pdf_obj *
pdf_parse_stm_obj(pdf_document *doc, fz_stream *file, pdf_lexbuf *buf)
{
pdf_token tok;
fz_context *ctx = file->ctx;
tok = pdf_lex(file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
return pdf_parse_array(doc, file, buf);
case PDF_TOK_OPEN_DICT:
return pdf_parse_dict(doc, file, buf);
case PDF_TOK_NAME: return pdf_new_name(doc, buf->scratch); break;
case PDF_TOK_REAL: return pdf_new_real(doc, buf->f); break;
case PDF_TOK_STRING: return pdf_new_string(doc, buf->scratch, buf->len); break;
case PDF_TOK_TRUE: return pdf_new_bool(doc, 1); break;
case PDF_TOK_FALSE: return pdf_new_bool(doc, 0); break;
case PDF_TOK_NULL: return pdf_new_null(doc); break;
case PDF_TOK_INT: return pdf_new_int(doc, buf->i); break;
default: fz_throw(ctx, FZ_ERROR_GENERIC, "unknown token in object stream");
}
return NULL; /* Stupid MSVC */
}
void
pdf_set_annot_contents(fz_context *ctx, pdf_annot *annot, const char *text)
{
pdf_document *doc = annot->page->doc;
pdf_dict_put_drop(ctx, annot->obj, PDF_NAME_Contents, pdf_new_string(ctx, doc, text, strlen(text)));
annot->changed = 1;
}
void pdf_new_page_label (pdf_file_handle pdf_file,
int page_index,
int base,
int count,
char style,
char *prefix)
{
struct pdf_obj *label_dict;
char style_str [2] = { style, '\0' };
if (! pdf_file->page_label_tree)
{
pdf_file->page_label_tree = pdf_new_name_tree (pdf_file, 1);
}
label_dict = pdf_new_obj (PT_DICTIONARY);
if (style)
pdf_set_dict_entry (label_dict, "S", pdf_new_name (style_str));
if (prefix)
pdf_set_dict_entry (label_dict, "P", pdf_new_string (prefix));
if (base > 1)
pdf_set_dict_entry (label_dict, "St", pdf_new_integer (base));
pdf_add_number_tree_element (pdf_file->page_label_tree,
page_index,
label_dict);
}
static void update_field_value(pdf_document *doc, pdf_obj *obj, char *text)
{
fz_context *ctx = doc->ctx;
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(obj);
if (grp)
obj = grp;
fz_var(sobj);
fz_try(ctx)
{
sobj = pdf_new_string(doc, text, strlen(text));
pdf_dict_puts(obj, "V", sobj);
}
fz_always(ctx)
{
pdf_drop_obj(sobj);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
pdf_field_mark_dirty(doc, obj);
}
void pdf_add_name_tree_element (struct pdf_name_tree *tree,
char *key,
struct pdf_obj *val)
{
struct pdf_obj *key_obj = pdf_new_string (key);
pdf_add_tree_element (tree, key_obj, val);
}
static void dict_put_string(fz_context *ctx,pdf_obj *dict,char *key,char *string)
{
pdf_obj *value;
value=pdf_new_string(ctx,string,strlen(string));
pdf_dict_puts(dict,key,value);
pdf_drop_obj(value);
}
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 {
/*
* Set up Indexed ColorSpace for color-type PALETTE:
*
* PNG allows only RGB color for base color space. If gAMA and/or cHRM
* chunk is available, we can use CalRGB color space instead of DeviceRGB
* for base color space.
