font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/* Encoding */
for (i = 0; i < 256; i++)
estrings[i] = NULL;
encoding = pdf_dict_gets(dict, "Encoding");
if (!encoding)
{
fz_throw(ctx, "syntaxerror: Type3 font missing Encoding");
}
if (pdf_is_name(encoding))
pdf_load_encoding(estrings, pdf_to_name(encoding));
if (pdf_is_dict(encoding))
{
pdf_obj *base, *diff, *item;
base = pdf_dict_gets(encoding, "BaseEncoding");
if (pdf_is_name(base))
pdf_load_encoding(estrings, pdf_to_name(base));
diff = pdf_dict_gets(encoding, "Differences");
if (pdf_is_array(diff))
{
n = pdf_array_len(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = pdf_array_get(diff, i);
if (pdf_is_int(item))
k = pdf_to_int(item);
if (pdf_is_name(item) && k >= 0 && k < nelem(estrings))
estrings[k++] = pdf_to_name(item);
}
}
}
fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 1);
fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);
pdf_load_to_unicode(xref, fontdesc, estrings, NULL, pdf_dict_gets(dict, "ToUnicode"));
/* Widths */
pdf_set_default_hmtx(ctx, fontdesc, 0);
first = pdf_to_int(pdf_dict_gets(dict, "FirstChar"));
last = pdf_to_int(pdf_dict_gets(dict, "LastChar"));
if (first < 0 || last > 255 || first > last)
first = last = 0;
widths = pdf_dict_gets(dict, "Widths");
if (!widths)
{
fz_throw(ctx, "syntaxerror: Type3 font missing Widths");
}
for (i = first; i <= last; i++)
{
float w = pdf_to_real(pdf_array_get(widths, i - first));
w = fontdesc->font->t3matrix.a * w * 1000;
fontdesc->font->t3widths[i] = w * 0.001f;
pdf_add_hmtx(ctx, fontdesc, i, i, w);
}
pdf_end_hmtx(ctx, fontdesc);
/* Resources -- inherit page resources if the font doesn't have its own */
fontdesc->font->t3freeres = pdf_t3_free_resources;
fontdesc->font->t3resources = pdf_dict_gets(dict, "Resources");
if (!fontdesc->font->t3resources)
fontdesc->font->t3resources = rdb;
if (fontdesc->font->t3resources)
pdf_keep_obj(fontdesc->font->t3resources);
if (!fontdesc->font->t3resources)
fz_warn(ctx, "no resource dictionary for type 3 font!");
fontdesc->font->t3doc = xref;
fontdesc->font->t3run = pdf_run_glyph_func;
/* CharProcs */
charprocs = pdf_dict_gets(dict, "CharProcs");
if (!charprocs)
{
fz_throw(ctx, "syntaxerror: Type3 font missing CharProcs");
}
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
obj = pdf_dict_gets(charprocs, estrings[i]);
if (pdf_is_stream(xref, pdf_to_num(obj), pdf_to_gen(obj)))
{
fontdesc->font->t3procs[i] = pdf_load_stream(xref, pdf_to_num(obj), pdf_to_gen(obj));
fontdesc->size += fontdesc->font->t3procs[i]->cap;
fontdesc->size += 0; // TODO: display list size calculation
}
}
}
}
fz_catch(ctx)
{
if (fontdesc)
pdf_drop_font(ctx, fontdesc);
fz_throw(ctx, "cannot load type3 font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
}
return fontdesc;
}
static void
gatherpatterns(int page, pdf_obj *pageref, pdf_obj *pageobj, pdf_obj *dict)
{
int i, n;
n = pdf_dict_len(dict);
for (i = 0; i < n; i++)
{
pdf_obj *patterndict;
pdf_obj *type;
pdf_obj *paint = NULL;
pdf_obj *tiling = NULL;
pdf_obj *shading = NULL;
int k;
patterndict = pdf_dict_get_val(dict, i);
if (!pdf_is_dict(patterndict))
{
