static int strip_outlines(fz_context *ctx, pdf_document *doc, pdf_obj *outlines, int page_count, int *page_object_nums, pdf_obj *names_list)
{
int nc;
pdf_obj *first;
pdf_obj *last;
first = pdf_dict_get(ctx, outlines, PDF_NAME_First);
if (first == NULL)
nc = 0;
else
nc = strip_outline(ctx, doc, first, page_count, page_object_nums, names_list, &first, &last);
if (nc == 0)
{
pdf_dict_del(ctx, outlines, PDF_NAME_First);
pdf_dict_del(ctx, outlines, PDF_NAME_Last);
pdf_dict_del(ctx, outlines, PDF_NAME_Count);
}
else
{
int old_count = pdf_to_int(ctx, pdf_dict_get(ctx, outlines, PDF_NAME_Count));
pdf_dict_put(ctx, outlines, PDF_NAME_First, first);
pdf_dict_put(ctx, outlines, PDF_NAME_Last, last);
pdf_dict_put(ctx, outlines, PDF_NAME_Count, pdf_new_int(ctx, doc, old_count > 0 ? nc : -nc));
}
return nc;
}
static void
pdf_load_name_tree_imp(pdf_obj *dict, pdf_document *xref, pdf_obj *node)
{
fz_context *ctx = xref->ctx;
pdf_obj *kids = pdf_dict_gets(node, "Kids");
pdf_obj *names = pdf_dict_gets(node, "Names");
int i;
if (kids && !pdf_dict_mark(node))
{
for (i = 0; i < pdf_array_len(kids); i++)
pdf_load_name_tree_imp(dict, xref, pdf_array_get(kids, i));
pdf_dict_unmark(node);
}
if (names)
{
for (i = 0; i + 1 < pdf_array_len(names); i += 2)
{
pdf_obj *key = pdf_array_get(names, i);
pdf_obj *val = pdf_array_get(names, i + 1);
if (pdf_is_string(key))
{
key = pdf_to_utf8_name(ctx, key);
pdf_dict_put(dict, key, val);
pdf_drop_obj(key);
}
else if (pdf_is_name(key))
{
pdf_dict_put(dict, key, val);
}
}
}
}
void pdf_set_portfolio_entry_info(fz_context *ctx, pdf_document *doc, int entry, int schema_entry, pdf_obj *data)
{
pdf_portfolio *p;
pdf_obj *obj, *lookup;
int ef = 0;
if (!doc)
fz_throw(ctx, FZ_ERROR_GENERIC, "Bad pdf_add_portfolio_entry call");
if (doc->portfolio == NULL)
load_portfolio(ctx, doc);
obj = pdf_portfolio_entry_obj_name(ctx, doc, entry, NULL);
if (!obj)
fz_throw(ctx, FZ_ERROR_GENERIC, "Can't set info on non existent portfolio entry");
for (p = doc->portfolio; p != NULL && schema_entry > 0; p = p->next, schema_entry--);
if (schema_entry)
fz_throw(ctx, FZ_ERROR_GENERIC, "schema_entry out of range");
switch (p->entry.type)
{
default:
case PDF_SCHEMA_TEXT:
case PDF_SCHEMA_DATE:
case PDF_SCHEMA_NUMBER:
lookup = NULL;
break;
case PDF_SCHEMA_FILENAME:
lookup = PDF_NAME_UF;
break;
case PDF_SCHEMA_DESC:
lookup = PDF_NAME_Desc;
break;
case PDF_SCHEMA_MODDATE:
lookup = PDF_NAME_ModDate;
ef = 1;
break;
case PDF_SCHEMA_CREATIONDATE:
lookup = PDF_NAME_CreationDate;
ef = 1;
break;
case PDF_SCHEMA_SIZE:
fz_throw(ctx, FZ_ERROR_GENERIC, "Can't set size!");
break;
}
if (lookup)
{
if (ef)
obj = pdf_dict_getl(ctx, obj, PDF_NAME_EF, PDF_NAME_F, PDF_NAME_Params, NULL);
pdf_dict_put(ctx, obj, lookup, data);
if (lookup == PDF_NAME_UF)
pdf_dict_put(ctx, obj, PDF_NAME_F, data);
return;
}
pdf_dict_putl(ctx, obj, data, PDF_NAME_CI, p->key, NULL);
