static void info_update(fz_context *ctx,pdf_document *xref,char *producer)
{
char moddate[64];
time_t now;
struct tm date;
pdf_obj *info;
int newinfo;
if (xref->trailer==NULL)
return;
time(&now);
date=(*localtime(&now));
sprintf(moddate,"D:%04d%02d%02d%02d%02d%02d%s",
date.tm_year+1900,date.tm_mon+1,date.tm_mday,
date.tm_hour,date.tm_min,date.tm_sec,
wsys_utc_string());
info=pdf_dict_gets(xref->trailer,"Info");
if (info==NULL)
{
newinfo=1;
info=pdf_new_dict(ctx,2);
}
else
newinfo=0;
dict_put_string(ctx,info,"Producer",producer);
dict_put_string(ctx,info,"ModDate",moddate);
if (newinfo)
{
pdf_dict_puts(xref->trailer,"Info",info);
pdf_drop_obj(info);
}
}
pdf_obj *
pdf_copy_dict(fz_context *ctx, pdf_obj *obj)
{
pdf_document *doc;
pdf_obj *dict;
int i, n;
RESOLVE(obj);
if (!OBJ_IS_DICT(obj))
fz_throw(ctx, FZ_ERROR_GENERIC, "not a dict (%s)", pdf_objkindstr(obj));
doc = DICT(obj)->doc;
n = pdf_dict_len(ctx, obj);
dict = pdf_new_dict(ctx, doc, n);
fz_try(ctx)
for (i = 0; i < n; i++)
pdf_dict_put(ctx, dict, pdf_dict_get_key(ctx, obj, i), pdf_dict_get_val(ctx, obj, i));
fz_catch(ctx)
{
pdf_drop_obj(ctx, dict);
fz_rethrow(ctx);
}
return dict;
}
/* Creates the PDF Encoding entry for the encoding.
* If baseenc is non-null, it is used as BaseEncoding entry.
*/
static pdf_obj *
create_encoding_resource (pdf_encoding *encoding, pdf_encoding *baseenc)
{
pdf_obj *differences;
ASSERT(encoding);
ASSERT(!encoding->resource);
differences = make_encoding_differences(encoding->glyphs,
baseenc ? baseenc->glyphs : NULL,
encoding->is_used);
if (differences) {
pdf_obj *resource = pdf_new_dict();
if (baseenc)
pdf_add_dict(resource, pdf_new_name("BaseEncoding"),
pdf_link_obj(baseenc->resource));
pdf_add_dict(resource, pdf_new_name("Differences"), differences);
return resource;
} else {
/* Fix a bug with the MinionPro package using MnSymbol fonts
* in its virtual fonts:
*
* Some font may have font_id even if no character is used.
* For example, suppose that a virtual file A.vf uses two
* other fonts, B and C. Even if only characters of B are used
* in a DVI document, C will have font_id too.
* In this case, both baseenc and differences can be NULL.
*
* Actually these fonts will be ignored in pdffont.c.
*/
return baseenc ? pdf_link_obj(baseenc->resource) : NULL;
}
}
static void
pdf_clean_stream_object(fz_context *ctx, pdf_document *doc, pdf_obj *obj, pdf_obj *orig_res, fz_cookie *cookie, int own_res,
pdf_text_filter_fn *text_filter, pdf_after_text_object_fn *after_text, void *arg,
int sanitize, int ascii)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *res = NULL;
pdf_obj *ref;
fz_buffer *buffer;
if (!obj)
return;
fz_var(res);
fz_var(proc_buffer);
fz_var(proc_filter);
buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
if (own_res)
{
pdf_obj *r = pdf_dict_get(ctx, obj, PDF_NAME(Resources));
if (r)
orig_res = r;
}
res = pdf_new_dict(ctx, doc, 1);
proc_buffer = pdf_new_buffer_processor(ctx, buffer, ascii);
proc_filter = pdf_new_filter_processor_with_text_filter(ctx, doc, proc_buffer, orig_res, res, text_filter, after_text, arg);
pdf_process_contents(ctx, proc_filter, doc, orig_res, obj, cookie);
pdf_close_processor(ctx, proc_filter);
pdf_close_processor(ctx, proc_buffer);
pdf_update_stream(ctx, doc, obj, buffer, 0);
if (own_res)
{
ref = pdf_add_object(ctx, doc, res);
pdf_dict_put_drop(ctx, obj, PDF_NAME(Resources), ref);
}
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
pdf_drop_obj(ctx, res);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
pdf_page *
pdf_create_page(pdf_document *doc, fz_rect mediabox, int res, int rotate)
{
pdf_page *page = NULL;
pdf_obj *pageobj;
float userunit = 1;
fz_context *ctx = doc->ctx;
fz_matrix ctm, tmp;
fz_rect realbox;
page = fz_malloc_struct(ctx, pdf_page);
fz_try(ctx)
{
page->resources = NULL;
page->contents = NULL;
page->transparency = 0;
page->links = NULL;
page->annots = NULL;
page->me = pageobj = pdf_new_dict(doc, 4);
pdf_dict_puts_drop(pageobj, "Type", pdf_new_name(doc, "Page"));
page->mediabox.x0 = fz_min(mediabox.x0, mediabox.x1) * userunit;
page->mediabox.y0 = fz_min(mediabox.y0, mediabox.y1) * userunit;
page->mediabox.x1 = fz_max(mediabox.x0, mediabox.x1) * userunit;
page->mediabox.y1 = fz_max(mediabox.y0, mediabox.y1) * userunit;
pdf_dict_puts_drop(pageobj, "MediaBox", pdf_new_rect(doc, &page->mediabox));
/* Snap page->rotate to 0, 90, 180 or 270 */
if (page->rotate < 0)
page->rotate = 360 - ((-page->rotate) % 360);
if (page->rotate >= 360)
page->rotate = page->rotate % 360;
page->rotate = 90*((page->rotate + 45)/90);
if (page->rotate > 360)
page->rotate = 0;