fz_warn(ctx, "not a pattern dict (%d %d R)", pdf_to_num(patterndict), pdf_to_gen(patterndict));
continue;
}
type = pdf_dict_gets(patterndict, "PatternType");
if (!pdf_is_int(type) || pdf_to_int(type) < 1 || pdf_to_int(type) > 2)
{
fz_warn(ctx, "not a pattern type (%d %d R)", pdf_to_num(patterndict), pdf_to_gen(patterndict));
type = NULL;
}
if (pdf_to_int(type) == 1)
{
paint = pdf_dict_gets(patterndict, "PaintType");
if (!pdf_is_int(paint) || pdf_to_int(paint) < 1 || pdf_to_int(paint) > 2)
{
fz_warn(ctx, "not a pattern paint type (%d %d R)", pdf_to_num(patterndict), pdf_to_gen(patterndict));
paint = NULL;
}
tiling = pdf_dict_gets(patterndict, "TilingType");
if (!pdf_is_int(tiling) || pdf_to_int(tiling) < 1 || pdf_to_int(tiling) > 3)
{
fz_warn(ctx, "not a pattern tiling type (%d %d R)", pdf_to_num(patterndict), pdf_to_gen(patterndict));
tiling = NULL;
}
}
else
{
shading = pdf_dict_gets(patterndict, "Shading");
}
for (k = 0; k < patterns; k++)
if (!pdf_objcmp(pattern[k].u.pattern.obj, patterndict))
break;
if (k < patterns)
continue;
pattern = fz_resize_array(ctx, pattern, patterns+1, sizeof(struct info));
patterns++;
pattern[patterns - 1].page = page;
pattern[patterns - 1].pageref = pageref;
pattern[patterns - 1].pageobj = pageobj;
pattern[patterns - 1].u.pattern.obj = patterndict;
pattern[patterns - 1].u.pattern.type = type;
pattern[patterns - 1].u.pattern.paint = paint;
pattern[patterns - 1].u.pattern.tiling = tiling;
pattern[patterns - 1].u.pattern.shading = shading;
}
}
pdf_font_desc *
pdf_load_type3_font(pdf_document *doc, pdf_obj *rdb, pdf_obj *dict)
{
char buf[256];
char *estrings[256];
pdf_font_desc *fontdesc = NULL;
pdf_obj *encoding;
pdf_obj *widths;
pdf_obj *charprocs;
pdf_obj *obj;
int first, last;
int i, k, n;
fz_rect bbox;
fz_matrix matrix;
fz_context *ctx = doc->ctx;
fz_var(fontdesc);
/* Make a new type3 font entry in the document */
if (doc->num_type3_fonts == doc->max_type3_fonts)
{
int new_max = doc->max_type3_fonts * 2;
if (new_max == 0)
new_max = 4;
doc->type3_fonts = fz_resize_array(doc->ctx, doc->type3_fonts, new_max, sizeof(*doc->type3_fonts));
doc->max_type3_fonts = new_max;
}
fz_try(ctx)
{
obj = pdf_dict_gets(dict, "Name");
if (pdf_is_name(obj))
fz_strlcpy(buf, pdf_to_name(obj), sizeof buf);
else
sprintf(buf, "Unnamed-T3");
fontdesc = pdf_new_font_desc(ctx);
obj = pdf_dict_gets(dict, "FontMatrix");
pdf_to_matrix(ctx, obj, &matrix);
obj = pdf_dict_gets(dict, "FontBBox");
fz_transform_rect(pdf_to_rect(ctx, obj, &bbox), &matrix);
fontdesc->font = fz_new_type3_font(ctx, buf, &matrix);
fontdesc->size += sizeof(fz_font) + 256 * (sizeof(fz_buffer*) + sizeof(float));
fz_set_font_bbox(ctx, fontdesc->font, bbox.x0, bbox.y0, bbox.x1, bbox.y1);
/* SumatraPDF: expose Type3 FontDescriptor flags */
fontdesc->flags = pdf_to_int(pdf_dict_gets(pdf_dict_gets(dict, "FontDescriptor"), "Flags"));
/* Encoding */
for (i = 0; i < 256; i++)
estrings[i] = NULL;
encoding = pdf_dict_gets(dict, "Encoding");
if (!encoding)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: Type3 font missing Encoding");
}
if (pdf_is_name(encoding))
pdf_load_encoding(estrings, pdf_to_name(encoding));