*
*/
static pdf_obj *
create_cspace_Indexed (png_structp png_ptr, png_infop info_ptr)
{
pdf_obj *colorspace;
pdf_obj *base, *lookup;
png_byte *data_ptr;
png_colorp plte;
int num_plte, i;
if (!png_get_valid(png_ptr, info_ptr, PNG_INFO_PLTE) ||
!png_get_PLTE(png_ptr, info_ptr, &plte, &num_plte)) {
WARN("%s: PNG does not have valid PLTE chunk.", PNG_DEBUG_STR);
return NULL;
}
/* Order is important. */
colorspace = pdf_new_array ();
pdf_add_array(colorspace, pdf_new_name("Indexed"));
if (png_get_valid(png_ptr, info_ptr, PNG_INFO_iCCP))
base = create_cspace_ICCBased(png_ptr, info_ptr);
else {
if (png_get_valid(png_ptr, info_ptr, PNG_INFO_sRGB))
base = create_cspace_sRGB(png_ptr, info_ptr);
else
base = create_cspace_CalRGB(png_ptr, info_ptr);
}
if (!base)
base = pdf_new_name("DeviceRGB");
pdf_add_array(colorspace, base);
pdf_add_array(colorspace, pdf_new_number(num_plte-1));
data_ptr = NEW(num_plte*3, png_byte);
for (i = 0; i < num_plte; i++) {
data_ptr[3*i] = plte[i].red;
data_ptr[3*i+1] = plte[i].green;
data_ptr[3*i+2] = plte[i].blue;
}
lookup = pdf_new_string(data_ptr, num_plte*3);
RELEASE(data_ptr);
pdf_add_array(colorspace, lookup);
return colorspace;
}
pdf_obj *
pdf_parse_stm_obj(pdf_document *xref, fz_stream *file, pdf_lexbuf *buf)
{
int tok;
fz_context *ctx = file->ctx;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse token in object stream") */
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
return pdf_parse_array(xref, file, buf);
/* RJW: "cannot parse object stream" */
case PDF_TOK_OPEN_DICT:
return pdf_parse_dict(xref, file, buf);
/* RJW: "cannot parse object stream" */
case PDF_TOK_NAME:
return fz_new_name(ctx, buf->scratch);
break;
case PDF_TOK_REAL:
return pdf_new_real(ctx, buf->f);
break;
case PDF_TOK_STRING:
return pdf_new_string(ctx, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
return pdf_new_bool(ctx, 1);
break;
case PDF_TOK_FALSE:
return pdf_new_bool(ctx, 0);
break;
case PDF_TOK_NULL:
return pdf_new_null(ctx);
break;
case PDF_TOK_INT:
return pdf_new_int(ctx, buf->i);
break;
default:
fz_throw(ctx, "unknown token in object stream");
}
return NULL; /* Stupid MSVC */
}
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);
}
/*
** 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);
}
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;
}
pdf_obj *
pdf_parse_array(pdf_document *xref, fz_stream *file, pdf_lexbuf *buf)
{
pdf_obj *ary = NULL;
pdf_obj *obj = NULL;
int a = 0, b = 0, n = 0;
int tok;
fz_context *ctx = file->ctx;
pdf_obj *op;
fz_var(obj);
ary = pdf_new_array(ctx, 4);
fz_try(ctx)
{
while (1)
{
tok = pdf_lex(file, buf);
if (tok != PDF_TOK_INT && tok != PDF_TOK_R)
{
if (n > 0)
{
obj = pdf_new_int(ctx, a);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
}
if (n > 1)
{
obj = pdf_new_int(ctx, b);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
}
n = 0;
}
if (tok == PDF_TOK_INT && n == 2)
{
obj = pdf_new_int(ctx, a);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
a = b;
n --;
}
switch (tok)
{
case PDF_TOK_CLOSE_ARRAY:
op = ary;
goto end;
case PDF_TOK_INT:
if (n == 0)
a = buf->i;
if (n == 1)
b = buf->i;
n ++;
break;
case PDF_TOK_R:
if (n != 2)
fz_throw(ctx, "cannot parse indirect reference in array");
obj = pdf_new_indirect(ctx, a, b, xref);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
n = 0;
break;
case PDF_TOK_OPEN_ARRAY:
obj = pdf_parse_array(xref, file, buf);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_OPEN_DICT:
obj = pdf_parse_dict(xref, file, buf);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_NAME:
obj = fz_new_name(ctx, buf->scratch);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_REAL:
obj = pdf_new_real(ctx, buf->f);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_STRING:
obj = pdf_new_string(ctx, buf->scratch, buf->len);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_TRUE:
obj = pdf_new_bool(ctx, 1);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_FALSE:
obj = pdf_new_bool(ctx, 0);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
case PDF_TOK_NULL:
obj = pdf_new_null(ctx);
pdf_array_push(ary, obj);
pdf_drop_obj(obj);
obj = NULL;
break;
default:
fz_throw(ctx, "cannot parse token in array");
}
}
end:
{}
}
fz_catch(ctx)
{
pdf_drop_obj(obj);
pdf_drop_obj(ary);
fz_throw(ctx, "cannot parse array");
}
return op;
}
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);
}
static void retainpages(int argc, char **argv)
{
pdf_obj *oldroot, *root, *pages, *kids, *countobj, *parent, *olddests;
/* Keep only pages/type and (reduced) dest entries to avoid
* references to unretained pages */
oldroot = pdf_dict_gets(xref->trailer, "Root");
pages = pdf_dict_gets(oldroot, "Pages");
olddests = pdf_load_name_tree(xref, "Dests");
root = pdf_new_dict(ctx, 2);
pdf_dict_puts(root, "Type", pdf_dict_gets(oldroot, "Type"));
pdf_dict_puts(root, "Pages", pdf_dict_gets(oldroot, "Pages"));
pdf_update_object(xref, pdf_to_num(oldroot), root);
pdf_drop_obj(root);
/* Create a new kids array with only the pages we want to keep */
parent = pdf_new_indirect(ctx, pdf_to_num(pages), pdf_to_gen(pages), xref);
kids = pdf_new_array(ctx, 1);
/* Retain pages specified */
while (argc - fz_optind)
{
int page, spage, epage, pagecount;
char *spec, *dash;
char *pagelist = argv[fz_optind];
pagecount = pdf_count_pages(xref);
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;
}
if (spage > epage)
page = spage, spage = epage, epage = page;
spage = fz_clampi(spage, 1, pagecount);
epage = fz_clampi(epage, 1, pagecount);
for (page = spage; page <= epage; page++)
{
pdf_obj *pageobj = xref->page_objs[page-1];
pdf_obj *pageref = xref->page_refs[page-1];
pdf_dict_puts(pageobj, "Parent", parent);
/* Store page object in new kids array */
pdf_array_push(kids, pageref);
}
spec = fz_strsep(&pagelist, ",");
}
fz_optind++;
}
pdf_drop_obj(parent);
/* Update page count and kids array */
countobj = pdf_new_int(ctx, pdf_array_len(kids));
pdf_dict_puts(pages, "Count", countobj);
pdf_drop_obj(countobj);
pdf_dict_puts(pages, "Kids", kids);
pdf_drop_obj(kids);
/* Also preserve the (partial) Dests name tree */
if (olddests)
{
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);
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 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_set_annot_contents(pdf_document *doc, pdf_annot *annot, char *text)
{
pdf_dict_puts_drop(annot->obj, "Contents", pdf_new_string(doc, text, strlen(text)));
}
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:
val = pdf_parse_array(doc, file, buf);
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;
}
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_drop(ctx, font, PDF_NAME_Type, PDF_NAME_Font);
pdf_dict_put_drop(ctx, font, PDF_NAME_Subtype, PDF_NAME_Type1);
pdf_dict_put_drop(ctx, font, PDF_NAME_BaseFont, pdf_new_name(ctx, doc, font_name));
pdf_dict_put_drop(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_drop(ctx, annot->obj, PDF_NAME_DA, pdf_new_string(ctx, doc, (char *)da_str, da_len));
/* FIXME: should convert to WinAnsiEncoding */
pdf_dict_put_drop(ctx, annot->obj, PDF_NAME_Contents, pdf_new_string(ctx, doc, text, strlen(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_drop(ctx, annot->obj, PDF_NAME_Rect, pdf_new_rect(ctx, doc, &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_set_free_text_details(pdf_document *doc, pdf_annot *annot, fz_point *pos, char *text, char *font_name, float font_size, float color[3])
