int
pdf_resolve_link(fz_context *ctx, pdf_document *doc, const char *uri, float *xp, float *yp)
{
if (uri && uri[0] == '#')
{
int page = fz_atoi(uri + 1) - 1;
if (xp || yp)
{
const char *x = strchr(uri, ',');
const char *y = strrchr(uri, ',');
if (x && y)
{
pdf_obj *obj;
fz_matrix ctm;
fz_point p;
p.x = x ? fz_atoi(x + 1) : 0;
p.y = y ? fz_atoi(y + 1) : 0;
obj = pdf_lookup_page_obj(ctx, doc, page);
pdf_page_obj_transform(ctx, obj, NULL, &ctm);
fz_transform_point(&p, &ctm);
if (xp) *xp = p.x;
if (yp) *yp = p.y;
}
}
return page;
}
fz_warn(ctx, "unknown link uri '%s'", uri);
return -1;
}
static int
showpage(fz_context *ctx, pdf_document *doc, fz_output *out, int page)
{
pdf_obj *pageref;
int failed = 0;
fz_printf(ctx, out, "<page pagenum=\"%d\">\n", page);
fz_try(ctx)
{
pageref = pdf_lookup_page_obj(ctx, doc, page-1);
if (!pageref)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot retrieve info from page %d", page);
}
fz_catch(ctx)
{
fz_printf(ctx, out, "Failed to gather information for page %d\n", page);
failed = 1;
}
if (!failed)
{
failed |= showbox(ctx, out, pageref, "MediaBox", PDF_NAME_MediaBox);
failed |= showbox(ctx, out, pageref, "CropBox", PDF_NAME_CropBox);
failed |= showbox(ctx, out, pageref, "ArtBox", PDF_NAME_ArtBox);
failed |= showbox(ctx, out, pageref, "BleedBox", PDF_NAME_BleedBox);
failed |= showbox(ctx, out, pageref, "TrimBox", PDF_NAME_TrimBox);
failed |= shownum(ctx, out, pageref, "Rotate", PDF_NAME_Rotate);
failed |= shownum(ctx, out, pageref, "UserUnit", PDF_NAME_UserUnit);
}
fz_printf(ctx, out, "</page>\n");
return failed;
}
static void retainpage(fz_context *ctx, pdf_document *doc, pdf_obj *parent, pdf_obj *kids, int page)
{
pdf_obj *pageref = pdf_lookup_page_obj(ctx, doc, page-1);
pdf_obj *pageobj = pdf_resolve_indirect(ctx, pageref);
pdf_dict_put(ctx, pageobj, PDF_NAME_Parent, parent);
/* Store page object in new kids array */
pdf_array_push(ctx, kids, pageref);
}
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);
}
}
static void retainpage(fz_context *ctx, pdf_document *doc, pdf_obj *parent, pdf_obj *kids, int page)
{
pdf_obj *pageref = pdf_lookup_page_obj(ctx, doc, page-1);
pdf_flatten_inheritable_page_items(ctx, pageref);
pdf_dict_put(ctx, pageref, PDF_NAME_Parent, parent);
/* Store page object in new kids array */
pdf_array_push(ctx, kids, pageref);
}
/* SumatraPDF: allow replacing potentially slow pdf_lookup_page_obj */
pdf_page *
pdf_load_page(pdf_document *doc, int number)
{
pdf_obj *pageref;
if (doc->file_reading_linearly)
{
pageref = pdf_progressive_advance(doc, number);
if (pageref == NULL)
fz_throw(doc->ctx, FZ_ERROR_TRYLATER, "page %d not available yet", number);
}
else
pageref = pdf_lookup_page_obj(doc, number);
return pdf_load_page_by_obj(doc, number, pageref);
}
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 void
gatherpageinfo(int page, int show)
{
pdf_obj *pageobj;
pdf_obj *pageref;
pdf_obj *rsrc;
pageref = pdf_lookup_page_obj(doc, page-1);
pageobj = pdf_resolve_indirect(pageref);
if (!pageobj)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot retrieve info from page %d", page);
gatherdimensions(page, pageref, pageobj);
rsrc = pdf_dict_gets(pageobj, "Resources");