pdf_dict_puts_drop(pageobj, "Rotate", pdf_new_int(doc, page->rotate));
fz_pre_rotate(fz_scale(&ctm, 1, -1), -page->rotate);
realbox = page->mediabox;
fz_transform_rect(&realbox, &ctm);
fz_pre_scale(fz_translate(&tmp, -realbox.x0, -realbox.y0), userunit, userunit);
fz_concat(&ctm, &ctm, &tmp);
page->ctm = ctm;
/* Do not create a Contents, as an empty Contents dict is not
* valid. See Bug 694712 */
}
fz_catch(ctx)
{
pdf_drop_obj(page->me);
fz_free(ctx, page);
fz_rethrow_message(ctx, "Failed to create page");
}
return page;
}
static void
pdf_clean_stream_object(fz_context *ctx, pdf_document *doc, pdf_obj *obj, pdf_obj *orig_res, fz_cookie *cookie, int own_res)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *res = NULL;
pdf_obj *ref = NULL;
fz_buffer *buffer;
if (!obj)
return;
fz_var(res);
fz_var(ref);
fz_var(proc_buffer);
fz_var(proc_filter);
buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
if (own_res)
{
pdf_obj *r = pdf_dict_get(ctx, obj, PDF_NAME_Resources);
if (r)
orig_res = r;
}
res = pdf_new_dict(ctx, doc, 1);
proc_buffer = pdf_new_buffer_processor(ctx, buffer);
proc_filter = pdf_new_filter_processor(ctx, proc_buffer, doc, orig_res, res);
pdf_process_contents(ctx, proc_filter, doc, orig_res, obj, cookie);
pdf_update_stream(ctx, doc, obj, buffer, 0);
if (own_res)
{
ref = pdf_new_ref(ctx, doc, res);
pdf_dict_put(ctx, obj, PDF_NAME_Resources, ref);
}
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
pdf_drop_obj(ctx, res);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow_message(ctx, "Failed while cleaning xobject");
}
}
static void
pdf_dev_alpha(fz_context *ctx, pdf_device *pdev, float alpha, int stroke)
{
int i;
pdf_document *doc = pdev->doc;
gstate *gs = CURRENT_GSTATE(pdev);
/* If the alpha is unchanged, nothing to do */
if (gs->alpha[stroke] == alpha)
return;
/* Have we sent such an alpha before? */
for (i = 0; i < pdev->num_alphas; i++)
if (pdev->alphas[i].alpha == alpha && pdev->alphas[i].stroke == stroke)
break;
if (i == pdev->num_alphas)
{
pdf_obj *o;
pdf_obj *ref = NULL;
fz_var(ref);
/* No. Need to make a new one */
if (pdev->num_alphas == pdev->max_alphas)
{
int newmax = pdev->max_alphas * 2;
if (newmax == 0)
newmax = 4;
pdev->alphas = fz_resize_array(ctx, pdev->alphas, newmax, sizeof(*pdev->alphas));
pdev->max_alphas = newmax;
}
pdev->alphas[i].alpha = alpha;
pdev->alphas[i].stroke = stroke;
o = pdf_new_dict(ctx, doc, 1);
fz_try(ctx)
{
char text[32];
pdf_dict_put_drop(ctx, o, (stroke ? PDF_NAME_CA : PDF_NAME_ca), pdf_new_real(ctx, doc, alpha));
ref = pdf_add_object(ctx, doc, o);
fz_snprintf(text, sizeof(text), "ExtGState/Alp%d", i);
pdf_dict_putp(ctx, pdev->resources, text, ref);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, o);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
pdev->num_alphas++;
}
static void page_merge(int page_from, int page_to, pdf_graft_map *graft_map)
{
pdf_obj *pageref = NULL;
pdf_obj *page_dict;
pdf_obj *obj = NULL, *ref = NULL;
/* Include minimal number of objects for page. Do not include items that
* reference other pages */
pdf_obj *known_page_objs[] = { PDF_NAME_Contents, PDF_NAME_Resources,
PDF_NAME_MediaBox, PDF_NAME_CropBox, PDF_NAME_BleedBox, PDF_NAME_TrimBox,
PDF_NAME_ArtBox, PDF_NAME_Rotate, PDF_NAME_UserUnit};
int n = nelem(known_page_objs);
int i;
int num;
fz_var(obj);
fz_var(ref);
fz_try(ctx)
{
pageref = pdf_lookup_page_obj(ctx, doc_src, page_from - 1);
/* Make a new dictionary and copy over the items from the source object to
* the new dict that we want to deep copy. */
page_dict = pdf_new_dict(ctx, doc_des, 4);
pdf_dict_put_drop(ctx, page_dict, PDF_NAME_Type, PDF_NAME_Page);
for (i = 0; i < n; i++)
{
obj = pdf_dict_get(ctx, pageref, known_page_objs[i]);
if (obj != NULL)
pdf_dict_put_drop(ctx, page_dict, known_page_objs[i], pdf_graft_object(ctx, doc_des, doc_src, obj, graft_map));
}
/* Add the dictionary */
obj = pdf_add_object_drop(ctx, doc_des, page_dict);
/* Get indirect ref */
num = pdf_to_num(ctx, obj);
ref = pdf_new_indirect(ctx, doc_des, num, 0);
/* Insert */
pdf_insert_page(ctx, doc_des, page_to - 1, ref);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, obj);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
pdf_obj *parse_pdf_dict (char **start, char *end)
{
pdf_obj *result, *tmp1, *tmp2;
char *save = *start;
skip_white(start, end);
if (*((*start)++) != '<' ||
*((*start)++) != '<') {
*start = save;
dump (*start, end);
return NULL;
}
result = pdf_new_dict ();
skip_white(start, end);
while (*start < end &&
**start != '>') {
if ((tmp1 = parse_pdf_name (start, end)) == NULL) {
pdf_release_obj (result);