}
/*
** If key doesn't exist, puts key,value pair.
** If key does exist, converts by merging (if dict) or adding (if array)
*/
static void wmupdf_dict_merge_keyval(fz_context *ctx,pdf_obj *dstdict,pdf_obj *key,pdf_obj *value)
{
pdf_obj *dstval;
dstval=pdf_dict_get(dstdict,key);
if (!dstval)
{
pdf_dict_put(dstdict,key,value);
return;
}
/* Values are same--no action required */
if (!pdf_objcmp(dstval,value))
return;
if (pdf_is_dict(dstval) && pdf_is_dict(value))
{
static char *okay_to_merge[] = {"Resources","XObject",""};
int i;
for (i=0;okay_to_merge[i][0]!='\0';i++)
if (!stricmp(okay_to_merge[i],pdf_to_name(key)))
break;
if (okay_to_merge[i][0]!='\0')
{
/* Merge source dict into dest dict */
wmupdf_dict_merge(ctx,pdf_to_name(key),dstval,value);
pdf_dict_put(dstdict,key,dstval);
}
else
/* Just overwrite dest dict with source dict */
pdf_dict_put(dstdict,key,value);
return;
}
/* This works for ProcSet array, but maybe not for any array (e.g. rectangle) */
if (pdf_is_array(dstval) && pdf_is_array(value))
{
wmupdf_array_merge(ctx,pdf_to_name(key),dstval,value);
return;
}
/* Last resort: overwrite with new value */
pdf_dict_put(dstdict,key,value);
/* This does NOT work--you can't just convert the value into an array of values */
/* PDF will become non-conformant. */
/*
array=pdf_new_array(ctx,2);
pdf_array_push(array,dstval);
pdf_array_push(array,value);
pdf_dict_put(dstdict,key,array);
pdf_drop_obj(array);
*/
}
static void update_field_value(fz_context *ctx, pdf_document *doc, pdf_obj *obj, const char *text)
{
pdf_obj *sobj = NULL;
pdf_obj *grp;
if (!text)
text = "";
/* All fields of the same name should be updated, so
* set the value at the head of the group */
grp = find_head_of_field_group(ctx, obj);
if (grp)
obj = grp;
fz_var(sobj);
fz_try(ctx)
{
sobj = pdf_new_string(ctx, doc, text, strlen(text));
pdf_dict_put(ctx, obj, PDF_NAME_V, sobj);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, sobj);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
pdf_field_mark_dirty(ctx, doc, obj);
}
static void
intersect_box(fz_context *ctx, pdf_document *doc, pdf_obj *page, pdf_obj *box_name, const fz_rect *mb)
{
pdf_obj *box = pdf_dict_get(ctx, page, box_name);
pdf_obj *newbox;
fz_rect old_rect;
if (box == NULL)
return;
old_rect.x0 = pdf_to_real(ctx, pdf_array_get(ctx, box, 0));
old_rect.y0 = pdf_to_real(ctx, pdf_array_get(ctx, box, 1));
old_rect.x1 = pdf_to_real(ctx, pdf_array_get(ctx, box, 2));
old_rect.y1 = pdf_to_real(ctx, pdf_array_get(ctx, box, 3));
if (old_rect.x0 < mb->x0)
old_rect.x0 = mb->x0;
if (old_rect.y0 < mb->y0)
old_rect.y0 = mb->y0;
if (old_rect.x1 > mb->x1)
old_rect.x1 = mb->x1;
if (old_rect.y1 > mb->y1)
old_rect.y1 = mb->y1;
newbox = pdf_new_array(ctx, doc, 4);
pdf_array_push(ctx, newbox, pdf_new_real(ctx, doc, old_rect.x0));
pdf_array_push(ctx, newbox, pdf_new_real(ctx, doc, old_rect.y0));
pdf_array_push(ctx, newbox, pdf_new_real(ctx, doc, old_rect.x1));
pdf_array_push(ctx, newbox, pdf_new_real(ctx, doc, old_rect.y1));
pdf_dict_put(ctx, page, box_name, newbox);
}
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;
}
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 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);
}
void
pdf_update_xobject(fz_context *ctx, pdf_document *doc, pdf_obj *form, fz_rect bbox, fz_matrix matrix, pdf_obj *res, fz_buffer *contents)
{
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);
else
pdf_dict_del(ctx, form, PDF_NAME(Resources));
pdf_update_stream(ctx, doc, form, contents, 0);
}
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);
}
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;
}
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);
}
}
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++)
pdf_dict_put(dict, pdf_dict_get_key(obj, i), pdf_dict_get_val(obj, i));
return dict;
}