gatherresourceinfo(page, rsrc, show);
}
static void
gatherresourceinfo(int page, pdf_obj *rsrc, int show)
{
pdf_obj *pageobj;
pdf_obj *pageref;
pdf_obj *font;
pdf_obj *xobj;
pdf_obj *shade;
pdf_obj *pattern;
pdf_obj *subrsrc;
int i;
pageref = pdf_lookup_page_obj(doc, page-1);
pageobj = pdf_resolve_indirect(pageref);
if (!pageobj)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot retrieve info from page %d", page);
font = pdf_dict_gets(rsrc, "Font");
if (show & FONTS && font)
{
int n;
gatherfonts(page, pageref, pageobj, font);
n = pdf_dict_len(font);
for (i = 0; i < n; i++)
{
pdf_obj *obj = pdf_dict_get_val(font, i);
subrsrc = pdf_dict_gets(obj, "Resources");
if (subrsrc && pdf_objcmp(rsrc, subrsrc))
gatherresourceinfo(page, subrsrc, show);
}
}
xobj = pdf_dict_gets(rsrc, "XObject");
if (show & XOBJS && xobj)
{
int n;
gatherimages(page, pageref, pageobj, xobj);
gatherforms(page, pageref, pageobj, xobj);
gatherpsobjs(page, pageref, pageobj, xobj);
n = pdf_dict_len(xobj);
for (i = 0; i < n; i++)
{
pdf_obj *obj = pdf_dict_get_val(xobj, i);
subrsrc = pdf_dict_gets(obj, "Resources");
if (subrsrc && pdf_objcmp(rsrc, subrsrc))
gatherresourceinfo(page, subrsrc, show);
}
}
shade = pdf_dict_gets(rsrc, "Shading");
if (show & SHADINGS && shade)
gathershadings(page, pageref, pageobj, shade);
pattern = pdf_dict_gets(rsrc, "Pattern");
if (show & PATTERNS && pattern)
{
int n;
gatherpatterns(page, pageref, pageobj, pattern);
n = pdf_dict_len(pattern);
for (i = 0; i < n; i++)
{
pdf_obj *obj = pdf_dict_get_val(pattern, i);
subrsrc = pdf_dict_gets(obj, "Resources");
if (subrsrc && pdf_objcmp(rsrc, subrsrc))
gatherresourceinfo(page, subrsrc, show);
}
}
}
static void retainpages(fz_context *ctx, globals *glo, int argc, char **argv)
{
pdf_obj *oldroot, *root, *pages, *kids, *countobj, *parent, *olddests;
pdf_document *doc = glo->doc;
int argidx = 0;
pdf_obj *names_list = NULL;
int pagecount;
int i;
/* Keep only pages/type and (reduced) dest entries to avoid
* references to unretained pages */
oldroot = pdf_dict_get(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root);
pages = pdf_dict_get(ctx, oldroot, PDF_NAME_Pages);
olddests = pdf_load_name_tree(ctx, doc, PDF_NAME_Dests);
root = pdf_new_dict(ctx, doc, 2);
pdf_dict_put(ctx, root, PDF_NAME_Type, pdf_dict_get(ctx, oldroot, PDF_NAME_Type));
pdf_dict_put(ctx, root, PDF_NAME_Pages, pdf_dict_get(ctx, oldroot, PDF_NAME_Pages));
pdf_update_object(ctx, doc, pdf_to_num(ctx, oldroot), root);
pdf_drop_obj(ctx, root);
/* Create a new kids array with only the pages we want to keep */
parent = pdf_new_indirect(ctx, doc, pdf_to_num(ctx, pages), pdf_to_gen(ctx, pages));
kids = pdf_new_array(ctx, doc, 1);
/* Retain pages specified */
while (argc - argidx)
{
int page, spage, epage;
char *spec, *dash;
char *pagelist = argv[argidx];
pagecount = pdf_count_pages(ctx, doc);
spec = fz_strsep(&pagelist, ",");
while (spec)
{
dash = strchr(spec, '-');
if (dash == spec)
spage = epage = pagecount;
else
spage = epage = atoi(spec);
if (dash)
{
if (strlen(dash) > 1)
epage = atoi(dash + 1);
else
epage = pagecount;
}
spage = fz_clampi(spage, 1, pagecount);
epage = fz_clampi(epage, 1, pagecount);
if (spage < epage)
