ErrorCode juggler_add_pages_from_file(juggler_t *dest, juggler_t *src, int dest_index)
{
pdf_obj *dest_pages = pdf_dict_getp(dest->ctx,
pdf_trailer(dest->ctx, dest->pdf), "Root/Pages");
int dest_pages_index = pdf_array_len(dest->ctx,
pdf_dict_gets(dest->ctx, dest_pages, "Kids"));
/* be aware that this function does not change the two variables if the page
index is greater than the number of pages */
find_destination_pages(dest->ctx,
dest_pages, dest_index, &dest_pages, &dest_pages_index);
pdf_obj *dest_kids = pdf_dict_gets(dest->ctx, dest_pages, "Kids");
if(!pdf_is_indirect(dest->ctx, dest_pages) ||
!pdf_is_dict(dest->ctx, dest_pages) || !pdf_is_array(dest->ctx, dest_kids))
{
return(ERROR_INVALID_RANGE);
}
pdf_obj *pages_root = pdf_dict_getp(src->ctx, pdf_trailer(src->ctx, src->pdf), "Root/Pages");
if(!pdf_is_indirect(src->ctx, pages_root) || !pdf_is_dict(src->ctx, pages_root))
return(ERROR_NO_PAGES);
/* if we copy the root pages-node and it's referenced objects, we will copy
all pages and all objects those pages need */
pdf_obj *new_pages_ref = copy_object_single(dest->ctx, dest->pdf, src->ctx, src->pdf, pages_root);
/* insert new pages-node */
pdf_array_insert_drop(dest->ctx, dest_kids, new_pages_ref, dest_pages_index);
/* update the parent */
pdf_obj *new_pages_parent = pdf_new_indirect(dest->ctx, dest->pdf,
pdf_to_num(dest->ctx, dest_pages), pdf_to_gen(dest->ctx, dest_pages));
pdf_dict_puts_drop(dest->ctx, new_pages_ref, "Parent", new_pages_parent);
/* TODO: If dest_pages contains anything inheritable but not the new node
we need to insert empty items to prevent this inerhitance */
/* update count */
int new_count = pdf_to_int(dest->ctx,
pdf_dict_gets(dest->ctx, dest_pages, "Count")) + src->pagecount;
pdf_dict_puts_drop(dest->ctx, dest_pages, "Count",
pdf_new_int(dest->ctx, dest->pdf, new_count));
/* let MuPDF rebuild the page tree */
pdf_finish_edit(dest->ctx, dest->pdf);
dest->pdf->page_count = new_count;
/* update juggler's state */
juggler_page_tree_changed_due_to_insert(dest, dest_index, src->pagecount);
return(NoError);
}
static pdf_obj *
resolve_dest_rec(fz_context *ctx, pdf_document *doc, pdf_obj *dest, int depth)
{
if (depth > 10) /* Arbitrary to avoid infinite recursion */
return NULL;
if (pdf_is_name(ctx, dest) || pdf_is_string(ctx, dest))
{
dest = pdf_lookup_dest(ctx, doc, dest);
dest = resolve_dest_rec(ctx, doc, dest, depth+1);
return dest;
}
else if (pdf_is_array(ctx, dest))
{
return dest;
}
else if (pdf_is_dict(ctx, dest))
{
dest = pdf_dict_get(ctx, dest, PDF_NAME_D);
return resolve_dest_rec(ctx, doc, dest, depth+1);
}
else if (pdf_is_indirect(ctx, dest))
return dest;
return NULL;
}
/* returns the number of attachments saved */
int save_attachments(int pageno, char *targetdir)
{
pdf_page *page = pdf_load_page(doc, pageno-1);
pdf_annot *annot;
int saved_count = 0;
for (annot = page->annots; annot ; annot = annot->next) {
pdf_obj *fs_obj = pdf_dict_gets(annot->obj, "FS");
if (fs_obj) {
pdf_obj *ef_obj;
char *name = basename(strdup(pdf_to_str_buf(pdf_dict_gets(fs_obj, "F"))));
ef_obj = pdf_dict_gets(fs_obj, "EF");
if (ef_obj) {
pdf_obj *f_obj = pdf_dict_gets(ef_obj, "F");
if (f_obj && pdf_is_indirect(f_obj)) {
static char pathname[PATH_MAX];
sprintf(pathname, "%s/%s", targetdir, name);
FILE *fout = fopen(pathname, "w");
if (!fout) {
fprintf(stderr, "extr: cannot write to file %s\n", pathname);
exit(1);
}
dump_stream(pdf_to_num(f_obj), fout);
fclose(fout);
saved_count++;
}
}
}
}
return saved_count;
}
static pdf_obj *
resolve_dest_rec(pdf_document *xref, pdf_obj *dest, int depth)
{
if (depth > 10) /* Arbitrary to avoid infinite recursion */
return NULL;
if (pdf_is_name(dest) || pdf_is_string(dest))
{
dest = pdf_lookup_dest(xref, dest);
return resolve_dest_rec(xref, dest, depth+1);
}
else if (pdf_is_array(dest))
{
return dest;
}
else if (pdf_is_dict(dest))
{
dest = pdf_dict_gets(dest, "D");
return resolve_dest_rec(xref, dest, depth+1);
}
else if (pdf_is_indirect(dest))
return dest;
return NULL;
}
/*
* Create a filter given a name and param dictionary.