{
fz_context *ctx = doc->ctx;
char nbuf[32];
pdf_obj *dr;
pdf_obj *form_fonts;
pdf_obj *font = NULL;
pdf_obj *ref;
pdf_font_desc *font_desc = NULL;
pdf_da_info da_info;
fz_buffer *fzbuf = NULL;
fz_matrix ctm;
fz_point page_pos;
fz_invert_matrix(&ctm, &annot->page->ctm);
dr = pdf_dict_gets(annot->page->me, "Resources");
if (!dr)
{
dr = pdf_new_dict(doc, 1);
pdf_dict_putp_drop(annot->page->me, "Resources", dr);
}
/* Ensure the resource dictionary includes a font dict */
form_fonts = pdf_dict_gets(dr, "Font");
if (!form_fonts)
{
form_fonts = pdf_new_dict(doc, 1);
pdf_dict_puts_drop(dr, "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;
int da_len;
fz_rect bounds;
find_free_font_name(form_fonts, nbuf, sizeof(nbuf));
font = pdf_new_dict(doc, 5);
ref = pdf_new_ref(doc, font);
pdf_dict_puts_drop(form_fonts, nbuf, ref);
pdf_dict_puts_drop(font, "Type", pdf_new_name(doc, "Font"));
pdf_dict_puts_drop(font, "Subtype", pdf_new_name(doc, "Type1"));
pdf_dict_puts_drop(font, "BaseFont", pdf_new_name(doc, font_name));
pdf_dict_puts_drop(font, "Encoding", pdf_new_name(doc, "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_puts_drop(annot->obj, "DA", pdf_new_string(doc, (char *)da_str, da_len));
/* FIXME: should convert to WinAnsiEncoding */
pdf_dict_puts_drop(annot->obj, "Contents", pdf_new_string(doc, text, strlen(text)));
font_desc = pdf_load_font(doc, NULL, font, 0);
pdf_measure_text(ctx, font_desc, (unsigned char *)text, strlen(text), &bounds);
page_pos = *pos;
fz_transform_point(&page_pos, &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_puts_drop(annot->obj, "Rect", pdf_new_rect(doc, &bounds));
update_rect(ctx, annot);
}
fz_always(ctx)
{
pdf_drop_obj(font);
fz_drop_buffer(ctx, fzbuf);
pdf_drop_font(ctx, font_desc);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
pdf_obj *
pdf_parse_ind_obj(fz_context *ctx, pdf_document *doc,
fz_stream *file, pdf_lexbuf *buf,
int *onum, int *ogen, fz_off_t *ostmofs, int *try_repair)
{
pdf_obj *obj = NULL;
int num = 0, gen = 0;
fz_off_t stm_ofs;
pdf_token tok;
fz_off_t a, b;
fz_var(obj);
tok = pdf_lex(ctx, file, buf);
if (tok != PDF_TOK_INT)
{
if (try_repair)
*try_repair = 1;
fz_throw(ctx, FZ_ERROR_GENERIC, "expected object number");
}
num = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok != PDF_TOK_INT)
{
if (try_repair)
*try_repair = 1;
fz_throw(ctx, FZ_ERROR_GENERIC, "expected generation number (%d ? obj)", num);
}
gen = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok != PDF_TOK_OBJ)
{
if (try_repair)
*try_repair = 1;
fz_throw(ctx, FZ_ERROR_GENERIC, "expected 'obj' keyword (%d %d ?)", num, gen);
}
tok = pdf_lex(ctx, file, buf);
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
obj = pdf_parse_array(ctx, doc, file, buf);
break;
case PDF_TOK_OPEN_DICT:
obj = pdf_parse_dict(ctx, doc, file, buf);
break;
case PDF_TOK_NAME: obj = pdf_new_name(ctx, doc, buf->scratch); break;
case PDF_TOK_REAL: obj = pdf_new_real(ctx, doc, buf->f); break;
case PDF_TOK_STRING: obj = pdf_new_string(ctx, doc, buf->scratch, buf->len); break;
case PDF_TOK_TRUE: obj = pdf_new_bool(ctx, doc, 1); break;
case PDF_TOK_FALSE: obj = pdf_new_bool(ctx, doc, 0); break;
case PDF_TOK_NULL: obj = pdf_new_null(ctx, doc); break;
case PDF_TOK_INT:
a = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_STREAM || tok == PDF_TOK_ENDOBJ)
{
obj = pdf_new_int_offset(ctx, doc, a);
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_R)
{
obj = pdf_new_indirect(ctx, doc, a, b);