void pdf_rename_portfolio_schema(fz_context *ctx, pdf_document *doc, int entry, const char *name, int name_len)
{
pdf_portfolio *p;
pdf_obj *s;
if (!doc)
fz_throw(ctx, FZ_ERROR_GENERIC, "Bad pdf_rename_portfolio_schema call");
if (doc->portfolio == NULL)
load_portfolio(ctx, doc);
p = doc->portfolio;
while (p && entry > 0)
p = p->next, entry--;
if (p == NULL || entry)
fz_throw(ctx, FZ_ERROR_GENERIC, "entry out of range in pdf_rename_portfolio_schema");
s = pdf_new_string(ctx, doc, name, name_len);
pdf_drop_obj(ctx, p->entry.name);
p->entry.name = s;
pdf_dict_put(ctx, p->val, PDF_NAME_N, s);
}
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 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);
}
}
/*
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 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--;
}
}
}
void pdf_clean_page_contents(fz_context *ctx, pdf_document *doc, pdf_page *page, fz_cookie *cookie, pdf_page_contents_process_fn *proc_fn, void *proc_arg)
{
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 *ref = NULL;
pdf_obj *obj;
pdf_obj *contents;
fz_buffer *buffer;
fz_var(new_obj);
fz_var(new_ref);
fz_var(res);
fz_var(ref);
fz_var(proc_buffer);
fz_var(proc_filter);
buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
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, page->resources, res);
pdf_process_contents(ctx, proc_filter, doc, page->resources, page->contents, cookie);
contents = page->contents;
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_new_ref(ctx, doc, new_obj);
page->contents = contents = new_ref;
}
else
{
pdf_dict_del(ctx, contents, PDF_NAME_Filter);
pdf_dict_del(ctx, contents, PDF_NAME_DecodeParms);
}
/* 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, page->resources, cookie, 1);
}
}
/* ColorSpace - no cleaning possible */
/* 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 = pdf_dict_get_val(ctx, obj, i);
if (!pat)
continue;
if (pdf_to_int(ctx, pdf_dict_get(ctx, pat, PDF_NAME_PatternType)) == 1)
pdf_clean_stream_object(ctx, doc, pat, page->resources, cookie, 0);
}
}
/* Shading - no cleaning possible */
/* 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 = pdf_dict_get_val(ctx, obj, i);
if (!pdf_name_eq(ctx, PDF_NAME_Form, pdf_dict_get(ctx, xobj, PDF_NAME_Subtype)))
continue;
pdf_clean_stream_object(ctx, doc, xobj, page->resources, cookie, 1);
}
}
/* 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 (pdf_name_eq(ctx, PDF_NAME_Type3, pdf_dict_get(ctx, o, PDF_NAME_Subtype)))
{
pdf_clean_type3(ctx, doc, o, page->resources, cookie);
}
}
}
/* ProcSet - no cleaning possible. Inherit this from the old dict. */
obj = pdf_dict_get(ctx, page->resources, PDF_NAME_ProcSet);
if (obj)
pdf_dict_put(ctx, res, PDF_NAME_ProcSet, obj);
/* Properties - no cleaning possible. */
if (proc_fn)
(*proc_fn)(ctx, buffer, res, proc_arg);
pdf_update_stream(ctx, doc, contents, buffer, 0);
pdf_drop_obj(ctx, page->resources);
ref = pdf_new_ref(ctx, doc, res);
page->resources = pdf_keep_obj(ctx, ref);
pdf_dict_put(ctx, page->me, 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, new_obj);
pdf_drop_obj(ctx, new_ref);
pdf_drop_obj(ctx, res);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow_message(ctx, "Failed while cleaning page");
}
}
pdf_obj *
pdf_parse_dict(pdf_document *doc, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *dict;
pdf_obj *key = NULL;
pdf_obj *val = NULL;
pdf_token tok;
int a, b;
fz_context *ctx = file->ctx;
dict = pdf_new_dict(doc, 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, FZ_ERROR_GENERIC, "invalid key in dict");
key = pdf_new_name(doc, 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(doc, 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, FZ_ERROR_GENERIC, "cannot make sense of broken array after all");
}
break;
case PDF_TOK_OPEN_DICT:
val = pdf_parse_dict(doc, file, buf);
break;
case PDF_TOK_NAME:
val = pdf_new_name(doc, buf->scratch);
break;