*/
static fz_stream *
build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, pdf_obj *p, int num, int gen, fz_compression_params *params)
{
fz_compression_params local_params;
if (params == NULL)
params = &local_params;
build_compression_params(ctx, f, p, params);
/* If we were using params we were passed in, and we successfully
* recognised the image type, we can use the existing filter and
* shortstop here. */
if (params != &local_params && params->type != FZ_IMAGE_RAW)
return chain;
if (params->type != FZ_IMAGE_RAW)
return fz_open_image_decomp_stream(ctx, chain, params, NULL);
if (pdf_name_eq(ctx, f, PDF_NAME_ASCIIHexDecode) || pdf_name_eq(ctx, f, PDF_NAME_AHx))
return fz_open_ahxd(ctx, chain);
else if (pdf_name_eq(ctx, f, PDF_NAME_ASCII85Decode) || pdf_name_eq(ctx, f, PDF_NAME_A85))
return fz_open_a85d(ctx, chain);
else if (pdf_name_eq(ctx, f, PDF_NAME_JBIG2Decode))
{
fz_jbig2_globals *globals = NULL;
pdf_obj *obj = pdf_dict_get(ctx, p, PDF_NAME_JBIG2Globals);
if (pdf_is_indirect(ctx, obj))
globals = pdf_load_jbig2_globals(ctx, doc, obj);
/* fz_open_jbig2d takes possession of globals */
return fz_open_jbig2d(ctx, chain, globals);
}
else if (pdf_name_eq(ctx, f, PDF_NAME_JPXDecode))
return chain; /* JPX decoding is special cased in the image loading code */
else if (pdf_name_eq(ctx, f, PDF_NAME_Crypt))
{
pdf_obj *name;
if (!doc->crypt)
{
fz_warn(ctx, "crypt filter in unencrypted document");
return chain;
}
name = pdf_dict_get(ctx, p, PDF_NAME_Name);
if (pdf_is_name(ctx, name))
return pdf_open_crypt_with_filter(ctx, chain, doc->crypt, name, num, gen);
return chain;
}
fz_warn(ctx, "unknown filter name (%s)", pdf_to_name(ctx, f));
return chain;
}
static void
pdf_print_key(fz_context *ctx, fz_output *out, void *key_)
{
pdf_obj *key = (pdf_obj *)key_;
if (pdf_is_indirect(ctx, key))
fz_printf(ctx, out, "(%d %d R) ", pdf_to_num(ctx, key), pdf_to_gen(ctx, key));
else
pdf_print_obj(ctx, out, key, 0);
}
static void
pdf_debug_key(void *key_)
{
pdf_obj *key = (pdf_obj *)key_;
if (pdf_is_indirect(key))
{
printf("(%d %d R) ", pdf_to_num(key), pdf_to_gen(key));
} else
pdf_print_obj(key);
}
static int
pdf_make_hash_key(fz_store_hash *hash, void *key_)
{
pdf_obj *key = (pdf_obj *)key_;
if (!pdf_is_indirect(key))
return 0;
hash->u.i.i0 = pdf_to_num(key);
hash->u.i.i1 = pdf_to_gen(key);
return 1;
}
static void
pdf_debug_key(FILE *out, void *key_)
{
pdf_obj *key = (pdf_obj *)key_;
if (pdf_is_indirect(key))
{
fprintf(out, "(%d %d R) ", pdf_to_num(key), pdf_to_gen(key));
}
else
pdf_fprint_obj(out, key, 0);
}
static int
pdf_make_hash_key(fz_context *ctx, fz_store_hash *hash, void *key_)
{
pdf_obj *key = (pdf_obj *)key_;
if (!pdf_is_indirect(ctx, key))
return 0;
hash->u.i.i0 = pdf_to_num(ctx, key);
hash->u.i.i1 = pdf_to_gen(ctx, key);
hash->u.i.ptr = pdf_get_indirect_document(ctx, key);
return 1;
}
// 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);
}
int
pdf_repair_obj(fz_context *ctx, pdf_document *doc, pdf_lexbuf *buf, fz_off_t *stmofsp, int *stmlenp, pdf_obj **encrypt, pdf_obj **id, pdf_obj **page, fz_off_t *tmpofs, pdf_obj **root)
{
fz_stream *file = doc->file;
pdf_token tok;
int stm_len;
*stmofsp = 0;
if (stmlenp)
*stmlenp = -1;
stm_len = 0;
/* On entry to this function, we know that we've just seen