if (pdf_is_dict(encoding))
{
pdf_obj *base, *diff, *item;
base = pdf_dict_gets(encoding, "BaseEncoding");
if (pdf_is_name(base))
pdf_load_encoding(estrings, pdf_to_name(base));
diff = pdf_dict_gets(encoding, "Differences");
if (pdf_is_array(diff))
{
n = pdf_array_len(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = pdf_array_get(diff, i);
if (pdf_is_int(item))
k = pdf_to_int(item);
if (pdf_is_name(item) && k >= 0 && k < nelem(estrings))
estrings[k++] = pdf_to_name(item);
}
}
}
fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 1);
fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);
pdf_load_to_unicode(doc, fontdesc, estrings, NULL, pdf_dict_gets(dict, "ToUnicode"));
/* SumatraPDF: trying to match Adobe Reader's behavior */
if (!(fontdesc->flags & PDF_FD_SYMBOLIC) && fontdesc->cid_to_ucs_len >= 128)
for (i = 32; i < 128; i++)
if (fontdesc->cid_to_ucs[i] == '?' || fontdesc->cid_to_ucs[i] == '\0')
fontdesc->cid_to_ucs[i] = i;
/* Widths */
pdf_set_default_hmtx(ctx, fontdesc, 0);
first = pdf_to_int(pdf_dict_gets(dict, "FirstChar"));
last = pdf_to_int(pdf_dict_gets(dict, "LastChar"));
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1966 */
if (first >= 256 && last - first < 256)
{
fz_warn(ctx, "ignoring out-of-bound values for FirstChar/LastChar: %d/%d", first, last);
last -= first;
first = 0;
}
if (first < 0 || last > 255 || first > last)
first = last = 0;
widths = pdf_dict_gets(dict, "Widths");
if (!widths)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: Type3 font missing Widths");
}
for (i = first; i <= last; i++)
{
float w = pdf_to_real(pdf_array_get(widths, i - first));
w = fontdesc->font->t3matrix.a * w * 1000;
fontdesc->font->t3widths[i] = w * 0.001f;
pdf_add_hmtx(ctx, fontdesc, i, i, w);
}
pdf_end_hmtx(ctx, fontdesc);
/* Resources -- inherit page resources if the font doesn't have its own */
fontdesc->font->t3freeres = pdf_t3_free_resources;
fontdesc->font->t3resources = pdf_dict_gets(dict, "Resources");
if (!fontdesc->font->t3resources)
fontdesc->font->t3resources = rdb;
if (fontdesc->font->t3resources)
pdf_keep_obj(fontdesc->font->t3resources);
if (!fontdesc->font->t3resources)
fz_warn(ctx, "no resource dictionary for type 3 font!");
fontdesc->font->t3doc = doc;
fontdesc->font->t3run = pdf_run_glyph_func;
/* CharProcs */
charprocs = pdf_dict_gets(dict, "CharProcs");
if (!charprocs)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: Type3 font missing CharProcs");
}
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
/* SumatraPDF: don't reject fonts with few broken glyphs */
fz_try(ctx)
{
obj = pdf_dict_gets(charprocs, estrings[i]);
if (pdf_is_stream(doc, pdf_to_num(obj), pdf_to_gen(obj)))
{
fontdesc->font->t3procs[i] = pdf_load_stream(doc, pdf_to_num(obj), pdf_to_gen(obj));
fontdesc->size += fontdesc->font->t3procs[i]->cap;
fontdesc->size += 0; // TODO: display list size calculation
}
}
fz_catch(ctx)
{
fz_warn(ctx, "failed to get data for type 3 glyph '%s'", estrings[i]);
}
}
}
}
fz_catch(ctx)
{
if (fontdesc)
pdf_drop_font(ctx, fontdesc);
fz_rethrow_message(ctx, "cannot load type3 font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