for (page = spage; page <= epage; ++page)
retainpage(ctx, doc, parent, kids, page);
else
for (page = spage; page >= epage; --page)
retainpage(ctx, doc, parent, kids, page);
spec = fz_strsep(&pagelist, ",");
}
argidx++;
}
pdf_drop_obj(ctx, parent);
/* Update page count and kids array */
countobj = pdf_new_int(ctx, doc, pdf_array_len(ctx, kids));
pdf_dict_put(ctx, pages, PDF_NAME_Count, countobj);
pdf_drop_obj(ctx, countobj);
pdf_dict_put(ctx, pages, PDF_NAME_Kids, kids);
pdf_drop_obj(ctx, kids);
/* Also preserve the (partial) Dests name tree */
if (olddests)
{
pdf_obj *names = pdf_new_dict(ctx, doc, 1);
pdf_obj *dests = pdf_new_dict(ctx, doc, 1);
int len = pdf_dict_len(ctx, olddests);
names_list = pdf_new_array(ctx, doc, 32);
for (i = 0; i < len; i++)
{
pdf_obj *key = pdf_dict_get_key(ctx, olddests, i);
pdf_obj *val = pdf_dict_get_val(ctx, olddests, i);
pdf_obj *dest = pdf_dict_get(ctx, val, PDF_NAME_D);
dest = pdf_array_get(ctx, dest ? dest : val, 0);
if (pdf_array_contains(ctx, pdf_dict_get(ctx, pages, PDF_NAME_Kids), dest))
{
pdf_obj *key_str = pdf_new_string(ctx, doc, pdf_to_name(ctx, key), strlen(pdf_to_name(ctx, key)));
pdf_array_push(ctx, names_list, key_str);
pdf_array_push(ctx, names_list, val);
pdf_drop_obj(ctx, key_str);
}
}
root = pdf_dict_get(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root);
pdf_dict_put(ctx, dests, PDF_NAME_Names, names_list);
pdf_dict_put(ctx, names, PDF_NAME_Dests, dests);
pdf_dict_put(ctx, root, PDF_NAME_Names, names);
pdf_drop_obj(ctx, names);
pdf_drop_obj(ctx, dests);
pdf_drop_obj(ctx, names_list);
pdf_drop_obj(ctx, olddests);
}
/* Force the next call to pdf_count_pages to recount */
glo->doc->page_count = 0;
/* Edit each pages /Annot list to remove any links that point to
* nowhere. */
pagecount = pdf_count_pages(ctx, doc);
for (i = 0; i < pagecount; i++)
{
pdf_obj *pageref = pdf_lookup_page_obj(ctx, doc, i);
pdf_obj *pageobj = pdf_resolve_indirect(ctx, pageref);
pdf_obj *annots = pdf_dict_get(ctx, pageobj, PDF_NAME_Annots);
int len = pdf_array_len(ctx, annots);
int j;
for (j = 0; j < len; j++)
{
pdf_obj *o = pdf_array_get(ctx, annots, j);
pdf_obj *p;
if (!pdf_name_eq(ctx, pdf_dict_get(ctx, o, PDF_NAME_Subtype), PDF_NAME_Link))
continue;
p = pdf_dict_get(ctx, o, PDF_NAME_A);
if (!pdf_name_eq(ctx, pdf_dict_get(ctx, p, PDF_NAME_S), PDF_NAME_GoTo))
continue;
if (string_in_names_list(ctx, pdf_dict_get(ctx, p, PDF_NAME_D), names_list))
continue;
/* FIXME: Should probably look at Next too */
/* Remove this annotation */
pdf_array_delete(ctx, annots, j);
j--;
}
}
}
pdf_page *
pdf_load_page(pdf_document *doc, int number)
{
fz_context *ctx = doc->ctx;
pdf_page *page;
pdf_annot *annot;
pdf_obj *pageobj, *pageref, *obj;
fz_rect mediabox, cropbox, realbox;
float userunit;
fz_matrix mat;
pageref = pdf_lookup_page_obj(doc, number);
pageobj = pdf_resolve_indirect(pageref);
page = fz_malloc_struct(ctx, pdf_page);
page->resources = NULL;
page->contents = NULL;
page->transparency = 0;
page->links = NULL;
page->annots = NULL;
page->deleted_annots = NULL;
page->tmp_annots = NULL;
page->me = pdf_keep_obj(pageobj);
obj = pdf_dict_gets(pageobj, "UserUnit");
if (pdf_is_real(obj))