* '<int> <int> obj'. We expect the next thing we see to be a
* pdf object. Regardless of the type of thing we meet next
* we only need to fully parse it if it is a dictionary. */
tok = pdf_lex(ctx, file, buf);
if (tok == PDF_TOK_OPEN_DICT)
{
pdf_obj *dict, *obj;
fz_try(ctx)
{
dict = pdf_parse_dict(ctx, doc, file, buf);
}
fz_catch(ctx)
{
fz_rethrow_if(ctx, FZ_ERROR_TRYLATER);
/* Don't let a broken object at EOF overwrite a good one */
if (file->eof)
fz_rethrow(ctx);
/* Silently swallow the error */
dict = pdf_new_dict(ctx, NULL, 2);
}
/* We must be careful not to try to resolve any indirections
* here. We have just read dict, so we know it to be a non
* indirected dictionary. Before we look at any values that
* we get back from looking up in it, we need to check they
* aren't indirected. */
if (encrypt || id || root)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Type);
if (!pdf_is_indirect(ctx, obj) && pdf_name_eq(ctx, obj, PDF_NAME_XRef))
{
if (encrypt)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Encrypt);
if (obj)
{
pdf_drop_obj(ctx, *encrypt);
*encrypt = pdf_keep_obj(ctx, obj);
}
}
if (id)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_ID);
if (obj)
{
pdf_drop_obj(ctx, *id);
*id = pdf_keep_obj(ctx, obj);
}
}
if (root)
*root = pdf_keep_obj(ctx, pdf_dict_get(ctx, dict, PDF_NAME_Root));
}
}
obj = pdf_dict_get(ctx, dict, PDF_NAME_Length);
if (!pdf_is_indirect(ctx, obj) && pdf_is_int(ctx, obj))
stm_len = pdf_to_int(ctx, obj);
if (doc->file_reading_linearly && page)
{
obj = pdf_dict_get(ctx, dict, PDF_NAME_Type);
if (!pdf_is_indirect(ctx, obj) && pdf_name_eq(ctx, obj, PDF_NAME_Page))
{
pdf_drop_obj(ctx, *page);
*page = pdf_keep_obj(ctx, dict);
}
}
pdf_drop_obj(ctx, dict);
}
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;
}
/*
* Load CMap stream in PDF file
*/
pdf_cmap *
pdf_load_embedded_cmap(fz_context *ctx, pdf_document *doc, pdf_obj *stmobj)
{
fz_stream *file = NULL;
pdf_cmap *cmap = NULL;
pdf_cmap *usecmap = NULL;
pdf_obj *obj;
fz_var(file);
fz_var(cmap);
fz_var(usecmap);
if (pdf_obj_marked(ctx, stmobj))
fz_throw(ctx, FZ_ERROR_GENERIC, "Recursion in embedded cmap");
if ((cmap = pdf_find_item(ctx, pdf_drop_cmap_imp, stmobj)) != NULL)
return cmap;
fz_try(ctx)
{
file = pdf_open_stream(ctx, stmobj);
cmap = pdf_load_cmap(ctx, file);
obj = pdf_dict_get(ctx, stmobj, PDF_NAME_WMode);
if (pdf_is_int(ctx, obj))
pdf_set_cmap_wmode(ctx, cmap, pdf_to_int(ctx, obj));
obj = pdf_dict_get(ctx, stmobj, PDF_NAME_UseCMap);
if (pdf_is_name(ctx, obj))
{
usecmap = pdf_load_system_cmap(ctx, pdf_to_name(ctx, obj));
pdf_set_usecmap(ctx, cmap, usecmap);
}
else if (pdf_is_indirect(ctx, obj))
{
if (pdf_mark_obj(ctx, obj))
fz_throw(ctx, FZ_ERROR_GENERIC, "recursive CMap");
fz_try(ctx)
usecmap = pdf_load_embedded_cmap(ctx, doc, obj);
fz_always(ctx)
pdf_unmark_obj(ctx, obj);
fz_catch(ctx)
fz_rethrow(ctx);
pdf_set_usecmap(ctx, cmap, usecmap);
}
pdf_store_item(ctx, stmobj, cmap, pdf_cmap_size(ctx, cmap));
}
fz_always(ctx)
{
fz_drop_stream(ctx, file);
pdf_drop_cmap(ctx, usecmap);
}
fz_catch(ctx)
{
pdf_drop_cmap(ctx, cmap);
fz_rethrow(ctx);
}
return cmap;
}
static void wmupdf_convert_single_page_to_form(pdf_document *xref,fz_context *ctx,int pageno)
{
pdf_obj *array,*srcpageobj,*srcpagecontents;
int i,len,streamlen,pageref,pagegen,compressed;