{
*start = save;
dump (*start, end);
return NULL;
}
};
if ((tmp2 = parse_pdf_object (start, end)) == NULL) {
pdf_release_obj (result);
pdf_release_obj (tmp1);
{
*start = save;
dump (*start, end);
return NULL;
}
}
pdf_add_dict (result, tmp1, tmp2);
skip_white(start, end);
}
if (*start >= end) {
pdf_release_obj (result);
*start = save;
dump (*start, end);
return NULL;
}
if (*((*start)++) == '>' &&
*((*start)++) == '>') {
return result;
} else {
pdf_release_obj (result);
fprintf (stderr, "\nDictionary object ended prematurely\n");
*start = save;
dump (*start, end);
return NULL;
}
}
static void html_make_link_dict (char *name)
{
pdf_obj *color;
if (!link_dict) {
link_dict = pdf_new_dict();
pdf_add_dict(link_dict, pdf_new_name("Type"), pdf_new_name ("Annot"));
pdf_add_dict(link_dict, pdf_new_name("Subtype"), pdf_new_name ("Link"));
color = pdf_new_array ();
pdf_add_array (color, pdf_new_number (0));
pdf_add_array (color, pdf_new_number (1));
pdf_add_array (color, pdf_new_number (1));
pdf_add_dict(link_dict, pdf_new_name("C"), color);
if (name && *name == '#' && !(base_value)) {
pdf_add_dict (link_dict, pdf_new_name("Dest"), pdf_new_string(name+1,strlen(name+1)));
} else if (name) { /* Assume its a URL */
char *url;
int len;
pdf_obj *action;
len = strlen(name)+1;
if (base_value)
len+=strlen(base_value);
url = NEW (len, char);
if (base_value)
strcpy (url, base_value);
else
url[0] = 0;
strcat (url, name);
action = pdf_new_dict();
pdf_add_dict (action, pdf_new_name ("Type"), pdf_new_name ("Action"));
pdf_add_dict (action, pdf_new_name ("S"), pdf_new_name ("URI"));
pdf_add_dict (action, pdf_new_name ("URI"),
pdf_new_string (url, len));
pdf_add_dict (link_dict, pdf_new_name ("A"), pdf_ref_obj (action));
pdf_release_obj (action);
RELEASE (url);
}
pdf_doc_begin_annot (link_dict);
} else {
/*
** From MuPDF pdfclean.c
*/
static void wmupdf_preserve_old_dests(pdf_obj *olddests,fz_context *ctx,pdf_document *xref,
pdf_obj *pages)
{
int i;
pdf_obj *names = pdf_new_dict(ctx,1);
pdf_obj *dests = pdf_new_dict(ctx,1);
pdf_obj *names_list = pdf_new_array(ctx,32);
int len = pdf_dict_len(olddests);
pdf_obj *root;
for (i=0;i<len;i++)
{
pdf_obj *key = pdf_dict_get_key(olddests,i);
pdf_obj *val = pdf_dict_get_val(olddests,i);
pdf_obj *key_str = pdf_new_string(ctx,pdf_to_name(key),strlen(pdf_to_name(key)));
pdf_obj *dest = pdf_dict_gets(val,"D");
dest = pdf_array_get(dest ? dest : val, 0);
if (pdf_array_contains(pdf_dict_gets(pages,"Kids"),dest))
{
pdf_array_push(names_list, key_str);
pdf_array_push(names_list, val);
}
pdf_drop_obj(key_str);
}
root = pdf_dict_gets(xref->trailer,"Root");
pdf_dict_puts(dests,"Names",names_list);
pdf_dict_puts(names,"Dests",dests);
pdf_dict_puts(root,"Names",names);
pdf_drop_obj(names);
pdf_drop_obj(dests);
pdf_drop_obj(names_list);
pdf_drop_obj(olddests);
}
static pdf_obj *
hash_to_pdf_dict (fz_context *ctx, pdf_document *doc, pdfout_data *hash,
int *page)
{
pdf_obj *dict_obj = pdf_new_dict (ctx, doc, 3);
int len = pdfout_data_hash_len (ctx, hash);
for (int j = 0; j < len; ++j)
{
char *key, *value;
int value_len;
pdfout_data_hash_get_key_value (ctx, hash, &key, &value, &value_len, j);
if (streq (key, "page"))
*page = pdfout_strtoint_null (ctx, value);
else if (streq (key, "prefix"))
{
pdf_obj *string = pdfout_utf8_to_str_obj (ctx, doc, value,
value_len);
pdf_dict_puts_drop (ctx, dict_obj, "P", string);
}
else if (streq (key, "first"))
{
int first = pdfout_strtoint_null (ctx, value);
pdf_dict_puts_drop (ctx, dict_obj, "St",
pdf_new_int (ctx, doc, first));
}
else if (streq (key, "style"))
{
const char *name;
if (streq (value, "arabic"))
name = "D";
else if (streq (value, "Roman"))
name = "R";
else if (streq (value, "roman"))
name = "r";
else if (streq (value, "Letters"))
name = "A";
else if (streq (value, "letters"))
name = "a";
else
abort ();
pdf_obj *name_obj = pdf_new_name (ctx, doc, name);
pdf_dict_puts_drop (ctx, dict_obj, "S", name_obj);
}
}
return dict_obj;
}
/*
put_pages_on_new_sheet() - put the contents of pages onto one page in dest_doc
How does it work?
1. Create a page for the new sheet that will contain all source-pages
2. Analyze their sizes! If too small, just center. If too big, warn
3. Analye their userunit-value... if one of them != 1 we have to insert
scale-operations to all others and need to use userunit in the sheet too
4. Copy each entry of the resource-dicts of every page of the sheet
(if it has not already been copied for a previous sheet)
5. Create a dict that contains src-resource-name to dest-resource-name for
this sheet (this is an n:m relation where n >= m)