case PDF_TOK_REAL:
val = pdf_new_real(doc, buf->f);
break;
case PDF_TOK_STRING:
val = pdf_new_string(doc, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
val = pdf_new_bool(doc, 1);
break;
case PDF_TOK_FALSE:
val = pdf_new_bool(doc, 0);
break;
case PDF_TOK_NULL:
val = pdf_new_null(doc);
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(doc, a);
pdf_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(doc, a, b);
break;
}
}
fz_throw(ctx, FZ_ERROR_GENERIC, "invalid indirect reference in dict");
default:
fz_throw(ctx, FZ_ERROR_GENERIC, "unknown token in dict");
}
pdf_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_rethrow_message(ctx, "cannot parse dict");
}
return dict;
}
pdf_obj *
pdf_parse_dict(fz_context *ctx, pdf_document *doc, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *dict;
pdf_obj *key = NULL;
pdf_obj *val = NULL;
pdf_token tok;
fz_off_t a, b;
dict = pdf_new_dict(ctx, doc, 8);
fz_var(key);
fz_var(val);
fz_try(ctx)
{
while (1)
{
tok = pdf_lex(ctx, 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, FZ_ERROR_GENERIC, "invalid key in dict");
key = pdf_new_name(ctx, doc, buf->scratch);
tok = pdf_lex(ctx, file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
val = pdf_parse_array(ctx, doc, file, buf);
break;
case PDF_TOK_OPEN_DICT:
val = pdf_parse_dict(ctx, doc, file, buf);
break;
case PDF_TOK_NAME: val = pdf_new_name(ctx, doc, buf->scratch); break;
case PDF_TOK_REAL: val = pdf_new_real(ctx, doc, buf->f); break;
case PDF_TOK_STRING: val = pdf_new_string(ctx, doc, buf->scratch, buf->len); break;
case PDF_TOK_TRUE: val = pdf_new_bool(ctx, doc, 1); break;
case PDF_TOK_FALSE: val = pdf_new_bool(ctx, doc, 0); break;
case PDF_TOK_NULL: val = pdf_new_null(ctx, doc); break;
case PDF_TOK_INT:
/* 64-bit to allow for numbers > INT_MAX and overflow */
a = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_CLOSE_DICT || tok == PDF_TOK_NAME ||
(tok == PDF_TOK_KEYWORD && !strcmp(buf->scratch, "ID")))
{
val = pdf_new_int_offset(ctx, doc, a);
pdf_dict_put(ctx, dict, key, val);
pdf_drop_obj(ctx, val);
val = NULL;
pdf_drop_obj(ctx, key);
key = NULL;
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_R)
{
val = pdf_new_indirect(ctx, doc, a, b);
break;
}
}
fz_throw(ctx, FZ_ERROR_GENERIC, "invalid indirect reference in dict");
default:
fz_throw(ctx, FZ_ERROR_GENERIC, "unknown token in dict");
}
pdf_dict_put(ctx, dict, key, val);
pdf_drop_obj(ctx, val);
val = NULL;
pdf_drop_obj(ctx, key);
key = NULL;
}
}
fz_catch(ctx)
{
pdf_drop_obj(ctx, dict);
pdf_drop_obj(ctx, key);
pdf_drop_obj(ctx, val);
fz_rethrow(ctx);
}
return dict;
}
// just copy one of the resource sub-entries (e.g. /Font)
static int copy_and_rename_resource(fz_context *dest_ctx, pdf_obj *dest,
fz_context *src_ctx, pdf_obj *src, char *prefix, struct put_info *info)
{
char new_name[64]; /* this buffer is big enough up to hold all digits for two 16-bit numbers */
int i;
for(i = 0; i < pdf_dict_len(src_ctx, src); i++) {
pdf_obj *src_key = pdf_dict_get_key(src_ctx, src, i);
pdf_obj *src_val = pdf_dict_get_val(src_ctx, src, i);
if(!pdf_is_name(src_ctx, src_key)) {
return(2);
}
/* if this is an inline resource, just copy this object into the new
resource dict */
if(!pdf_is_indirect(src_ctx, src_val)) {
if(snprintf(new_name, sizeof(new_name) / sizeof(new_name[0]),
"%sinline_%d", prefix, info->next_inline_id++) >= sizeof(new_name) / sizeof(new_name[0]))
return(1); // not enough space
pdf_obj *new_res = copy_unassigned_object_continue(dest_ctx,
info->dest_doc, src_ctx, info->src_doc, src_val, &info->new_ids);
//pdf_obj *new_res = pdf_new_dict(dest_ctx, info->dest_doc, 10);
printf("dump it...\n");
pdf_fprint_obj(dest_ctx, stdout, new_res, 0);
/* now reference this new object in the resource object of this sheet */
pdf_obj *dest_key = pdf_new_name(dest_ctx, info->dest_doc, new_name);
pdf_dict_put(dest_ctx, dest, dest_key, new_res);