double bbox_array[4];
double matrix[6];
/* New source page, so get the source page objects */
srcpageobj = xref->page_objs[pageno-1];
pageref=pdf_to_num(xref->page_refs[pageno-1]);
pagegen=pdf_to_gen(xref->page_refs[pageno-1]);
wmupdf_page_bbox(srcpageobj,bbox_array);
for (i=0;i<6;i++)
matrix[i]=0.;
matrix[0]=matrix[3]=1.;
srcpagecontents=pdf_dict_gets(srcpageobj,"Contents");
/* Concatenate all indirect streams from source page directly into it. */
// printf("Adding streams to source page %d (pageref=%d, pagegen=%d)...\n",pageno,pageref,pagegen);
streamlen=0;
if (pdf_is_array(srcpagecontents))
{
int k;
for (k=0;k<pdf_array_len(srcpagecontents);k++)
{
pdf_obj *obj;
obj=pdf_array_get(srcpagecontents,k);
if (pdf_is_indirect(obj))
pdf_resolve_indirect(obj);
streamlen=add_to_srcpage_stream(xref,ctx,pageref,pagegen,obj);
}
}
else
{
if (pdf_is_indirect(srcpagecontents))
pdf_resolve_indirect(srcpagecontents);
streamlen=add_to_srcpage_stream(xref,ctx,pageref,pagegen,srcpagecontents);
}
compressed=stream_deflate(xref,ctx,pageref,pagegen,&streamlen);
srcpageobj = xref->page_objs[pageno-1];
pageref=pdf_to_num(xref->page_refs[pageno-1]);
len=pdf_dict_len(srcpageobj);
for (i=0;i<len;i++)
{
pdf_obj *key; /* *value */
key=pdf_dict_get_key(srcpageobj,i);
/*
if (pdf_is_name(key))
printf("key[%d] = name = %s\n",i,pdf_to_name(key));
else
printf("key[%d] = ??\n",i);
*/
/* value=pdf_dict_get_val(srcpageobj,i); */
/* Keep same resources */
if (!pdf_is_name(key))
continue;
if (pdf_is_name(key) && !stricmp("Resources",pdf_to_name(key)))
continue;
/* Drop dictionary entry otherwise */
// printf("Deleting key %s.\n",pdf_to_name(key));
pdf_dict_del(srcpageobj,key);
i=-1;
len=pdf_dict_len(srcpageobj);
}
pdf_dict_puts(srcpageobj,"Type",pdf_new_name(ctx,"XObject"));
pdf_dict_puts(srcpageobj,"Subtype",pdf_new_name(ctx,"Form"));
pdf_dict_puts(srcpageobj,"FormType",pdf_new_int(ctx,1));
if (compressed)
pdf_dict_puts(srcpageobj,"Filter",pdf_new_name(ctx,"FlateDecode"));
pdf_dict_puts(srcpageobj,"Length",pdf_new_int(ctx,streamlen));
array=pdf_new_array(ctx,4);
for (i=0;i<4;i++)
pdf_array_push(array,pdf_new_real(ctx,bbox_array[i]));
pdf_dict_puts(srcpageobj,"BBox",array);
array=pdf_new_array(ctx,6);
for (i=0;i<6;i++)
pdf_array_push(array,pdf_new_real(ctx,matrix[i]));
pdf_dict_puts(srcpageobj,"Matrix",array);
/* (It's no longer a "page"--it's a Form-type XObject) */
/* I don't think this call should be made since it will call fz_drop_object on srcpageobj */
/* pdf_update_object(xref,pageref,srcpageobj); */
}
/*
* Load CMap stream in PDF file
*/
pdf_cmap *
pdf_load_embedded_cmap(pdf_document *doc, pdf_obj *stmobj)
{
fz_stream *file = NULL;
pdf_cmap *cmap = NULL;
pdf_cmap *usecmap;
pdf_obj *wmode;
pdf_obj *obj = NULL;
fz_context *ctx = doc->ctx;
int phase = 0;
fz_var(phase);
fz_var(obj);
fz_var(file);
fz_var(cmap);
if (pdf_obj_marked(stmobj))
fz_throw(ctx, FZ_ERROR_GENERIC, "Recursion in embedded cmap");
if ((cmap = pdf_find_item(ctx, pdf_free_cmap_imp, stmobj)) != NULL)
{
return cmap;
}
fz_try(ctx)
{
file = pdf_open_stream(doc, pdf_to_num(stmobj), pdf_to_gen(stmobj));
phase = 1;
cmap = pdf_load_cmap(ctx, file);
phase = 2;
fz_close(file);
file = NULL;
wmode = pdf_dict_gets(stmobj, "WMode");
if (pdf_is_int(wmode))
pdf_set_cmap_wmode(ctx, cmap, pdf_to_int(wmode));