6. For each source-page-stream create a new content-stream and change
scale (if user-unit is needed), translation and rotation... And add a
clipping-path around the page
7. Copy all source content-streams of the current page to this new content-
stream renaming all its resources to the new names
8. TODO: Annots?
9. TODO: Preseparated pages
10. TODO: Merge procedure sets
*/
static int put_pages_on_new_sheet(fz_context *dest_ctx, pdf_document *dest_doc,
fz_context *src_ctx, pdf_document *src_doc,
struct pos_info *positions, size_t put_count)
{
/* copy each entry of the resource dict */
static const int RENAME_INITIAL_CAP = 128;
struct put_info put_info = { dest_doc, src_doc, NULL, NULL, 0 };
/* what destianation page is currently opened? */
int sheet_pagenum = -1;
pdf_page *sheet = NULL;
size_t i;
for(i = 0; i < put_count; i++) {
/* if the current sheet changes, we need to close the current and open
the new one*/
if(sheet_pagenum != positions[i].sheet_pagenum) {
if(sheet != NULL)
pdf_drop_page(dest_ctx, sheet);
sheet_pagenum = positions[i].sheet_pagenum;
sheet = pdf_load_page(dest_ctx, dest_doc, sheet_pagenum);
// TODO: We assume that the destination-page uses arrays for the
// content and has a usable resoruce-dict... TODO: add those checks
}
/* load the source-page */
pdf_page *src_page = pdf_load_page(src_ctx, src_doc, positions[i].src_pagenum);
/* create a rename_dict */
put_info.rename_dict = pdf_new_dict(dest_ctx, dest_doc, RENAME_INITIAL_CAP);
/* copy all resources, adjust the page and finally copy the content */
copy_and_rename_resources(dest_ctx, sheet->resources,
src_ctx, src_page->resources, &put_info);
adjust_page_position(src_ctx, src_doc, src_page, positions + i);
copy_content_streams_of_page(
dest_ctx, sheet, src_ctx, src_page, &put_info, positions + i);
/* free everything we created for that source-page */
pdf_drop_obj(dest_ctx, put_info.rename_dict);
pdf_drop_page(src_ctx, src_page);
}
if(sheet != NULL)
pdf_drop_page(dest_ctx, sheet);
return(0);
}
static pdf_obj *start_new_destpage(fz_context *ctx,double width_pts,double height_pts)
{
pdf_obj *pageobj;
pdf_obj *mbox;
pageobj=pdf_new_dict(ctx,2);
pdf_dict_puts(pageobj,"Type",pdf_new_name(ctx,"Page"));
mbox=pdf_new_array(ctx,4);
pdf_array_push(mbox,pdf_new_real(ctx,0.));
pdf_array_push(mbox,pdf_new_real(ctx,0.));
pdf_array_push(mbox,pdf_new_real(ctx,width_pts));
pdf_array_push(mbox,pdf_new_real(ctx,height_pts));
pdf_dict_puts(pageobj,"MediaBox",mbox);
return(pageobj);
}
pdf_obj *
pdf_load_name_tree(pdf_document *xref, char *which)
{
fz_context *ctx = xref->ctx;
pdf_obj *root = pdf_dict_gets(xref->trailer, "Root");
pdf_obj *names = pdf_dict_gets(root, "Names");
pdf_obj *tree = pdf_dict_gets(names, which);
if (pdf_is_dict(tree))
{
pdf_obj *dict = pdf_new_dict(ctx, 100);
pdf_load_name_tree_imp(dict, xref, tree);
return dict;
}
return NULL;
}
static void page_merge(int page_from, int page_to, pdf_graft_map *graft_map)
{
pdf_obj *page_ref;
pdf_obj *page_dict;
pdf_obj *obj;
pdf_obj *ref = NULL;
int i;
/* Copy as few key/value pairs as we can. Do not include items that reference other pages. */
static pdf_obj * const copy_list[] = { PDF_NAME(Contents), PDF_NAME(Resources),
PDF_NAME(MediaBox), PDF_NAME(CropBox), PDF_NAME(BleedBox), PDF_NAME(TrimBox), PDF_NAME(ArtBox),
PDF_NAME(Rotate), PDF_NAME(UserUnit) };
fz_var(ref);
fz_try(ctx)
{
page_ref = pdf_lookup_page_obj(ctx, doc_src, page_from - 1);
pdf_flatten_inheritable_page_items(ctx, page_ref);
/* Make a new page object dictionary to hold the items we copy from the source page. */
page_dict = pdf_new_dict(ctx, doc_des, 4);
pdf_dict_put(ctx, page_dict, PDF_NAME(Type), PDF_NAME(Page));
for (i = 0; i < nelem(copy_list); i++)
{
obj = pdf_dict_get(ctx, page_ref, copy_list[i]);
if (obj != NULL)
pdf_dict_put_drop(ctx, page_dict, copy_list[i], pdf_graft_mapped_object(ctx, graft_map, obj));
}
/* Add the page object to the destination document. */
ref = pdf_add_object_drop(ctx, doc_des, page_dict);
/* Insert it into the page tree. */
pdf_insert_page(ctx, doc_des, page_to - 1, ref);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
static ErrorCode juggler_impose_create_sheet(fz_context *ctx, pdf_document *dest_doc, int width, int height)
{
/* create and insert page to dest_doc */
fz_rect rect = { 0, 0, width, height };// { 0, 0, 1390, 1684 };
pdf_page *sheet = pdf_create_page(ctx, dest_doc, rect, 0, 0);
sheet->resources = pdf_new_dict(ctx, dest_doc, 16);
pdf_dict_puts(ctx, sheet->me, "Resources", sheet->resources); /* will be droped when freeing the page */
sheet->contents = pdf_new_array(ctx, dest_doc, 8);
pdf_dict_puts(ctx, sheet->me, "Contents", sheet->contents); /* will be droped when freeing the page */
pdf_insert_page(ctx, dest_doc, sheet, INT_MAX);
pdf_drop_page(ctx, sheet);
return(NoError);
}
static void insert_resource_name(pdf_csi *csi, pdf_filter_state *state, const char *key, const char *name)
{
pdf_obj *xobj;
pdf_obj *obj;
if (!state->resources || !name || name[0] == 0)
return;
xobj = pdf_dict_gets(csi->rdb, key);
obj = pdf_dict_gets(xobj, name);
xobj = pdf_dict_gets(state->resources, key);
if (xobj == NULL) {
xobj = pdf_new_dict(csi->doc, 1);
pdf_dict_puts_drop(state->resources, key, xobj);
}
pdf_dict_putp(xobj, name, obj);
}
pdf_obj *
pdf_copy_dict(fz_context *ctx, pdf_obj *obj)
{
pdf_obj *dict;
int i, n;
RESOLVE(obj);
if (!obj)
return NULL; /* Can't warn :( */
if (obj->kind != PDF_DICT)
fz_warn(ctx, "assert: not a dict (%s)", pdf_objkindstr(obj));
n = pdf_dict_len(obj);
dict = pdf_new_dict(ctx, n);