userunit = pdf_to_real(obj);
else
userunit = 1;
pdf_to_rect(ctx, pdf_lookup_inherited_page_item(doc, pageobj, "MediaBox"), &mediabox);
if (fz_is_empty_rect(&mediabox))
{
fz_warn(ctx, "cannot find page size for page %d", number + 1);
mediabox.x0 = 0;
mediabox.y0 = 0;
mediabox.x1 = 612;
mediabox.y1 = 792;
}
pdf_to_rect(ctx, pdf_lookup_inherited_page_item(doc, pageobj, "CropBox"), &cropbox);
if (!fz_is_empty_rect(&cropbox))
fz_intersect_rect(&mediabox, &cropbox);
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;
if (page->mediabox.x1 - page->mediabox.x0 < 1 || page->mediabox.y1 - page->mediabox.y0 < 1)
{
fz_warn(ctx, "invalid page size in page %d", number + 1);
page->mediabox = fz_unit_rect;
}
page->rotate = pdf_to_int(pdf_lookup_inherited_page_item(doc, pageobj, "Rotate"));
/* 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;
fz_pre_rotate(fz_scale(&page->ctm, 1, -1), -page->rotate);
realbox = page->mediabox;
fz_transform_rect(&realbox, &page->ctm);
fz_pre_scale(fz_translate(&mat, -realbox.x0, -realbox.y0), userunit, userunit);
fz_concat(&page->ctm, &page->ctm, &mat);
obj = pdf_dict_gets(pageobj, "Annots");
if (obj)
{
/* SumatraPDF: ignore annotations in case of unexpected errors */
fz_try(ctx)
{
page->links = pdf_load_link_annots(doc, obj, &page->ctm);
page->annots = pdf_load_annots(doc, obj, page);
}
fz_catch(ctx)
{
fz_warn(ctx, "unexpectedly failed to load page annotations");
}
}
page->duration = pdf_to_real(pdf_dict_gets(pageobj, "Dur"));
obj = pdf_dict_gets(pageobj, "Trans");
page->transition_present = (obj != NULL);
if (obj)
{
pdf_load_transition(doc, page, obj);
}
// TODO: inherit
page->resources = pdf_lookup_inherited_page_item(doc, pageobj, "Resources");
if (page->resources)
pdf_keep_obj(page->resources);
obj = pdf_dict_gets(pageobj, "Contents");
fz_try(ctx)
{
page->contents = pdf_keep_obj(obj);
if (pdf_resources_use_blending(doc, page->resources))
page->transparency = 1;
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=2107 */
else if (!strcmp(pdf_to_name(pdf_dict_getp(pageobj, "Group/S")), "Transparency"))
page->transparency = 1;
for (annot = page->annots; annot && !page->transparency; annot = annot->next)
if (annot->ap && pdf_resources_use_blending(doc, annot->ap->resources))
page->transparency = 1;
}
fz_catch(ctx)
{
pdf_free_page(doc, page);
fz_rethrow_message(ctx, "cannot load page %d contents (%d 0 R)", number + 1, pdf_to_num(pageref));
}
return page;
}
static void decimatepages(fz_context *ctx, pdf_document *doc)
{
pdf_obj *oldroot, *root, *pages, *kids, *parent;
int num_pages = pdf_count_pages(ctx, doc);
int page, kidcount;
oldroot = pdf_dict_get(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root);
pages = pdf_dict_get(ctx, oldroot, PDF_NAME_Pages);
root = pdf_new_dict(ctx, doc, 2);
pdf_dict_put(ctx, root, PDF_NAME_Type, pdf_dict_get(ctx, oldroot, PDF_NAME_Type));
pdf_dict_put(ctx, root, PDF_NAME_Pages, pdf_dict_get(ctx, oldroot, PDF_NAME_Pages));
pdf_update_object(ctx, doc, pdf_to_num(ctx, oldroot), root);
pdf_drop_obj(ctx, root);
/* Create a new kids array with our new pages in */
parent = pdf_new_indirect(ctx, doc, pdf_to_num(ctx, pages), pdf_to_gen(ctx, pages));
kids = pdf_new_array(ctx, doc, 1);