obj = pdf_dict_gets(stmobj, "UseCMap");
if (pdf_is_name(obj))
{
usecmap = pdf_load_system_cmap(ctx, pdf_to_name(obj));
pdf_set_usecmap(ctx, cmap, usecmap);
pdf_drop_cmap(ctx, usecmap);
}
else if (pdf_is_indirect(obj))
{
phase = 3;
pdf_mark_obj(obj);
usecmap = pdf_load_embedded_cmap(doc, obj);
pdf_unmark_obj(obj);
phase = 4;
pdf_set_usecmap(ctx, cmap, usecmap);
pdf_drop_cmap(ctx, usecmap);
}
pdf_store_item(ctx, stmobj, cmap, pdf_cmap_size(ctx, cmap));
}
fz_catch(ctx)
{
if (file)
fz_close(file);
if (cmap)
pdf_drop_cmap(ctx, cmap);
if (phase < 1)
fz_rethrow_message(ctx, "cannot open cmap stream (%d %d R)", pdf_to_num(stmobj), pdf_to_gen(stmobj));
else if (phase < 2)
fz_rethrow_message(ctx, "cannot parse cmap stream (%d %d R)", pdf_to_num(stmobj), pdf_to_gen(stmobj));
else if (phase < 3)
fz_rethrow_message(ctx, "cannot load system usecmap '%s'", pdf_to_name(obj));
else
{
if (phase == 3)
pdf_unmark_obj(obj);
fz_rethrow_message(ctx, "cannot load embedded usecmap (%d %d R)", pdf_to_num(obj), pdf_to_gen(obj));
}
}
return cmap;
}
/* dest points to the new pages content-streams-dict, src is a reference to
one source content-stream */
int copy_content_stream_of_page(fz_context *dest_ctx, pdf_obj *dest,
fz_context *src_ctx, pdf_obj *src,
struct put_info *info, struct pos_info *pos)
{
if(!pdf_is_array(dest_ctx, dest) && !pdf_is_indirect(src_ctx, src))
return(-1);
/*
translation: 1 0 0 1 diff_x diff_y
scale: scale 0 0 scale 0 0
rotation: cos sin -sin cos 0 0
-------------------------------------------------
rotation 0: 1 0 0 1 0 0
rotation 90: 0 1 -1 0 0 0
rotation 180: -1 0 0 -1 0 0
rotation 270: 0 -1 1 0 0 0
*/
fz_buffer *buffer = fz_new_buffer(dest_ctx, 1024);
fz_output *output = fz_new_output_with_buffer(dest_ctx, buffer);
fz_printf(dest_ctx, output, "q\n");
/* set the outer clip region */
fz_printf(dest_ctx, output, "%f %f %f %f re W n\n",
pos->outer_clip_x, pos->outer_clip_y,
pos->outer_clip_width, pos->outer_clip_height);
/* position the page correctly */
if(pos->rotate == 0) {
fz_printf(dest_ctx, output, "1 0 0 1 %f %f cm\n",
pos->x + pos->content_translate_x, pos->y + pos->content_translate_y);
} else if(pos->rotate == 90) {
fz_printf(dest_ctx, output, "0 1 -1 0 %f %f cm\n", pos->x + pos->width, pos->y);
} else if(pos->rotate == 180) {
fz_printf(dest_ctx, output, "-1 0 0 -1 %f %f cm\n",
pos->width + pos->x - pos->content_translate_x,
pos->height + pos->y - pos->content_translate_y);
} else if(pos->rotate == 270) {
fz_printf(dest_ctx, output, "0 -1 1 0 %f %f cm\n", pos->x, pos->y + pos->height);
}
if(pos->bleed_clip_x != 0.0 || pos->bleed_clip_y != 0.0 ||
pos->bleed_clip_width != 0.0 || pos->bleed_clip_height != 0.0)
{
fz_printf(dest_ctx, output, "%f %f %f %f re W n\n",
pos->bleed_clip_x, pos->bleed_clip_y, pos->bleed_clip_width, pos->bleed_clip_height);
}
int src_num = pdf_to_num(src_ctx, src);
int src_gen = pdf_to_gen(src_ctx, src);
fz_stream *input = pdf_open_stream(src_ctx, info->src_doc, src_num, src_gen);
rename_res_in_content_stream(src_ctx, input, dest_ctx, output, info->rename_dict);
fz_printf(dest_ctx, output, "Q");
fz_drop_output(dest_ctx, output);
fz_drop_stream(dest_ctx, input);
int new_num = pdf_create_object(dest_ctx, info->dest_doc);
pdf_obj *new_ref = pdf_new_indirect(dest_ctx, info->dest_doc, new_num, 0);