for (i = 0; i < n; i++)
fz_dict_put(dict, pdf_dict_get_key(obj, i), pdf_dict_get_val(obj, i));
return dict;
}
static int new_stream_object(pdf_document *xref,fz_context *ctx,char *buf)
{
int ref;
pdf_obj *obj,*len;
fz_buffer *fzbuf;
ref = pdf_create_object(xref);
obj = pdf_new_dict(ctx,1);
len=pdf_new_int(ctx,strlen(buf));
pdf_dict_puts(obj,"Length",len);
pdf_drop_obj(len);
pdf_update_object(xref,ref,obj);
pdf_drop_obj(obj);
fzbuf=fz_new_buffer(ctx,strlen(buf));
fz_write_buffer(ctx,fzbuf,(unsigned char *)buf,strlen(buf));
pdf_update_stream(xref,ref,fzbuf);
fz_drop_buffer(ctx,fzbuf);
return(ref);
}
static void
copy_resource(fz_context *ctx, pdf_filter_processor *p, pdf_obj *key, const char *name)
{
pdf_obj *res, *obj;
if (!name || name[0] == 0)
return;
res = pdf_dict_get(ctx, p->old_rdb, key);
obj = pdf_dict_gets(ctx, res, name);
if (obj)
{
res = pdf_dict_get(ctx, p->new_rdb, key);
if (!res)
{
res = pdf_new_dict(ctx, p->doc, 1);
pdf_dict_put_drop(ctx, p->new_rdb, key, res);
}
pdf_dict_putp(ctx, res, name, obj);
}
}
pdf_obj *
pdf_new_xobject(fz_context *ctx, pdf_document *doc, fz_rect bbox, fz_matrix matrix, pdf_obj *res, fz_buffer *contents)
{
pdf_obj *ind = NULL;
pdf_obj *form = pdf_new_dict(ctx, doc, 5);
fz_try(ctx)
{
pdf_dict_put(ctx, form, PDF_NAME(Type), PDF_NAME(XObject));
pdf_dict_put(ctx, form, PDF_NAME(Subtype), PDF_NAME(Form));
pdf_dict_put_rect(ctx, form, PDF_NAME(BBox), bbox);
pdf_dict_put_matrix(ctx, form, PDF_NAME(Matrix), matrix);
if (res)
pdf_dict_put(ctx, form, PDF_NAME(Resources), res);
ind = pdf_add_stream(ctx, doc, contents, form, 0);
}
fz_always(ctx)
pdf_drop_obj(ctx, form);
fz_catch(ctx)
fz_rethrow(ctx);
return ind;
}
void
pdfout_page_labels_set (fz_context *ctx, pdf_document *doc,
pdfout_data *labels)
{
if (labels)
check_page_labels (ctx, labels);
pdf_obj *root = pdf_dict_get (ctx, pdf_trailer (ctx, doc), PDF_NAME_Root);
if (root == NULL)
pdfout_throw (ctx, "no document catalog, cannot set/unset page labels");
if (labels == NULL)
{
/* Remove page labels. */
pdf_dict_dels (ctx, root, "PageLabels");
return;
}
int num = pdfout_data_array_len (ctx, labels);
pdf_obj *array_obj = pdf_new_array (ctx, doc, 2 * num);
for (int i = 0; i < num; ++i)
{
pdfout_data *hash = pdfout_data_array_get (ctx, labels, i);
int page;
pdf_obj *dict_obj = hash_to_pdf_dict (ctx, doc, hash, &page);
pdf_obj *page_obj = pdf_new_int (ctx, doc, page);
pdf_array_push_drop (ctx, array_obj, page_obj);
pdf_array_push_drop (ctx, array_obj, dict_obj);
}
pdf_obj *labels_obj = pdf_new_dict (ctx, doc, 1);
pdf_dict_puts_drop (ctx, labels_obj, "Nums", array_obj);
pdf_dict_puts_drop (ctx, root, "PageLabels", labels_obj);
}
pdf_obj *
pdf_copy_dict(fz_context *ctx, pdf_obj *obj)
{
pdf_obj *dict;
int i, n;
RESOLVE(obj);
if (obj >= PDF_OBJ__LIMIT)
{
pdf_document *doc = DICT(obj)->doc;
if (obj->kind != PDF_DICT)
fz_warn(ctx, "assert: not a dict (%s)", pdf_objkindstr(obj));
n = pdf_dict_len(ctx, obj);
dict = pdf_new_dict(ctx, doc, n);
for (i = 0; i < n; i++)
pdf_dict_put(ctx, dict, pdf_dict_get_key(ctx, obj, i), pdf_dict_get_val(ctx, obj, i));
return dict;
}
return NULL; /* Can't warn :( */
}
static int copy_and_rename_resources(fz_context *dest_ctx, pdf_obj *dest,
fz_context *src_ctx, pdf_obj *src, struct put_info *info)
{
static char *RESOURCE_TYPES[] =
{ "ExtGState", "ColorSpace", "Pattern", "Shading", "XObject",
"Font", "Properties" };
static char *RESOURCE_PREFIXES[] =
{ "E_", "C_", "P_", "S_", "X_", "F_", "T_" };
static const size_t RESOURCE_TYPES_COUNT =
sizeof(RESOURCE_TYPES) / sizeof(RESOURCE_TYPES[0]);
size_t i;
for(i = 0; i < RESOURCE_TYPES_COUNT; i++) {
pdf_obj *src_type = pdf_dict_gets(src_ctx, src, RESOURCE_TYPES[i]);
pdf_obj *dest_type = pdf_dict_gets(dest_ctx, dest, RESOURCE_TYPES[i]);
/* we only copy resource-dicts that exist in the source-page ;) */
if(pdf_is_dict(src_ctx, src_type)) {printf("TODO: REMOVEME: Copying and renaming resources of type %s!!!\n", RESOURCE_TYPES[i]);
/* if this kind of dict does not exists in the dest resources,
we must create it */
if(!pdf_is_dict(dest_ctx, dest_type)) {
dest_type = pdf_new_dict(dest_ctx, info->dest_doc, 8);
pdf_dict_puts_drop(dest_ctx, dest, RESOURCE_TYPES[i], dest_type);
}
if(copy_and_rename_resource(dest_ctx, dest_type,
src_ctx, src_type, RESOURCE_PREFIXES[i], info))
{
return(2);
}
}
}
// TODO: Merge Procedure-Sets (although they are obsolete)
return(0);
}
static void
pdf_clean_type3(fz_context *ctx, pdf_document *doc, pdf_obj *obj, pdf_obj *orig_res, fz_cookie *cookie, int sanitize, int ascii)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *res = NULL;
pdf_obj *ref;
pdf_obj *charprocs;
int i, l;
fz_var(res);
fz_var(proc_buffer);
fz_var(proc_filter);
fz_try(ctx)
{
res = pdf_dict_get(ctx, obj, PDF_NAME(Resources));
if (res)
orig_res = res;
res = NULL;
res = pdf_new_dict(ctx, doc, 1);
charprocs = pdf_dict_get(ctx, obj, PDF_NAME(CharProcs));
l = pdf_dict_len(ctx, charprocs);
for (i = 0; i < l; i++)
{
pdf_obj *val = pdf_dict_get_val(ctx, charprocs, i);
fz_buffer *buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
proc_buffer = pdf_new_buffer_processor(ctx, buffer, ascii);
if (sanitize)
{
proc_filter = pdf_new_filter_processor(ctx, doc, proc_buffer, orig_res, res);
pdf_process_contents(ctx, proc_filter, doc, orig_res, val, cookie);
pdf_close_processor(ctx, proc_filter);
}
else
{
pdf_process_contents(ctx, proc_filter, doc, orig_res, val, cookie);
}
pdf_close_processor(ctx, proc_buffer);
pdf_update_stream(ctx, doc, val, buffer, 0);
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