break;
}
}
fz_throw(ctx, FZ_ERROR_GENERIC, "expected 'R' keyword (%d %d R)", num, gen);
case PDF_TOK_ENDOBJ:
obj = pdf_new_null(ctx, doc);
goto skip;
default:
fz_throw(ctx, FZ_ERROR_GENERIC, "syntax error in object (%d %d R)", num, gen);
}
fz_try(ctx)
{
tok = pdf_lex(ctx, file, buf);
}
fz_catch(ctx)
{
pdf_drop_obj(ctx, obj);
fz_rethrow(ctx);
}
skip:
if (tok == PDF_TOK_STREAM)
{
int c = fz_read_byte(ctx, file);
while (c == ' ')
c = fz_read_byte(ctx, file);
if (c == '\r')
{
c = fz_peek_byte(ctx, file);
if (c != '\n')
fz_warn(ctx, "line feed missing after stream begin marker (%d %d R)", num, gen);
else
fz_read_byte(ctx, file);
}
stm_ofs = fz_tell(ctx, file);
}
else if (tok == PDF_TOK_ENDOBJ)
{
stm_ofs = 0;
}
else
{
fz_warn(ctx, "expected 'endobj' or 'stream' keyword (%d %d R)", num, gen);
stm_ofs = 0;
}
if (onum) *onum = num;
if (ogen) *ogen = gen;
if (ostmofs) *ostmofs = stm_ofs;
return obj;
}
int
bmp_include_image (pdf_ximage *ximage, FILE *fp)
{
pdf_obj *stream, *stream_dict, *colorspace;
ximage_info info;
struct hdr_info hdr;
int num_palette, flip;
int i;
pdf_ximage_init_image_info(&info);
stream = stream_dict = colorspace = NULL;
rewind(fp);
if (read_header(fp, &hdr) < 0)
return -1;
get_density(&info.xdensity, &info.ydensity, &hdr);
info.width = hdr.width;
info.height = hdr.height;
if (info.height < 0) {
info.height = -info.height;
flip = 0;
} else {
flip = 1;
}
if (hdr.bit_count < 24) {
if (hdr.bit_count != 1 &&
hdr.bit_count != 4 && hdr.bit_count != 8) {
WARN("Unsupported palette size: %ld", hdr.bit_count);
return -1;
}
num_palette = (hdr.offset - hdr.hsize - DIB_FILE_HEADER_SIZE) / hdr.psize;
info.bits_per_component = hdr.bit_count;
info.num_components = 1;
} else if (hdr.bit_count == 24) { /* full color */
num_palette = 1; /* dummy */
info.bits_per_component = 8;
info.num_components = 3;
} else {
WARN("Unkown/Unsupported BMP bitCount value: %ld", hdr.bit_count);
return -1;
}
if (info.width == 0 || info.height == 0 || num_palette < 1) {
WARN("Invalid BMP file: width=%ld, height=%ld, #palette=%d",
info.width, info.height, num_palette);
return -1;
}
/* Start reading raster data */
stream = pdf_new_stream(STREAM_COMPRESS);
stream_dict = pdf_stream_dict(stream);
/* Color space: Indexed or DeviceRGB */
if (hdr.bit_count < 24) {
pdf_obj *lookup;
unsigned char *palette, bgrq[4];
palette = NEW(num_palette*3+1, unsigned char);
for (i = 0; i < num_palette; i++) {
if (fread(bgrq, 1, hdr.psize, fp) != hdr.psize) {
WARN("Reading file failed...");
RELEASE(palette);
return -1;
}
/* BGR data */
palette[3*i ] = bgrq[2];
palette[3*i+1] = bgrq[1];
palette[3*i+2] = bgrq[0];
}
lookup = pdf_new_string(palette, num_palette*3);
RELEASE(palette);
colorspace = pdf_new_array();
pdf_add_array(colorspace, pdf_new_name("Indexed"));
pdf_add_array(colorspace, pdf_new_name("DeviceRGB"));
pdf_add_array(colorspace, pdf_new_number(num_palette-1));
pdf_add_array(colorspace, lookup);
} else {
void pdf_set_annot_contents(fz_context *ctx, pdf_document *doc, pdf_annot *annot, char *text)
{
pdf_dict_put_drop(ctx, annot->obj, PDF_NAME_Contents, pdf_new_string(ctx, doc, text, strlen(text)));
}
pdf_obj *
CMap_create_stream (CMap *cmap)
{
pdf_obj *stream;
pdf_obj *stream_dict;
CIDSysInfo *csi;
struct sbuf wbuf;
struct rangeDef *ranges;
unsigned char *codestr;
int i, j, count = 0;
if (!cmap || !CMap_is_valid(cmap)) {
WARN("Invalid CMap");
return NULL;
}
if (cmap->type == CMAP_TYPE_IDENTITY)
return NULL;
stream = pdf_new_stream(STREAM_COMPRESS);
stream_dict = pdf_stream_dict(stream);
csi = CMap_get_CIDSysInfo(cmap);
if (!csi) {
csi = (cmap->type != CMAP_TYPE_TO_UNICODE) ?