}
doc->type3_fonts[doc->num_type3_fonts++] = fz_keep_font(ctx, fontdesc->font);
return fontdesc;
}
static pdf_font_desc *
pdf_load_simple_font(pdf_document *xref, pdf_obj *dict)
{
pdf_obj *descriptor;
pdf_obj *encoding;
pdf_obj *widths;
unsigned short *etable = NULL;
pdf_font_desc *fontdesc = NULL;
char *subtype;
FT_Face face;
FT_CharMap cmap;
int symbolic;
int kind;
char *basefont;
char *estrings[256];
char ebuffer[256][32];
int i, k, n;
int fterr;
fz_context *ctx = xref->ctx;
fz_var(fontdesc);
fz_var(etable);
basefont = pdf_to_name(pdf_dict_gets(dict, "BaseFont"));
/* Load font file */
fz_try(ctx)
{
fontdesc = pdf_new_font_desc(ctx);
descriptor = pdf_dict_gets(dict, "FontDescriptor");
if (descriptor)
pdf_load_font_descriptor(fontdesc, xref, descriptor, NULL, basefont);
else
pdf_load_builtin_font(ctx, fontdesc, basefont);
/* Some chinese documents mistakenly consider WinAnsiEncoding to be codepage 936 */
if (descriptor && pdf_is_string(pdf_dict_gets(descriptor, "FontName")) &&
!pdf_dict_gets(dict, "ToUnicode") &&
!strcmp(pdf_to_name(pdf_dict_gets(dict, "Encoding")), "WinAnsiEncoding") &&
pdf_to_int(pdf_dict_gets(descriptor, "Flags")) == 4)
{
char *cp936fonts[] = {
"\xCB\xCE\xCC\xE5", "SimSun,Regular",
"\xBA\xDA\xCC\xE5", "SimHei,Regular",
"\xBF\xAC\xCC\xE5_GB2312", "SimKai,Regular",
"\xB7\xC2\xCB\xCE_GB2312", "SimFang,Regular",
"\xC1\xA5\xCA\xE9", "SimLi,Regular",
NULL
};
for (i = 0; cp936fonts[i]; i += 2)
if (!strcmp(basefont, cp936fonts[i]))
break;
if (cp936fonts[i])
{
fz_warn(ctx, "workaround for S22PDF lying about chinese font encodings");
pdf_drop_font(ctx, fontdesc);
fontdesc = pdf_new_font_desc(ctx);
pdf_load_font_descriptor(fontdesc, xref, descriptor, "Adobe-GB1", cp936fonts[i+1]);
fontdesc->encoding = pdf_load_system_cmap(ctx, "GBK-EUC-H");
fontdesc->to_unicode = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2");
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2");
face = fontdesc->font->ft_face;
kind = ft_kind(face);
goto skip_encoding;
}
}
face = fontdesc->font->ft_face;
kind = ft_kind(face);
/* Encoding */
symbolic = fontdesc->flags & 4;
if (face->num_charmaps > 0)
cmap = face->charmaps[0];
else
cmap = NULL;
for (i = 0; i < face->num_charmaps; i++)
{
FT_CharMap test = face->charmaps[i];
if (kind == TYPE1)
{
if (test->platform_id == 7)
cmap = test;
}
if (kind == TRUETYPE)
{
if (test->platform_id == 1 && test->encoding_id == 0)
cmap = test;
if (test->platform_id == 3 && test->encoding_id == 1)
cmap = test;
if (symbolic && test->platform_id == 3 && test->encoding_id == 0)
cmap = test;
}
}
if (cmap)
{
fterr = FT_Set_Charmap(face, cmap);
if (fterr)
fz_warn(ctx, "freetype could not set cmap: %s", ft_error_string(fterr));
}
else
fz_warn(ctx, "freetype could not find any cmaps");
etable = fz_malloc_array(ctx, 256, sizeof(unsigned short));
fontdesc->size += 256 * sizeof(unsigned short);
for (i = 0; i < 256; i++)
{
estrings[i] = NULL;
etable[i] = 0;
}
encoding = pdf_dict_gets(dict, "Encoding");
if (encoding)
{
if (pdf_is_name(encoding))
pdf_load_encoding(estrings, pdf_to_name(encoding));
if (pdf_is_dict(encoding))
{
pdf_obj *base, *diff, *item;