/* ProcSet - no cleaning possible. Inherit this from the old dict. */
pdf_dict_put(ctx, res, PDF_NAME(ProcSet), pdf_dict_get(ctx, orig_res, PDF_NAME(ProcSet)));
ref = pdf_add_object(ctx, doc, res);
pdf_dict_put_drop(ctx, obj, PDF_NAME(Resources), ref);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, res);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
int
pdf_font_load_pkfont (pdf_font *font)
{
pdf_obj *fontdict;
char *usedchars;
char *ident;
unsigned dpi;
FILE *fp;
double point_size, pix2charu;
int opcode, code, firstchar, lastchar, prev;
pdf_obj *charprocs, *procset, *encoding, *tmp_array;
double widths[256];
pdf_rect bbox;
char charavail[256];
#if ENABLE_GLYPHENC
int encoding_id;
char **enc_vec;
#endif /* ENABLE_GLYPHENC */
int error = 0;
if (!pdf_font_is_in_use(font)) {
return 0;
}
ident = pdf_font_get_ident(font);
point_size = pdf_font_get_param(font, PDF_FONT_PARAM_POINT_SIZE);
usedchars = pdf_font_get_usedchars(font);
#if ENABLE_GLYPHENC
encoding_id = pdf_font_get_encoding(font);
if (encoding_id < 0)
enc_vec = NULL;
else {
enc_vec = pdf_encoding_get_encoding(encoding_id);
}
#endif /* ENABLE_GLYPHENC */
ASSERT(ident && usedchars && point_size > 0.0);
dpi = truedpi(ident, point_size, base_dpi);
{
char *fontfile = pdf_font_get_fontfile (font);
if (fontfile)
fp = MFOPEN(fontfile, FOPEN_RBIN_MODE);
else
fp = dpx_open_pk_font_at(ident, dpi);
}
if (!fp) {
ERROR("Could not find/open PK font file: %s (at %udpi)", ident, dpi);
}
memset(charavail, 0, 256);
charprocs = pdf_new_dict();
/* Include bitmap as 72dpi image:
* There seems to be problems in "scaled" bitmap glyph
* rendering in several viewers.
*/
pix2charu = 72. * 1000. / ((double) base_dpi) / point_size;
bbox.llx = bbox.lly = HUGE_VAL;
bbox.urx = bbox.ury = -HUGE_VAL;
while ((opcode = fgetc(fp)) >= 0 && opcode != PK_POST) {
if (opcode < 240) {
struct pk_header_ pkh;
error = read_pk_char_header(&pkh, opcode, fp);
if (error)
ERROR("Error in reading PK character header.");
else if (charavail[pkh.chrcode & 0xff])
WARN("More than two bitmap image for single glyph?: font=\"%s\" code=0x%02x",
ident, pkh.chrcode);
if (!usedchars[pkh.chrcode & 0xff])
do_skip(fp, pkh.pkt_len);
else {
char *charname;
pdf_obj *charproc;
unsigned char *pkt_ptr;
size_t bytesread;
double charwidth;
/* Charwidth in PDF units */
charwidth = ROUND(1000.0 * pkh.wd / (((double) (1<<20))*pix2charu), 0.1);
widths[pkh.chrcode & 0xff] = charwidth;
/* Update font BBox info */
bbox.llx = MIN(bbox.llx, -pkh.bm_hoff);
bbox.lly = MIN(bbox.lly, pkh.bm_voff - pkh.bm_ht);
bbox.urx = MAX(bbox.urx, pkh.bm_wd - pkh.bm_hoff);
bbox.ury = MAX(bbox.ury, pkh.bm_voff);
pkt_ptr = NEW(pkh.pkt_len, unsigned char);
if ((bytesread = fread(pkt_ptr, 1, pkh.pkt_len, fp))!= pkh.pkt_len) {
ERROR("Only %ld bytes PK packet read. (expected %ld bytes)",
bytesread, pkh.pkt_len);
}
charproc = create_pk_CharProc_stream(&pkh, charwidth, pkt_ptr, bytesread);
RELEASE(pkt_ptr);
if (!charproc)
ERROR("Unpacking PK character data failed.");
#if ENABLE_GLYPHENC
if (encoding_id >= 0 && enc_vec) {
charname = (char *) enc_vec[pkh.chrcode & 0xff];
if (!charname) {
WARN("\".notdef\" glyph used in font (code=0x%02x): %s", pkh.chrcode, ident);
charname = work_buffer;
pk_char2name(charname, pkh.chrcode);
}
}
else
#endif /* ENABLE_GLYPHENC */
{
charname = work_buffer;
pk_char2name(charname, pkh.chrcode);
}
pdf_add_dict(charprocs, pdf_new_name(charname), pdf_ref_obj(charproc)); /* _FIXME_ */
pdf_release_obj(charproc);
}
charavail[pkh.chrcode & 0xff] = 1;
} else { /* A command byte */
switch (opcode) {
case PK_NO_OP: break;
case PK_XXX1: do_skip(fp, get_unsigned_byte(fp)); break;
case PK_XXX2: do_skip(fp, get_unsigned_pair(fp)); break;
case PK_XXX3: do_skip(fp, get_unsigned_triple(fp)); break;
case PK_XXX4: do_skip(fp, get_unsigned_quad(fp)); break;
case PK_YYY: do_skip(fp, 4); break;
case PK_PRE: do_preamble(fp); break;
}
}
}
pdf_obj *
pdf_parse_dict(pdf_document *xref, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *dict;
pdf_obj *key = NULL;
pdf_obj *val = NULL;
int tok;
int a, b;
fz_context *ctx = file->ctx;
dict = pdf_new_dict(ctx, 8);
fz_var(key);
fz_var(val);
fz_try(ctx)
{
while (1)
{
tok = pdf_lex(file, buf);
skip:
if (tok == PDF_TOK_CLOSE_DICT)
break;
/* for BI .. ID .. EI in content streams */
if (tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID"))
break;
if (tok != PDF_TOK_NAME)
fz_throw(ctx, "invalid key in dict");
key = fz_new_name(ctx, buf->scratch);
tok = pdf_lex(file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1643 */
fz_try(ctx)
{
val = pdf_parse_array(xref, file, buf);
}
fz_catch(ctx)
{
fz_warn(ctx, "ignoring broken array for '%s'", pdf_to_name(key));
pdf_drop_obj(key);
val = key = NULL;
do
tok = pdf_lex(file, buf);
while (tok != PDF_TOK_CLOSE_DICT && tok != PDF_TOK_CLOSE_ARRAY &&
tok != PDF_TOK_EOF && tok != PDF_TOK_OPEN_ARRAY && tok != PDF_TOK_OPEN_DICT);
if (tok == PDF_TOK_CLOSE_DICT)
goto skip;
if (tok == PDF_TOK_CLOSE_ARRAY)
continue;
fz_throw(ctx, "cannot make sense of broken array after all");
}
break;
case PDF_TOK_OPEN_DICT:
val = pdf_parse_dict(xref, file, buf);
break;
case PDF_TOK_NAME:
val = fz_new_name(ctx, buf->scratch);
break;
case PDF_TOK_REAL:
val = pdf_new_real(ctx, buf->f);
break;
case PDF_TOK_STRING:
val = pdf_new_string(ctx, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
val = pdf_new_bool(ctx, 1);
break;
case PDF_TOK_FALSE:
val = pdf_new_bool(ctx, 0);
break;
case PDF_TOK_NULL:
val = pdf_new_null(ctx);
break;
case PDF_TOK_INT:
/* 64-bit to allow for numbers > INT_MAX and overflow */