&CSI_IDENTITY : &CSI_UNICODE;
}
if (cmap->type != CMAP_TYPE_TO_UNICODE) {
pdf_obj *csi_dict;
csi_dict = pdf_new_dict();
pdf_add_dict(csi_dict,
pdf_new_name("Registry"),
pdf_new_string(csi->registry, strlen(csi->registry)));
pdf_add_dict(csi_dict,
pdf_new_name("Ordering"),
pdf_new_string(csi->ordering, strlen(csi->ordering)));
pdf_add_dict(csi_dict,
pdf_new_name("Supplement"),
pdf_new_number(csi->supplement));
pdf_add_dict(stream_dict,
pdf_new_name("Type"),
pdf_new_name("CMap"));
pdf_add_dict(stream_dict,
pdf_new_name("CMapName"),
pdf_new_name(cmap->name));
pdf_add_dict(stream_dict,
pdf_new_name("CIDSystemInfo"), csi_dict);
if (cmap->wmode != 0)
pdf_add_dict(stream_dict,
pdf_new_name("WMode"),
pdf_new_number(cmap->wmode));
}
/* TODO:
* Predefined CMaps need not to be embedded.
*/
if (cmap->useCMap) {
ERROR("UseCMap found (not supported yet)...");
if (CMap_is_Identity(cmap->useCMap)) { /* not sure */
if (CMap_get_wmode(cmap) == 1) {
pdf_add_dict(stream_dict,
pdf_new_name("UseCMap"),
pdf_new_name("Identity-V"));
} else {
pdf_add_dict(stream_dict,
pdf_new_name("UseCMap"),
pdf_new_name("Identity-H"));
}
} else {
int res_id;
pdf_obj *ucmap_ref;
res_id = pdf_findresource("CMap", CMap_get_name(cmap->useCMap));
if (res_id >= 0) {
ucmap_ref = pdf_get_resource_reference(res_id);
} else {
pdf_obj *ucmap_obj;
ucmap_obj = CMap_create_stream(cmap->useCMap);
if (!ucmap_obj) {
ERROR("Uh ah. I cannot continue...");
}
res_id = pdf_defineresource("CMap",
CMap_get_name(cmap->useCMap),
ucmap_obj, PDF_RES_FLUSH_IMMEDIATE);
ucmap_ref = pdf_get_resource_reference(res_id);
}
pdf_add_dict(stream_dict, pdf_new_name("UseCMap"), ucmap_ref);
}
}
#define WBUF_SIZE 40960
wbuf.buf = NEW(WBUF_SIZE, char);
codestr = NEW(cmap->profile.maxBytesIn, unsigned char);
memset(codestr, 0, cmap->profile.maxBytesIn);
wbuf.curptr = wbuf.buf;
wbuf.limptr = wbuf.buf + WBUF_SIZE -
2 * (cmap->profile.maxBytesIn + cmap->profile.maxBytesOut) + 16;
/* Start CMap */
pdf_add_stream(stream, (const void *) CMAP_BEGIN, strlen(CMAP_BEGIN));
wbuf.curptr += sprintf(wbuf.curptr, "/CMapName ");
write_name(&wbuf.curptr, wbuf.limptr, cmap->name);
wbuf.curptr += sprintf(wbuf.curptr, " def\n");
wbuf.curptr += sprintf(wbuf.curptr, "/CMapType %d def\n" , cmap->type);
if (cmap->wmode != 0 &&
cmap->type != CMAP_TYPE_TO_UNICODE)
wbuf.curptr += sprintf(wbuf.curptr, "/WMode %d def\n", cmap->wmode);
#define CMAP_CSI_FMT "/CIDSystemInfo <<\n\
/Registry (%s)\n\
/Ordering (%s)\n\
/Supplement %d\n\
>> def\n"
wbuf.curptr += sprintf(wbuf.curptr, "/CIDSystemInfo <<\n");
wbuf.curptr += sprintf(wbuf.curptr, " /Registry ");
write_string(&wbuf.curptr, wbuf.limptr, csi->registry);
wbuf.curptr += sprintf(wbuf.curptr, "\n");
wbuf.curptr += sprintf(wbuf.curptr, " /Ordering ");
write_string(&wbuf.curptr, wbuf.limptr, csi->ordering);
wbuf.curptr += sprintf(wbuf.curptr, "\n");
wbuf.curptr += sprintf(wbuf.curptr, " /Supplement %d\n>> def\n",
csi->supplement);
pdf_add_stream(stream, wbuf.buf, (int)(wbuf.curptr - wbuf.buf));
wbuf.curptr = wbuf.buf;
/* codespacerange */