base = pdf_dict_gets(encoding, "BaseEncoding");
if (pdf_is_name(base))
pdf_load_encoding(estrings, pdf_to_name(base));
else if (!fontdesc->is_embedded && !symbolic)
pdf_load_encoding(estrings, "StandardEncoding");
diff = pdf_dict_gets(encoding, "Differences");
if (pdf_is_array(diff))
{
n = pdf_array_len(diff);
k = 0;
for (i = 0; i < n; i++)
{
item = pdf_array_get(diff, i);
if (pdf_is_int(item))
k = pdf_to_int(item);
if (pdf_is_name(item) && k >= 0 && k < nelem(estrings))
estrings[k++] = pdf_to_name(item);
}
}
}
}
/* start with the builtin encoding */
for (i = 0; i < 256; i++)
etable[i] = ft_char_index(face, i);
fz_lock(ctx, FZ_LOCK_FREETYPE);
/* built-in and substitute fonts may be a different type than what the document expects */
subtype = pdf_to_name(pdf_dict_gets(dict, "Subtype"));
if (!strcmp(subtype, "Type1"))
kind = TYPE1;
else if (!strcmp(subtype, "MMType1"))
kind = TYPE1;
else if (!strcmp(subtype, "TrueType"))
kind = TRUETYPE;
else if (!strcmp(subtype, "CIDFontType0"))
kind = TYPE1;
else if (!strcmp(subtype, "CIDFontType2"))
kind = TRUETYPE;
/* encode by glyph name where we can */
if (kind == TYPE1)
{
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
etable[i] = FT_Get_Name_Index(face, estrings[i]);
if (etable[i] == 0)
{
int aglcode = pdf_lookup_agl(estrings[i]);
const char **dupnames = pdf_lookup_agl_duplicates(aglcode);
while (*dupnames)
{
etable[i] = FT_Get_Name_Index(face, (char*)*dupnames);
if (etable[i])
break;
dupnames++;
}
}
}
}
}
/* encode by glyph name where we can */
if (kind == TRUETYPE)
{
/* Unicode cmap */
if (!symbolic && face->charmap && face->charmap->platform_id == 3)
{
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
int aglcode = pdf_lookup_agl(estrings[i]);
if (!aglcode)
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ft_char_index(face, aglcode);
}
}
}
/* MacRoman cmap */
else if (!symbolic && face->charmap && face->charmap->platform_id == 1)
{
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
k = lookup_mre_code(estrings[i]);
if (k <= 0)
etable[i] = FT_Get_Name_Index(face, estrings[i]);
else
etable[i] = ft_char_index(face, k);
}
}
}
/* Symbolic cmap */
else if (!face->charmap || face->charmap->encoding != FT_ENCODING_MS_SYMBOL)
{
for (i = 0; i < 256; i++)
{
if (estrings[i])
{
etable[i] = FT_Get_Name_Index(face, estrings[i]);
if (etable[i] == 0)
etable[i] = ft_char_index(face, i);
}
}
}
}
/* try to reverse the glyph names from the builtin encoding */
for (i = 0; i < 256; i++)
{
if (etable[i] && !estrings[i])
{
if (FT_HAS_GLYPH_NAMES(face))
{
fterr = FT_Get_Glyph_Name(face, etable[i], ebuffer[i], 32);
if (fterr)
fz_warn(ctx, "freetype get glyph name (gid %d): %s", etable[i], ft_error_string(fterr));
if (ebuffer[i][0])
estrings[i] = ebuffer[i];
}
else
{
estrings[i] = (char*) pdf_win_ansi[i]; /* discard const */
}
}
}
/* symbolic Type 1 fonts with an implicit encoding and non-standard glyph names */
if (kind == TYPE1 && symbolic)
{
for (i = 0; i < 256; i++)
if (etable[i] && estrings[i] && !pdf_lookup_agl(estrings[i]))
estrings[i] = (char*) pdf_standard[i];
}
fz_unlock(ctx, FZ_LOCK_FREETYPE);
fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 1);
fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);
fontdesc->cid_to_gid_len = 256;
fontdesc->cid_to_gid = etable;
fz_try(ctx)
{
pdf_load_to_unicode(xref, fontdesc, estrings, NULL, pdf_dict_gets(dict, "ToUnicode"));
}
fz_catch(ctx)
{
fz_warn(ctx, "cannot load ToUnicode CMap");
}
skip_encoding:
/* Widths */
pdf_set_default_hmtx(ctx, fontdesc, fontdesc->missing_width);
widths = pdf_dict_gets(dict, "Widths");
if (widths)
{
int first, last;
first = pdf_to_int(pdf_dict_gets(dict, "FirstChar"));
last = pdf_to_int(pdf_dict_gets(dict, "LastChar"));
if (first < 0 || last > 255 || first > last)
first = last = 0;
for (i = 0; i < last - first + 1; i++)
{
int wid = pdf_to_int(pdf_array_get(widths, i));
pdf_add_hmtx(ctx, fontdesc, i + first, i + first, wid);
}
}
else
{
fz_lock(ctx, FZ_LOCK_FREETYPE);
fterr = FT_Set_Char_Size(face, 1000, 1000, 72, 72);
if (fterr)
fz_warn(ctx, "freetype set character size: %s", ft_error_string(fterr));
for (i = 0; i < 256; i++)
{
pdf_add_hmtx(ctx, fontdesc, i, i, ft_width(ctx, fontdesc, i));
}
fz_unlock(ctx, FZ_LOCK_FREETYPE);
}
pdf_end_hmtx(ctx, fontdesc);
}
fz_catch(ctx)
{
if (fontdesc && etable != fontdesc->cid_to_gid)
fz_free(ctx, etable);
pdf_drop_font(ctx, fontdesc);
fz_throw(ctx, "cannot load simple font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
}
return fontdesc;
}
char *
pdf_parse_link_dest(fz_context *ctx, pdf_document *doc, pdf_obj *dest)
{
pdf_obj *obj, *pageobj;
fz_rect mediabox;
fz_matrix pagectm;
const char *ld;
int page, x, y, h;
dest = resolve_dest(ctx, doc, dest);
if (dest == NULL)
{
fz_warn(ctx, "undefined link destination");
return NULL;
}
if (pdf_is_name(ctx, dest))
{
ld = pdf_to_name(ctx, dest);
return fz_strdup(ctx, ld);
}
else if (pdf_is_string(ctx, dest))
{
ld = pdf_to_str_buf(ctx, dest);
return fz_strdup(ctx, ld);
}
pageobj = pdf_array_get(ctx, dest, 0);
if (pdf_is_int(ctx, pageobj))
{
page = pdf_to_int(ctx, pageobj);
pageobj = pdf_lookup_page_obj(ctx, doc, page);
}
else
{
fz_try(ctx)
page = pdf_lookup_page_number(ctx, doc, pageobj);
fz_catch(ctx)
page = -1;
}
if (page < 0)
return NULL;
obj = pdf_array_get(ctx, dest, 1);
if (obj)
{
/* Link coords use a coordinate space that does not seem to respect Rotate or UserUnit. */
/* All we need to do is figure out the page height to flip the coordinate space. */
pdf_page_obj_transform(ctx, pageobj, &mediabox, &pagectm);
mediabox = fz_transform_rect(mediabox, pagectm);
h = mediabox.y1 - mediabox.y0;
if (pdf_name_eq(ctx, obj, PDF_NAME(XYZ)))
{
x = pdf_array_get_int(ctx, dest, 2);
y = h - pdf_array_get_int(ctx, dest, 3);
}
else if (pdf_name_eq(ctx, obj, PDF_NAME(FitR)))
{
x = pdf_array_get_int(ctx, dest, 2);
y = h - pdf_array_get_int(ctx, dest, 5);
}
else if (pdf_name_eq(ctx, obj, PDF_NAME(FitH)) || pdf_name_eq(ctx, obj, PDF_NAME(FitBH)))
{
x = 0;
y = h - pdf_array_get_int(ctx, dest, 2);
}
else if (pdf_name_eq(ctx, obj, PDF_NAME(FitV)) || pdf_name_eq(ctx, obj, PDF_NAME(FitBV)))
{
x = pdf_array_get_int(ctx, dest, 2);
y = 0;
}
else
{
x = 0;
y = 0;
}
return fz_asprintf(ctx, "#%d,%d,%d", page + 1, x, y);
}
return fz_asprintf(ctx, "#%d", page + 1);
}