a = buf->i;
tok = pdf_lex(file, buf);
if (tok == PDF_TOK_CLOSE_DICT || tok == PDF_TOK_NAME ||
(tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID")))
{
val = pdf_new_int(ctx, a);
fz_dict_put(dict, key, val);
pdf_drop_obj(val);
val = NULL;
pdf_drop_obj(key);
key = NULL;
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(file, buf);
if (tok == PDF_TOK_R)
{
val = pdf_new_indirect(ctx, a, b, xref);
break;
}
}
fz_throw(ctx, "invalid indirect reference in dict");
default:
fz_throw(ctx, "unknown token in dict");
}
fz_dict_put(dict, key, val);
pdf_drop_obj(val);
val = NULL;
pdf_drop_obj(key);
key = NULL;
}
}
fz_catch(ctx)
{
pdf_drop_obj(dict);
pdf_drop_obj(key);
pdf_drop_obj(val);
fz_throw(ctx, "cannot parse dict");
}
return dict;
}
/*
Performs the same task as
pdf_clean_page_contents, but with an optional text filter
function.
text_filter: Function to assess whether a given character
should be kept (return 0) or removed (return 1).
after_text: Function called after each text object is closed
to allow other output to be sent.
arg: Opaque value to be passed to callback functions.
*/
void pdf_filter_page_contents(fz_context *ctx, pdf_document *doc, pdf_page *page, fz_cookie *cookie,
pdf_page_contents_process_fn *proc_fn, pdf_text_filter_fn *text_filter, pdf_after_text_object_fn *after_text, void *proc_arg,
int sanitize, int ascii)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *new_obj = NULL;
pdf_obj *new_ref = NULL;
pdf_obj *res = NULL;
pdf_obj *obj;
pdf_obj *contents;
pdf_obj *resources;
fz_buffer *buffer;
fz_var(new_obj);
fz_var(new_ref);
fz_var(res);
fz_var(proc_buffer);
fz_var(proc_filter);
buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
contents = pdf_page_contents(ctx, page);
resources = pdf_page_resources(ctx, page);
proc_buffer = pdf_new_buffer_processor(ctx, buffer, ascii);
if (sanitize)
{
res = pdf_new_dict(ctx, doc, 1);
proc_filter = pdf_new_filter_processor_with_text_filter(ctx, doc, proc_buffer, resources, res, text_filter, after_text, proc_arg);
pdf_process_contents(ctx, proc_filter, doc, resources, contents, cookie);
pdf_close_processor(ctx, proc_filter);
}
else
{
res = pdf_keep_obj(ctx, resources);
pdf_process_contents(ctx, proc_buffer, doc, resources, contents, cookie);
}
pdf_close_processor(ctx, proc_buffer);
/* Deal with page content stream. */
if (pdf_is_array(ctx, contents))
{
/* create a new object to replace the array */
new_obj = pdf_new_dict(ctx, doc, 1);
new_ref = pdf_add_object(ctx, doc, new_obj);
contents = new_ref;
pdf_dict_put(ctx, page->obj, PDF_NAME(Contents), contents);
}
else
{
pdf_dict_del(ctx, contents, PDF_NAME(Filter));
pdf_dict_del(ctx, contents, PDF_NAME(DecodeParms));
}
pdf_update_stream(ctx, doc, contents, buffer, 0);
/* Now deal with resources. The spec allows for Type3 fonts and form
* XObjects to omit a resource dictionary and look in the parent.
* Avoid that by flattening here as part of the cleaning. This could
* conceivably cause changes in rendering, but we don't care. */
/* ExtGState */
obj = pdf_dict_get(ctx, res, PDF_NAME(ExtGState));
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *o = pdf_dict_get(ctx, pdf_dict_get_val(ctx, obj, i), PDF_NAME(SMask));
if (!o)
continue;
o = pdf_dict_get(ctx, o, PDF_NAME(G));
if (!o)
continue;
/* Transparency group XObject */
pdf_clean_stream_object(ctx, doc, o, resources, cookie, 1, text_filter, after_text, proc_arg, sanitize, ascii);
}
}
/* Pattern */
obj = pdf_dict_get(ctx, res, PDF_NAME(Pattern));
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *pat_res;
pdf_obj *pat = pdf_dict_get_val(ctx, obj, i);
if (!pat)
continue;
pat_res = pdf_dict_get(ctx, pat, PDF_NAME(Resources));
if (pat_res == NULL)
pat_res = resources;
if (pdf_dict_get_int(ctx, pat, PDF_NAME(PatternType)) == 1)
pdf_clean_stream_object(ctx, doc, pat, pat_res, cookie, 0, text_filter, after_text, proc_arg, sanitize, ascii);
}
}
/* XObject */
obj = pdf_dict_get(ctx, res, PDF_NAME(XObject));
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *xobj_res;
pdf_obj *xobj = pdf_dict_get_val(ctx, obj, i);
if (!xobj)
continue;
xobj_res = pdf_dict_get(ctx, xobj, PDF_NAME(Resources));
if (xobj_res == NULL)
xobj_res = resources;
if (pdf_name_eq(ctx, PDF_NAME(Form), pdf_dict_get(ctx, xobj, PDF_NAME(Subtype))))
pdf_clean_stream_object(ctx, doc, xobj, xobj_res, cookie, 1, text_filter, after_text, proc_arg, sanitize, ascii);
}
}
/* Font */
obj = pdf_dict_get(ctx, res, PDF_NAME(Font));
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *o = pdf_dict_get_val(ctx, obj, i);
if (!o)
continue;
if (pdf_name_eq(ctx, PDF_NAME(Type3), pdf_dict_get(ctx, o, PDF_NAME(Subtype))))
pdf_clean_type3(ctx, doc, o, resources, cookie, sanitize, ascii);
}
}
/* ProcSet - no cleaning possible. Inherit this from the old dict. */
obj = pdf_dict_get(ctx, resources, PDF_NAME(ProcSet));
if (obj)
pdf_dict_put(ctx, res, PDF_NAME(ProcSet), obj);
/* ColorSpace - no cleaning possible. */
/* Properties - no cleaning possible. */
if (proc_fn)
(*proc_fn)(ctx, buffer, res, proc_arg);
/* Update resource dictionary */
if (sanitize)
{
pdf_dict_put(ctx, page->obj, PDF_NAME(Resources), res);
}
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
pdf_drop_obj(ctx, new_obj);
pdf_drop_obj(ctx, new_ref);
pdf_drop_obj(ctx, res);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
static pdf_obj*
pdf_get_page_obj (pdf_file *pf, int page_no,
pdf_obj **ret_bbox, pdf_obj **ret_resources)
{
pdf_obj *page_tree;
pdf_obj *bbox = NULL, *resources = NULL, *rotate = NULL;
int page_idx;
/*
* Get Page Tree.