ranges = cmap->codespace.ranges;
wbuf.curptr += sprintf(wbuf.curptr,
"%d begincodespacerange\n", cmap->codespace.num);
for (i = 0; i < cmap->codespace.num; i++) {
*(wbuf.curptr)++ = '<';
for (j = 0; j < ranges[i].dim; j++) {
sputx(ranges[i].codeLo[j], &(wbuf.curptr), wbuf.limptr);
}
*(wbuf.curptr)++ = '>';
*(wbuf.curptr)++ = ' ';
*(wbuf.curptr)++ = '<';
for (j = 0; j < ranges[i].dim; j++) {
sputx(ranges[i].codeHi[j], &(wbuf.curptr), wbuf.limptr);
}
*(wbuf.curptr)++ = '>';
*(wbuf.curptr)++ = '\n';
}
pdf_add_stream(stream, wbuf.buf, (int)(wbuf.curptr - wbuf.buf));
wbuf.curptr = wbuf.buf;
pdf_add_stream(stream,
"endcodespacerange\n", strlen("endcodespacerange\n"));
/* CMap body */
if (cmap->mapTbl) {
count = write_map(cmap->mapTbl,
0, codestr, 0, &wbuf, stream); /* Top node */
if (count > 0) { /* Flush */
char fmt_buf[32];
if (count > 100)
ERROR("Unexpected error....: %d", count);
sprintf(fmt_buf, "%d beginbfchar\n", count);
pdf_add_stream(stream, fmt_buf, strlen(fmt_buf));
pdf_add_stream(stream,
wbuf.buf, (int) (wbuf.curptr - wbuf.buf));
pdf_add_stream(stream,
"endbfchar\n", strlen("endbfchar\n"));
count = 0;
wbuf.curptr = wbuf.buf;
}
}
/* End CMap */
pdf_add_stream(stream, CMAP_END, strlen(CMAP_END));
RELEASE(codestr);
RELEASE(wbuf.buf);
return stream;
}
pdf_obj *
pdf_parse_ind_obj(pdf_document *xref,
fz_stream *file, pdf_lexbuf *buf,
int *onum, int *ogen, int *ostmofs)
{
pdf_obj *obj = NULL;
int num = 0, gen = 0, stm_ofs;
int tok;
int a, b;
fz_context *ctx = file->ctx;
fz_var(obj);
tok = pdf_lex(file, buf);
/* RJW: cannot parse indirect object (%d %d R)", num, gen */
if (tok != PDF_TOK_INT)
fz_throw(ctx, "expected object number (%d %d R)", num, gen);
num = buf->i;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
if (tok != PDF_TOK_INT)
fz_throw(ctx, "expected generation number (%d %d R)", num, gen);
gen = buf->i;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
if (tok != PDF_TOK_OBJ)
fz_throw(ctx, "expected 'obj' keyword (%d %d R)", num, gen);
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
switch (tok)
{
case PDF_TOK_OPEN_ARRAY:
obj = pdf_parse_array(xref, file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
break;
case PDF_TOK_OPEN_DICT:
obj = pdf_parse_dict(xref, file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen */
break;
case PDF_TOK_NAME:
obj = fz_new_name(ctx, buf->scratch);
break;
case PDF_TOK_REAL:
obj = pdf_new_real(ctx, buf->f);
break;
case PDF_TOK_STRING:
obj = pdf_new_string(ctx, buf->scratch, buf->len);
break;
case PDF_TOK_TRUE:
obj = pdf_new_bool(ctx, 1);
break;
case PDF_TOK_FALSE:
obj = pdf_new_bool(ctx, 0);
break;
case PDF_TOK_NULL:
obj = pdf_new_null(ctx);
break;
case PDF_TOK_INT:
a = buf->i;
tok = pdf_lex(file, buf);
/* "cannot parse indirect object (%d %d R)", num, gen */
if (tok == PDF_TOK_STREAM || tok == PDF_TOK_ENDOBJ)
{
obj = pdf_new_int(ctx, a);
goto skip;
}
if (tok == PDF_TOK_INT)
{
b = buf->i;
tok = pdf_lex(file, buf);
/* RJW: "cannot parse indirect object (%d %d R)", num, gen); */
if (tok == PDF_TOK_R)
{
obj = pdf_new_indirect(ctx, a, b, xref);