pdf_drop_obj(dest_ctx, dest_key);
pdf_drop_obj(dest_ctx, new_res);
} else {
/* The new name of resource objects is always the num/gen of the
referenced object in the src-file. Thus we can check by that name
if the object was already referenced by another page of this sheet. */
if(snprintf(new_name, sizeof(new_name) / sizeof(new_name[0]),
"%s%d_%d", prefix, pdf_to_num(dest_ctx, src_val), pdf_to_gen(dest_ctx, src_val)) >=
sizeof(new_name) / sizeof(new_name[0]))
return(1); // not enough space
if(pdf_dict_gets(dest_ctx, dest, new_name) == NULL) {
/* if this resource is not inlined and not already in the resource-dict
of the current sheet... */
/* ...copy the referenced resource to the new document!
If this object has copied already (for another sheet in dest_doc),
copy_object_continue() will do nothing */
pdf_obj *new_res = copy_object_continue(dest_ctx, info->dest_doc,
src_ctx, info->src_doc, src_val, &info->new_ids);
/* now reference this new object in the resource object of this sheet */
pdf_obj *dest_key = pdf_new_name(dest_ctx, info->dest_doc, new_name);
pdf_dict_put(dest_ctx, dest, dest_key, new_res);
pdf_drop_obj(dest_ctx, dest_key);
pdf_drop_obj(dest_ctx, new_res);
}
}
/* even if it was used on another sheet or on this sheet, add it to the
rename-dict for this sheet! Because it could have different names
on different source-pages */
pdf_obj *rename_key = pdf_new_name(dest_ctx, info->dest_doc, pdf_to_name(dest_ctx, src_key));
pdf_obj *rename_val = pdf_new_name(dest_ctx, info->dest_doc, new_name);
pdf_dict_put(dest_ctx, info->rename_dict, rename_key, rename_val);
pdf_drop_obj(dest_ctx, rename_key);
pdf_drop_obj(dest_ctx, rename_val);
}
return(0);
}
/*
When resetting or submitting a form, the fields to act upon are defined
by an array of either field references or field names, plus a flag determining
whether to act upon the fields in the array, or all fields other than those in
the array. specified_fields interprets this information and produces the array
of fields to be acted upon.
*/
static pdf_obj *specified_fields(fz_context *ctx, pdf_document *doc, pdf_obj *fields, int exclude)
{
pdf_obj *form = pdf_dict_getl(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root, PDF_NAME_AcroForm, PDF_NAME_Fields, NULL);
int i, n;
pdf_obj *result = pdf_new_array(ctx, doc, 0);
pdf_obj *nil = NULL;
fz_var(nil);
fz_try(ctx)
{
/* The 'fields' array not being present signals that all fields
* should be acted upon, so handle it using the exclude case - excluding none */
if (exclude || !fields)
{
/* mark the fields we don't want to act upon */
nil = pdf_new_null(ctx, doc);
n = pdf_array_len(ctx, fields);
for (i = 0; i < n; i++)
{
pdf_obj *field = pdf_array_get(ctx, fields, i);
if (pdf_is_string(ctx, field))
field = pdf_lookup_field(ctx, form, pdf_to_str_buf(ctx, field));
if (field)
pdf_dict_put(ctx, field, PDF_NAME_Exclude, nil);
}
/* Act upon all unmarked fields */
n = pdf_array_len(ctx, form);
for (i = 0; i < n; i++)
add_field_hierarchy_to_array(ctx, result, pdf_array_get(ctx, form, i));
/* Unmark the marked fields */
n = pdf_array_len(ctx, fields);
for (i = 0; i < n; i++)
{
pdf_obj *field = pdf_array_get(ctx, fields, i);
if (pdf_is_string(ctx, field))
field = pdf_lookup_field(ctx, form, pdf_to_str_buf(ctx, field));
if (field)
pdf_dict_del(ctx, field, PDF_NAME_Exclude);
}
}
else
{
n = pdf_array_len(ctx, fields);
for (i = 0; i < n; i++)
{
pdf_obj *field = pdf_array_get(ctx, fields, i);
if (pdf_is_string(ctx, field))
field = pdf_lookup_field(ctx, form, pdf_to_str_buf(ctx, field));
if (field)
add_field_hierarchy_to_array(ctx, result, field);
}
}
}
fz_always(ctx)
{
pdf_drop_obj(ctx, nil);
}
fz_catch(ctx)
{
pdf_drop_obj(ctx, result);
fz_rethrow(ctx);
}
return result;
}
static void reset_field(fz_context *ctx, pdf_document *doc, pdf_obj *field)
{
/* Set V to DV whereever DV is present, and delete V where DV is not.