/* each stream has a dict containing at least its length... */
pdf_obj *stream_info_dict = pdf_new_dict(dest_ctx, info->dest_doc, 1);
pdf_dict_puts_drop(dest_ctx, stream_info_dict, "Length", pdf_new_int(dest_ctx, info->dest_doc, buffer->len));
pdf_update_object(dest_ctx, info->dest_doc, new_num, stream_info_dict);
pdf_drop_obj(dest_ctx, stream_info_dict);
pdf_update_stream(dest_ctx, info->dest_doc, new_ref, buffer, 0);
fz_drop_buffer(dest_ctx, buffer);
pdf_array_push(dest_ctx, dest, new_ref);
pdf_drop_obj(dest_ctx, new_ref);
return(0);
}
static pdf_font_desc *
load_cid_font(pdf_document *xref, pdf_obj *dict, pdf_obj *encoding, pdf_obj *to_unicode)
{
pdf_obj *widths;
pdf_obj *descriptor;
pdf_font_desc *fontdesc = NULL;
FT_Face face;
int kind;
char collection[256];
char *basefont;
int i, k, fterr;
pdf_obj *obj;
int dw;
fz_context *ctx = xref->ctx;
fz_var(fontdesc);
fz_try(ctx)
{
/* Get font name and CID collection */
basefont = pdf_to_name(pdf_dict_gets(dict, "BaseFont"));
{
pdf_obj *cidinfo;
char tmpstr[64];
int tmplen;
cidinfo = pdf_dict_gets(dict, "CIDSystemInfo");
if (!cidinfo)
fz_throw(ctx, "cid font is missing info");
obj = pdf_dict_gets(cidinfo, "Registry");
tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
tmpstr[tmplen] = '\0';
fz_strlcpy(collection, tmpstr, sizeof collection);
fz_strlcat(collection, "-", sizeof collection);
obj = pdf_dict_gets(cidinfo, "Ordering");
tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
tmpstr[tmplen] = '\0';
fz_strlcat(collection, tmpstr, sizeof collection);
}
/* Load font file */
fontdesc = pdf_new_font_desc(ctx);
descriptor = pdf_dict_gets(dict, "FontDescriptor");
if (!descriptor)
fz_throw(ctx, "syntaxerror: missing font descriptor");
pdf_load_font_descriptor(fontdesc, xref, descriptor, collection, basefont);
face = fontdesc->font->ft_face;
kind = ft_kind(face);
/* Encoding */
if (pdf_is_name(encoding))
{
if (!strcmp(pdf_to_name(encoding), "Identity-H"))
fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 2);
else if (!strcmp(pdf_to_name(encoding), "Identity-V"))
fontdesc->encoding = pdf_new_identity_cmap(ctx, 1, 2);
else
fontdesc->encoding = pdf_load_system_cmap(ctx, pdf_to_name(encoding));
}
else if (pdf_is_indirect(encoding))
{
fontdesc->encoding = pdf_load_embedded_cmap(xref, encoding);
}
else
{
fz_throw(ctx, "syntaxerror: font missing encoding");
}
fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);
pdf_set_font_wmode(ctx, fontdesc, pdf_cmap_wmode(ctx, fontdesc->encoding));
if (kind == TRUETYPE)
{
pdf_obj *cidtogidmap;
cidtogidmap = pdf_dict_gets(dict, "CIDToGIDMap");
if (pdf_is_indirect(cidtogidmap))
{
fz_buffer *buf;
buf = pdf_load_stream(xref, pdf_to_num(cidtogidmap), pdf_to_gen(cidtogidmap));
fontdesc->cid_to_gid_len = (buf->len) / 2;
fontdesc->cid_to_gid = fz_malloc_array(ctx, fontdesc->cid_to_gid_len, sizeof(unsigned short));
fontdesc->size += fontdesc->cid_to_gid_len * sizeof(unsigned short);
for (i = 0; i < fontdesc->cid_to_gid_len; i++)
fontdesc->cid_to_gid[i] = (buf->data[i * 2] << 8) + buf->data[i * 2 + 1];
fz_drop_buffer(ctx, buf);
}
/* if truetype font is external, cidtogidmap should not be identity */
/* so we map from cid to unicode and then map that through the (3 1) */
/* unicode cmap to get a glyph id */