kidcount = 0;
for (page=0; page < num_pages; page++)
{
pdf_page *page_details = pdf_load_page(ctx, doc, page);
int xf = x_factor, yf = y_factor;
int x, y;
float w = page_details->mediabox.x1 - page_details->mediabox.x0;
float h = page_details->mediabox.y1 - page_details->mediabox.y0;
if (xf == 0 && yf == 0)
{
/* Nothing specified, so split along the long edge */
if (w > h)
xf = 2, yf = 1;
else
xf = 1, yf = 2;
}
else if (xf == 0)
xf = 1;
else if (yf == 0)
yf = 1;
for (y = yf-1; y >= 0; y--)
{
for (x = 0; x < xf; x++)
{
pdf_obj *newpageobj, *newpageref, *newmediabox;
fz_rect mb;
int num;
newpageobj = pdf_copy_dict(ctx, pdf_lookup_page_obj(ctx, doc, page));
num = pdf_create_object(ctx, doc);
pdf_update_object(ctx, doc, num, newpageobj);
newpageref = pdf_new_indirect(ctx, doc, num, 0);
newmediabox = pdf_new_array(ctx, doc, 4);
mb.x0 = page_details->mediabox.x0 + (w/xf)*x;
if (x == xf-1)
mb.x1 = page_details->mediabox.x1;
else
mb.x1 = page_details->mediabox.x0 + (w/xf)*(x+1);
mb.y0 = page_details->mediabox.y0 + (h/yf)*y;
if (y == yf-1)
mb.y1 = page_details->mediabox.y1;
else
mb.y1 = page_details->mediabox.y0 + (h/yf)*(y+1);
pdf_array_push(ctx, newmediabox, pdf_new_real(ctx, doc, mb.x0));
pdf_array_push(ctx, newmediabox, pdf_new_real(ctx, doc, mb.y0));
pdf_array_push(ctx, newmediabox, pdf_new_real(ctx, doc, mb.x1));
pdf_array_push(ctx, newmediabox, pdf_new_real(ctx, doc, mb.y1));
pdf_dict_put(ctx, newpageobj, PDF_NAME_Parent, parent);
pdf_dict_put(ctx, newpageobj, PDF_NAME_MediaBox, newmediabox);
/* Store page object in new kids array */
pdf_array_push(ctx, kids, newpageref);
kidcount++;
}
}
}
pdf_drop_obj(ctx, parent);
/* Update page count and kids array */
pdf_dict_put(ctx, pages, PDF_NAME_Count, pdf_new_int(ctx, doc, kidcount));
pdf_dict_put(ctx, pages, PDF_NAME_Kids, kids);
pdf_drop_obj(ctx, kids);
}
static void retainpages(fz_context *ctx, globals *glo, int argc, char **argv)
{
pdf_obj *oldroot, *root, *pages, *kids, *countobj, *parent, *olddests;
pdf_document *doc = glo->doc;
int argidx = 0;
pdf_obj *names_list = NULL;
pdf_obj *outlines;
int pagecount;
int i;
int *page_object_nums;
/* Keep only pages/type and (reduced) dest entries to avoid
* references to unretained pages */
oldroot = pdf_dict_get(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root);
pages = pdf_dict_get(ctx, oldroot, PDF_NAME_Pages);
olddests = pdf_load_name_tree(ctx, doc, PDF_NAME_Dests);
outlines = pdf_dict_get(ctx, oldroot, PDF_NAME_Outlines);
root = pdf_new_dict(ctx, doc, 3);
pdf_dict_put(ctx, root, PDF_NAME_Type, pdf_dict_get(ctx, oldroot, PDF_NAME_Type));
pdf_dict_put(ctx, root, PDF_NAME_Pages, pdf_dict_get(ctx, oldroot, PDF_NAME_Pages));
pdf_dict_put(ctx, root, PDF_NAME_Outlines, outlines);
pdf_update_object(ctx, doc, pdf_to_num(ctx, oldroot), root);
/* Create a new kids array with only the pages we want to keep */
parent = pdf_new_indirect(ctx, doc, pdf_to_num(ctx, pages), pdf_to_gen(ctx, pages));
kids = pdf_new_array(ctx, doc, 1);
/* Retain pages specified */
while (argc - argidx)
{
int page, spage, epage;
char *spec, *dash;
char *pagelist = argv[argidx];
pagecount = pdf_count_pages(ctx, doc);
spec = fz_strsep(&pagelist, ",");