* FIXME: we assume for now that V has not been set unequal
* to DV higher in the hierarchy than "field".
*
* At the bottom of the hierarchy we may find widget annotations
* that aren't also fields, but DV and V will not be present in their
* dictionaries, and attempts to remove V will be harmless. */
pdf_obj *dv = pdf_dict_get(ctx, field, PDF_NAME_DV);
pdf_obj *kids = pdf_dict_get(ctx, field, PDF_NAME_Kids);
if (dv)
pdf_dict_put(ctx, field, PDF_NAME_V, dv);
else
pdf_dict_del(ctx, field, PDF_NAME_V);
if (kids == NULL)
{
/* The leaves of the tree are widget annotations
* In some cases we need to update the appearance state;
* in others we need to mark the field as dirty so that
* the appearance stream will be regenerated. */
switch (pdf_field_type(ctx, doc, field))
{
case PDF_WIDGET_TYPE_RADIOBUTTON:
case PDF_WIDGET_TYPE_CHECKBOX:
{
pdf_obj *leafv = pdf_get_inheritable(ctx, doc, field, PDF_NAME_V);
if (leafv)
pdf_keep_obj(ctx, leafv);
else
leafv = PDF_NAME_Off;
fz_try(ctx)
{
pdf_dict_put(ctx, field, PDF_NAME_AS, leafv);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, leafv);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
break;
case PDF_WIDGET_TYPE_PUSHBUTTON:
break;
default:
pdf_field_mark_dirty(ctx, doc, field);
break;
}
}
if (pdf_field_dirties_document(ctx, doc, field))
doc->dirty = 1;
}
void
pdf_set_free_text_details(fz_context *ctx, pdf_annot *annot, fz_point *pos, char *text, char *font_name, float font_size, float color[3])
{
pdf_document *doc = annot->page->doc;
char nbuf[32];
pdf_obj *dr;
pdf_obj *form_fonts;
pdf_obj *font = NULL;
pdf_obj *ref;
pdf_font_desc *font_desc = NULL;
fz_matrix page_ctm, inv_page_ctm;
pdf_da_info da_info;
fz_buffer *fzbuf = NULL;
fz_point page_pos;
pdf_page_transform(ctx, annot->page, NULL, &page_ctm);
fz_invert_matrix(&inv_page_ctm, &page_ctm);
dr = pdf_dict_get(ctx, annot->page->obj, PDF_NAME(Resources));
if (!dr)
{
dr = pdf_new_dict(ctx, doc, 1);
pdf_dict_put_drop(ctx, annot->page->obj, PDF_NAME(Resources), dr);
}
/* Ensure the resource dictionary includes a font dict */
form_fonts = pdf_dict_get(ctx, dr, PDF_NAME(Font));
if (!form_fonts)
{
form_fonts = pdf_new_dict(ctx, doc, 1);
pdf_dict_put_drop(ctx, dr, PDF_NAME(Font), form_fonts);
/* form_fonts is still valid if execution continues past the above call */
}
fz_var(fzbuf);
fz_var(font);
fz_try(ctx)
{
unsigned char *da_str;
size_t da_len;
fz_rect bounds;
find_free_font_name(ctx, form_fonts, nbuf, sizeof(nbuf));
font = pdf_new_dict(ctx, doc, 5);
ref = pdf_add_object(ctx, doc, font);
pdf_dict_puts_drop(ctx, form_fonts, nbuf, ref);
pdf_dict_put(ctx, font, PDF_NAME(Type), PDF_NAME(Font));
pdf_dict_put(ctx, font, PDF_NAME(Subtype), PDF_NAME(Type1));
pdf_dict_put_name(ctx, font, PDF_NAME(BaseFont), font_name);
pdf_dict_put(ctx, font, PDF_NAME(Encoding), PDF_NAME(WinAnsiEncoding));
memcpy(da_info.col, color, sizeof(float)*3);
da_info.col_size = 3;
da_info.font_name = nbuf;
da_info.font_size = font_size;
fzbuf = fz_new_buffer(ctx, 0);
pdf_fzbuf_print_da(ctx, fzbuf, &da_info);