break;
}
}
fz_throw(ctx, "expected 'R' keyword (%d %d R)", num, gen);
case PDF_TOK_ENDOBJ:
obj = pdf_new_null(ctx);
goto skip;
default:
fz_throw(ctx, "syntax error in object (%d %d R)", num, gen);
}
fz_try(ctx)
{
tok = pdf_lex(file, buf);
}
fz_catch(ctx)
{
pdf_drop_obj(obj);
fz_throw(ctx, "cannot parse indirect object (%d %d R)", num, gen);
}
skip:
if (tok == PDF_TOK_STREAM)
{
int c = fz_read_byte(file);
while (c == ' ')
c = fz_read_byte(file);
if (c == '\r')
{
c = fz_peek_byte(file);
if (c != '\n')
fz_warn(ctx, "line feed missing after stream begin marker (%d %d R)", num, gen);
else
fz_read_byte(file);
}
stm_ofs = fz_tell(file);
}
else if (tok == PDF_TOK_ENDOBJ)
{
stm_ofs = 0;
}
else
{
fz_warn(ctx, "expected 'endobj' or 'stream' keyword (%d %d R)", num, gen);
stm_ofs = 0;
}
if (onum) *onum = num;
if (ogen) *ogen = gen;
if (ostmofs) *ostmofs = stm_ofs;
return obj;
}
int pdf_add_portfolio_entry(fz_context *ctx, pdf_document *doc,
const char *name, int name_len,
const char *desc, int desc_len,
const char *filename, int filename_len,
const char *unifile, int unifile_len, fz_buffer *buf)
{
int entry, len;
pdf_obj *ef, *f, *params, *s;
pdf_obj *key;
pdf_obj *val = NULL;
fz_var(val);
if (!doc)
fz_throw(ctx, FZ_ERROR_GENERIC, "Bad pdf_add_portfolio_entry call");
if (doc->portfolio == NULL)
load_portfolio(ctx, doc);
key = pdf_new_string(ctx, doc, name, name_len);
fz_try(ctx)
{
val = pdf_new_dict(ctx, doc, 6);
pdf_dict_put_drop(ctx, val, PDF_NAME_CI, pdf_new_dict(ctx, doc, 4));
pdf_dict_put_drop(ctx, val, PDF_NAME_EF, (ef = pdf_new_dict(ctx, doc, 4)));
pdf_dict_put_drop(ctx, val, PDF_NAME_F, pdf_new_string(ctx, doc, filename, filename_len));
pdf_dict_put_drop(ctx, val, PDF_NAME_UF, pdf_new_string(ctx, doc, unifile, unifile_len));
pdf_dict_put_drop(ctx, val, PDF_NAME_Desc, pdf_new_string(ctx, doc, desc, desc_len));
pdf_dict_put_drop(ctx, val, PDF_NAME_Type, PDF_NAME_Filespec);
pdf_dict_put_drop(ctx, ef, PDF_NAME_F, (f = pdf_add_stream(ctx, doc, buf, NULL, 0)));
len = fz_buffer_storage(ctx, buf, NULL);
pdf_dict_put_drop(ctx, f, PDF_NAME_DL, pdf_new_int(ctx, doc, len));
pdf_dict_put_drop(ctx, f, PDF_NAME_Length, pdf_new_int(ctx, doc, len));
pdf_dict_put_drop(ctx, f, PDF_NAME_Params, (params = pdf_new_dict(ctx, doc, 4)));
pdf_dict_put_drop(ctx, params, PDF_NAME_Size, pdf_new_int(ctx, doc, len));
s = pdf_dict_getl(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root, PDF_NAME_Collection, NULL);
if (s == NULL)
{
s = pdf_new_dict(ctx, doc, 4);
pdf_dict_putl_drop(ctx, pdf_trailer(ctx, doc), s, PDF_NAME_Root, PDF_NAME_Collection, NULL);
}
s = pdf_dict_getl(ctx, pdf_trailer(ctx, doc), PDF_NAME_Root, PDF_NAME_Names, PDF_NAME_EmbeddedFiles, NULL);
if (s == NULL)
{
s = pdf_new_dict(ctx, doc, 4);
pdf_dict_putl_drop(ctx, pdf_trailer(ctx, doc), s, PDF_NAME_Root, PDF_NAME_Names, PDF_NAME_EmbeddedFiles, NULL);
}
entry = pdf_name_tree_insert(ctx, doc, s, key, val);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, key);
pdf_drop_obj(ctx, val);
}
fz_catch(ctx)
fz_rethrow(ctx);
return entry;
}