*/
page_tree = NULL;
{
pdf_obj *trailer, *catalog;
pdf_obj *markinfo, *tmp;
trailer = pdf_file_get_trailer(pf);
if (pdf_lookup_dict(trailer, "Encrypt")) {
WARN("This PDF document is encrypted.");
pdf_release_obj(trailer);
return NULL;
}
catalog = pdf_deref_obj(pdf_lookup_dict(trailer, "Root"));
if (!PDF_OBJ_DICTTYPE(catalog)) {
WARN("Can't read document catalog.");
pdf_release_obj(trailer);
if (catalog)
pdf_release_obj(catalog);
return NULL;
}
pdf_release_obj(trailer);
markinfo = pdf_deref_obj(pdf_lookup_dict(catalog, "MarkInfo"));
if (markinfo) {
tmp = pdf_lookup_dict(markinfo, "Marked");
if (PDF_OBJ_BOOLEANTYPE(tmp) && pdf_boolean_value(tmp))
WARN("PDF file is tagged... Ignoring tags.");
pdf_release_obj(markinfo);
}
page_tree = pdf_deref_obj(pdf_lookup_dict(catalog, "Pages"));
pdf_release_obj(catalog);
}
if (!page_tree) {
WARN("Page tree not found.");
return NULL;
}
/*
* Negative page numbers are counted from the back.
*/
{
int count = pdf_number_value(pdf_lookup_dict(page_tree, "Count"));
page_idx = page_no + (page_no >= 0 ? -1 : count);
if (page_idx < 0 || page_idx >= count) {
WARN("Page %ld does not exist.", page_no);
pdf_release_obj(page_tree);
return NULL;
}
page_no = page_idx+1;
}
/*
* Seek correct page. Get Media/Crop Box.
* Media box and resources can be inherited.
*/
{
pdf_obj *kids_ref, *kids;
pdf_obj *crop_box = NULL;
pdf_obj *tmp;
tmp = pdf_lookup_dict(page_tree, "Resources");
resources = tmp ? pdf_deref_obj(tmp) : pdf_new_dict();
while (1) {
int kids_length, i;
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "MediaBox")))) {
if (bbox)
pdf_release_obj(bbox);
bbox = tmp;
}
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "BleedBox")))) {
if (!rect_equal(tmp, bbox)) {
if (bbox)
pdf_release_obj(bbox);
bbox = tmp;
} else {
pdf_release_obj(tmp);
}
}
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "TrimBox")))) {
if (!rect_equal(tmp, bbox)) {
if (bbox)
pdf_release_obj(bbox);
bbox = tmp;
} else {
pdf_release_obj(tmp);
}
}
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "ArtBox")))) {
if (!rect_equal(tmp, bbox)) {
if (bbox)
pdf_release_obj(bbox);
bbox = tmp;
} else {
pdf_release_obj(tmp);
}
}
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "CropBox")))) {
if (crop_box)
pdf_release_obj(crop_box);
crop_box = tmp;
}
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "Rotate")))) {
if (rotate)
pdf_release_obj(rotate);
rotate = tmp;
}
if ((tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "Resources")))) {
#if 0
pdf_merge_dict(tmp, resources);
#endif
if (resources)
pdf_release_obj(resources);
resources = tmp;
}
kids_ref = pdf_lookup_dict(page_tree, "Kids");
if (!kids_ref)
break;
kids = pdf_deref_obj(kids_ref);
kids_length = pdf_array_length(kids);
for (i = 0; i < kids_length; i++) {
int count;
pdf_release_obj(page_tree);
page_tree = pdf_deref_obj(pdf_get_array(kids, i));
tmp = pdf_deref_obj(pdf_lookup_dict(page_tree, "Count"));
if (tmp) {
/* Pages object */
count = pdf_number_value(tmp);
pdf_release_obj(tmp);
} else {
/* Page object */
count = 1;
}
if (page_idx < count)
break;
page_idx -= count;
}
pdf_release_obj(kids);
if (i == kids_length) {
WARN("Page %ld not found! Broken PDF file?", page_no);
if (bbox)
pdf_release_obj(bbox);
if (crop_box)
pdf_release_obj(crop_box);
if (rotate)
pdf_release_obj(rotate);
pdf_release_obj(resources);
pdf_release_obj(page_tree);
return NULL;
}
}
if (crop_box) {
pdf_release_obj(bbox);
bbox = crop_box;
}
}
if (!bbox) {
WARN("No BoundingBox information available.");
pdf_release_obj(page_tree);
pdf_release_obj(resources);
if (rotate)
pdf_release_obj(rotate);
return NULL;
}
if (rotate) {
if (pdf_number_value(rotate) != 0.0)
WARN("<< /Rotate %d >> found. (Not supported yet)",
(int)pdf_number_value(rotate));
pdf_release_obj(rotate);
rotate = NULL;
}
if (ret_bbox != NULL)
*ret_bbox = bbox;
if (ret_resources != NULL)
*ret_resources = resources;
return page_tree;
}