da_len = fz_buffer_storage(ctx, fzbuf, &da_str);
pdf_dict_put_string(ctx, annot->obj, PDF_NAME(DA), (char *)da_str, da_len);
/* FIXME: should convert to WinAnsiEncoding */
pdf_dict_put_text_string(ctx, annot->obj, PDF_NAME(Contents), text);
font_desc = pdf_load_font(ctx, doc, NULL, font, 0);
pdf_measure_text(ctx, font_desc, (unsigned char *)text, strlen(text), &bounds);
page_pos = *pos;
fz_transform_point(&page_pos, &inv_page_ctm);
bounds.x0 *= font_size;
bounds.x1 *= font_size;
bounds.y0 *= font_size;
bounds.y1 *= font_size;
bounds.x0 += page_pos.x;
bounds.x1 += page_pos.x;
bounds.y0 += page_pos.y;
bounds.y1 += page_pos.y;
pdf_dict_put_rect(ctx, annot->obj, PDF_NAME(Rect), &bounds);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, font);
fz_drop_buffer(ctx, fzbuf);
pdf_drop_font(ctx, font_desc);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
void pdf_add_portfolio_schema(fz_context *ctx, pdf_document *doc, int entry, const pdf_portfolio_schema *info)
{
pdf_portfolio **pp;
pdf_portfolio *p;
pdf_obj *s;
pdf_obj *sc;
int num;
char str_name[32];
pdf_obj *num_name = NULL;
if (!doc)
fz_throw(ctx, FZ_ERROR_GENERIC, "Bad pdf_add_portfolio_schema call");
if (doc->portfolio == NULL)
load_portfolio(ctx, doc);
fz_var(num_name);
pp = &doc->portfolio;
while (*pp && entry > 0)
pp = &(*pp)->next, entry--;
fz_try(ctx)
{
/* Find a name for the new schema entry */
num = 0;
do
{
pdf_drop_obj(ctx, num_name);
num_name = NULL;
num++;
sprintf(str_name, "%d", num);
num_name = pdf_new_name(ctx, doc, str_name);
p = doc->portfolio;
for (p = doc->portfolio; p; p = p->next)
if (pdf_name_eq(ctx, num_name, p->key))
break;
}
while (p);
sc = pdf_new_dict(ctx, doc, 4);
pdf_dict_put_drop(ctx, sc, PDF_NAME_E, pdf_new_bool(ctx, doc, !!info->editable));
pdf_dict_put_drop(ctx, sc, PDF_NAME_V, pdf_new_bool(ctx, doc, !!info->visible));
pdf_dict_put_drop(ctx, sc, PDF_NAME_N, info->name);
pdf_dict_put(ctx, sc, PDF_NAME_Subtype, PDF_NAME_S);
/* Add to our linked list (in the correct sorted place) */
p = fz_malloc_struct(ctx, pdf_portfolio);
p->entry = *info;
p->sort = 0; /* Will be rewritten in a mo */
p->key = pdf_keep_obj(ctx, num_name);
p->val = pdf_keep_obj(ctx, sc);
p->next = *pp;
*pp = p;
/* Add the key to the schema */
s = pdf_dict_getl(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root, PDF_NAME_Collection, PDF_NAME_Schema, NULL);
pdf_dict_put(ctx, s, num_name, sc);
/* Renumber the schema entries */
for (num = 0, p = doc->portfolio; p; num++, p = p->next)
{
pdf_dict_put_drop(ctx, p->val, PDF_NAME_O, pdf_new_int(ctx, doc, num));
p->sort = num;
}
}
fz_always(ctx)
pdf_drop_obj(ctx, num_name);
fz_catch(ctx)
fz_rethrow(ctx);
}
void
pdf_delete_annot(fz_context *ctx, pdf_document *doc, pdf_page *page, pdf_annot *annot)
{
pdf_annot **annotptr;
pdf_obj *old_annot_arr;
pdf_obj *annot_arr;
if (annot == NULL)
return;
/* Remove annot from page's list */
for (annotptr = &page->annots; *annotptr; annotptr = &(*annotptr)->next)
{
if (*annotptr == annot)
break;
}