else if (fontdesc->font->ft_substitute)
{
fterr = FT_Select_Charmap(face, ft_encoding_unicode);
if (fterr)
{
fz_throw(ctx, "fonterror: no unicode cmap when emulating CID font: %s", ft_error_string(fterr));
}
if (!strcmp(collection, "Adobe-CNS1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-CNS1-UCS2");
else if (!strcmp(collection, "Adobe-GB1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2");
else if (!strcmp(collection, "Adobe-Japan1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan1-UCS2");
else if (!strcmp(collection, "Adobe-Japan2"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan2-UCS2");
else if (!strcmp(collection, "Adobe-Korea1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Korea1-UCS2");
}
}
pdf_load_to_unicode(xref, fontdesc, NULL, collection, to_unicode);
/* Horizontal */
dw = 1000;
obj = pdf_dict_gets(dict, "DW");
if (obj)
dw = pdf_to_int(obj);
pdf_set_default_hmtx(ctx, fontdesc, dw);
widths = pdf_dict_gets(dict, "W");
if (widths)
{
int c0, c1, w, n, m;
n = pdf_array_len(widths);
for (i = 0; i < n; )
{
c0 = pdf_to_int(pdf_array_get(widths, i));
obj = pdf_array_get(widths, i + 1);
if (pdf_is_array(obj))
{
m = pdf_array_len(obj);
for (k = 0; k < m; k++)
{
w = pdf_to_int(pdf_array_get(obj, k));
pdf_add_hmtx(ctx, fontdesc, c0 + k, c0 + k, w);
}
i += 2;
}
else
{
c1 = pdf_to_int(obj);
w = pdf_to_int(pdf_array_get(widths, i + 2));
pdf_add_hmtx(ctx, fontdesc, c0, c1, w);
i += 3;
}
}
}
pdf_end_hmtx(ctx, fontdesc);
/* Vertical */
if (pdf_cmap_wmode(ctx, fontdesc->encoding) == 1)
{
int dw2y = 880;
int dw2w = -1000;
obj = pdf_dict_gets(dict, "DW2");
if (obj)
{
dw2y = pdf_to_int(pdf_array_get(obj, 0));
dw2w = pdf_to_int(pdf_array_get(obj, 1));
}
pdf_set_default_vmtx(ctx, fontdesc, dw2y, dw2w);
widths = pdf_dict_gets(dict, "W2");
if (widths)
{
int c0, c1, w, x, y, n;
n = pdf_array_len(widths);
for (i = 0; i < n; )
{
c0 = pdf_to_int(pdf_array_get(widths, i));
obj = pdf_array_get(widths, i + 1);
if (pdf_is_array(obj))
{
int m = pdf_array_len(obj);
for (k = 0; k * 3 < m; k ++)
{
w = pdf_to_int(pdf_array_get(obj, k * 3 + 0));
x = pdf_to_int(pdf_array_get(obj, k * 3 + 1));
y = pdf_to_int(pdf_array_get(obj, k * 3 + 2));
pdf_add_vmtx(ctx, fontdesc, c0 + k, c0 + k, x, y, w);
}
i += 2;
}
else
{
c1 = pdf_to_int(obj);
w = pdf_to_int(pdf_array_get(widths, i + 2));
x = pdf_to_int(pdf_array_get(widths, i + 3));
y = pdf_to_int(pdf_array_get(widths, i + 4));
pdf_add_vmtx(ctx, fontdesc, c0, c1, x, y, w);
i += 5;
}
}
}
pdf_end_vmtx(ctx, fontdesc);
}
}
fz_catch(ctx)
{
pdf_drop_font(ctx, fontdesc);
fz_throw(ctx, "cannot load cid font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
}
return fontdesc;
}
static pdf_font_desc *
load_cid_font(pdf_document *doc, pdf_obj *dict, pdf_obj *encoding, pdf_obj *to_unicode)
{
pdf_obj *widths;
pdf_obj *descriptor;
pdf_font_desc *fontdesc = NULL;
FT_Face face;
int kind;
char collection[256];
char *basefont;
int i, k, fterr;
pdf_obj *obj;
int dw;
fz_context *ctx = doc->ctx;
fz_var(fontdesc);
fz_try(ctx)
{
/* Get font name and CID collection */
basefont = pdf_to_name(pdf_dict_gets(dict, "BaseFont"));
{
pdf_obj *cidinfo;
char tmpstr[64];
int tmplen;
cidinfo = pdf_dict_gets(dict, "CIDSystemInfo");
if (!cidinfo)
fz_throw(ctx, FZ_ERROR_GENERIC, "cid font is missing info");