while (spec)
{
dash = strchr(spec, '-');
if (dash == spec)
spage = epage = pagecount;
else
spage = epage = atoi(spec);
if (dash)
{
if (strlen(dash) > 1)
epage = atoi(dash + 1);
else
epage = pagecount;
}
spage = fz_clampi(spage, 1, pagecount);
epage = fz_clampi(epage, 1, pagecount);
if (spage < epage)
for (page = spage; page <= epage; ++page)
retainpage(ctx, doc, parent, kids, page);
else
for (page = spage; page >= epage; --page)
retainpage(ctx, doc, parent, kids, page);
spec = fz_strsep(&pagelist, ",");
}
argidx++;
}
pdf_drop_obj(ctx, parent);
/* Update page count and kids array */
countobj = pdf_new_int(ctx, doc, pdf_array_len(ctx, kids));
pdf_dict_put(ctx, pages, PDF_NAME_Count, countobj);
pdf_drop_obj(ctx, countobj);
pdf_dict_put(ctx, pages, PDF_NAME_Kids, kids);
pdf_drop_obj(ctx, kids);
/* Force the next call to pdf_count_pages to recount */
glo->doc->page_count = 0;
pagecount = pdf_count_pages(ctx, doc);
page_object_nums = fz_calloc(ctx, pagecount, sizeof(*page_object_nums));
for (i = 0; i < pagecount; i++)
{
pdf_obj *pageref = pdf_lookup_page_obj(ctx, doc, i);
page_object_nums[i] = pdf_to_num(ctx, pageref);
}
/* If we had an old Dests tree (now reformed as an olddests
* dictionary), keep any entries in there that point to
* valid pages. This may mean we keep more than we need, but
* it's safe at least. */
if (olddests)
{
pdf_obj *names = pdf_new_dict(ctx, doc, 1);
pdf_obj *dests = pdf_new_dict(ctx, doc, 1);
int len = pdf_dict_len(ctx, olddests);
names_list = pdf_new_array(ctx, doc, 32);
for (i = 0; i < len; i++)
{
pdf_obj *key = pdf_dict_get_key(ctx, olddests, i);
pdf_obj *val = pdf_dict_get_val(ctx, olddests, i);
pdf_obj *dest = pdf_dict_get(ctx, val, PDF_NAME_D);
dest = pdf_array_get(ctx, dest ? dest : val, 0);
if (dest_is_valid_page(ctx, dest, page_object_nums, pagecount))
{
pdf_obj *key_str = pdf_new_string(ctx, doc, pdf_to_name(ctx, key), strlen(pdf_to_name(ctx, key)));
pdf_array_push(ctx, names_list, key_str);
pdf_array_push(ctx, names_list, val);
pdf_drop_obj(ctx, key_str);
}
}
pdf_dict_put(ctx, dests, PDF_NAME_Names, names_list);
pdf_dict_put(ctx, names, PDF_NAME_Dests, dests);
pdf_dict_put(ctx, root, PDF_NAME_Names, names);
pdf_drop_obj(ctx, names);
pdf_drop_obj(ctx, dests);
pdf_drop_obj(ctx, olddests);
}
/* Edit each pages /Annot list to remove any links that point to
* nowhere. */
for (i = 0; i < pagecount; i++)
{
pdf_obj *pageref = pdf_lookup_page_obj(ctx, doc, i);
pdf_obj *pageobj = pdf_resolve_indirect(ctx, pageref);
pdf_obj *annots = pdf_dict_get(ctx, pageobj, PDF_NAME_Annots);
int len = pdf_array_len(ctx, annots);
int j;
for (j = 0; j < len; j++)
{
pdf_obj *o = pdf_array_get(ctx, annots, j);
if (!pdf_name_eq(ctx, pdf_dict_get(ctx, o, PDF_NAME_Subtype), PDF_NAME_Link))
continue;
if (!dest_is_valid(ctx, o, pagecount, page_object_nums, names_list))
{
/* Remove this annotation */
pdf_array_delete(ctx, annots, j);
j--;
}
}
}
if (strip_outlines(ctx, doc, outlines, pagecount, page_object_nums, names_list) == 0)
{
pdf_dict_del(ctx, root, PDF_NAME_Outlines);
}
fz_free(ctx, page_object_nums);
pdf_drop_obj(ctx, names_list);
pdf_drop_obj(ctx, root);
}
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);
}