/* Check the passed annotation was of this page */
if (*annotptr == NULL)
return;
*annotptr = annot->next;
/* If the removed annotation was the last in the list adjust the end pointer */
if (*annotptr == NULL)
page->annot_tailp = annotptr;
/* Stick it in the deleted list */
annot->next = page->deleted_annots;
page->deleted_annots = annot;
pdf_drop_xobject(ctx, annot->ap);
annot->ap = NULL;
/* Recreate the "Annots" array with this annot removed */
old_annot_arr = pdf_dict_get(ctx, page->me, PDF_NAME_Annots);
if (old_annot_arr)
{
int i, n = pdf_array_len(ctx, old_annot_arr);
annot_arr = pdf_new_array(ctx, doc, n?(n-1):0);
fz_try(ctx)
{
for (i = 0; i < n; i++)
{
pdf_obj *obj = pdf_array_get(ctx, old_annot_arr, i);
if (obj != annot->obj)
pdf_array_push(ctx, annot_arr, obj);
}
if (pdf_is_indirect(ctx, old_annot_arr))
pdf_update_object(ctx, doc, pdf_to_num(ctx, old_annot_arr), annot_arr);
else
pdf_dict_put(ctx, page->me, PDF_NAME_Annots, annot_arr);
if (pdf_is_indirect(ctx, annot->obj))
pdf_delete_object(ctx, doc, pdf_to_num(ctx, annot->obj));
}
fz_always(ctx)
{
pdf_drop_obj(ctx, annot_arr);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
pdf_drop_obj(ctx, annot->obj);
annot->obj = NULL;
doc->dirty = 1;
}
pdf_annot *
pdf_create_annot(fz_context *ctx, pdf_page *page, enum pdf_annot_type type)
{
pdf_annot *annot = NULL;
pdf_document *doc = page->doc;
pdf_obj *annot_obj = pdf_new_dict(ctx, doc, 0);
pdf_obj *ind_obj = NULL;
fz_var(annot);
fz_var(ind_obj);
fz_try(ctx)
{
int ind_obj_num;
const char *type_str;
pdf_obj *annot_arr;
type_str = pdf_string_from_annot_type(ctx, type);
if (type == PDF_ANNOT_UNKNOWN)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot create unknown annotation");
annot_arr = pdf_dict_get(ctx, page->obj, PDF_NAME(Annots));
if (annot_arr == NULL)
{
annot_arr = pdf_new_array(ctx, doc, 0);
pdf_dict_put_drop(ctx, page->obj, PDF_NAME(Annots), annot_arr);
}
pdf_dict_put(ctx, annot_obj, PDF_NAME(Type), PDF_NAME(Annot));
pdf_dict_put_name(ctx, annot_obj, PDF_NAME(Subtype), type_str);
/* Make printable as default */
pdf_dict_put_int(ctx, annot_obj, PDF_NAME(F), PDF_ANNOT_IS_PRINT);
/*
Both annotation object and annotation structure are now created.
Insert the object in the hierarchy and the structure in the
page's array.
*/
ind_obj_num = pdf_create_object(ctx, doc);
pdf_update_object(ctx, doc, ind_obj_num, annot_obj);
ind_obj = pdf_new_indirect(ctx, doc, ind_obj_num, 0);
pdf_array_push(ctx, annot_arr, ind_obj);
annot = pdf_new_annot(ctx, page, ind_obj);
annot->ap = NULL;
/*
Linking must be done after any call that might throw because
pdf_drop_annots below actually frees a list. Put the new annot
at the end of the list, so that it will be drawn last.
*/
*page->annot_tailp = annot;
page->annot_tailp = &annot->next;
doc->dirty = 1;
}
fz_always(ctx)
{
pdf_drop_obj(ctx, annot_obj);
pdf_drop_obj(ctx, ind_obj);
}
fz_catch(ctx)
{
pdf_drop_annots(ctx, annot);
fz_rethrow(ctx);
}
return annot;
}
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);
}