obj = pdf_dict_gets(cidinfo, "Registry");
tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
tmpstr[tmplen] = '\0';
fz_strlcpy(collection, tmpstr, sizeof collection);
fz_strlcat(collection, "-", sizeof collection);
obj = pdf_dict_gets(cidinfo, "Ordering");
tmplen = fz_mini(sizeof tmpstr - 1, pdf_to_str_len(obj));
memcpy(tmpstr, pdf_to_str_buf(obj), tmplen);
tmpstr[tmplen] = '\0';
fz_strlcat(collection, tmpstr, sizeof collection);
}
/* Load font file */
fontdesc = pdf_new_font_desc(ctx);
descriptor = pdf_dict_gets(dict, "FontDescriptor");
if (!descriptor)
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: missing font descriptor");
pdf_load_font_descriptor(fontdesc, doc, descriptor, collection, basefont, 1, 1);
face = fontdesc->font->ft_face;
kind = ft_kind(face);
/* Encoding */
if (pdf_is_name(encoding))
{
if (!strcmp(pdf_to_name(encoding), "Identity-H"))
fontdesc->encoding = pdf_new_identity_cmap(ctx, 0, 2);
else if (!strcmp(pdf_to_name(encoding), "Identity-V"))
fontdesc->encoding = pdf_new_identity_cmap(ctx, 1, 2);
else
fontdesc->encoding = pdf_load_system_cmap(ctx, pdf_to_name(encoding));
}
else if (pdf_is_indirect(encoding))
{
fontdesc->encoding = pdf_load_embedded_cmap(doc, encoding);
}
else
{
fz_throw(ctx, FZ_ERROR_GENERIC, "syntaxerror: font missing encoding");
}
fontdesc->size += pdf_cmap_size(ctx, fontdesc->encoding);
pdf_set_font_wmode(ctx, fontdesc, pdf_cmap_wmode(ctx, fontdesc->encoding));
if (kind == TRUETYPE ||
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1565 */
!strcmp(pdf_to_name(pdf_dict_gets(dict, "Subtype")), "CIDFontType2") ||
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1997 */
pdf_is_indirect(pdf_dict_gets(dict, "CIDToGIDMap")))
{
pdf_obj *cidtogidmap;
cidtogidmap = pdf_dict_gets(dict, "CIDToGIDMap");
if (pdf_is_indirect(cidtogidmap))
{
fz_buffer *buf;
buf = pdf_load_stream(doc, pdf_to_num(cidtogidmap), pdf_to_gen(cidtogidmap));
fontdesc->cid_to_gid_len = (buf->len) / 2;
fontdesc->cid_to_gid = fz_malloc_array(ctx, fontdesc->cid_to_gid_len, sizeof(unsigned short));
fontdesc->size += fontdesc->cid_to_gid_len * sizeof(unsigned short);
for (i = 0; i < fontdesc->cid_to_gid_len; i++)
fontdesc->cid_to_gid[i] = (buf->data[i * 2] << 8) + buf->data[i * 2 + 1];
fz_drop_buffer(ctx, buf);
}
/* if truetype font is external, cidtogidmap should not be identity */
/* so we map from cid to unicode and then map that through the (3 1) */
/* unicode cmap to get a glyph id */
else if (fontdesc->font->ft_substitute)
{
fterr = FT_Select_Charmap(face, ft_encoding_unicode);
if (fterr)
{
fz_throw(ctx, FZ_ERROR_GENERIC, "fonterror: no unicode cmap when emulating CID font: %s", ft_error_string(fterr));
}
if (!strcmp(collection, "Adobe-CNS1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-CNS1-UCS2");
else if (!strcmp(collection, "Adobe-GB1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-GB1-UCS2");
else if (!strcmp(collection, "Adobe-Japan1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan1-UCS2");
else if (!strcmp(collection, "Adobe-Japan2"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Japan2-UCS2");
else if (!strcmp(collection, "Adobe-Korea1"))
fontdesc->to_ttf_cmap = pdf_load_system_cmap(ctx, "Adobe-Korea1-UCS2");
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=2318 */
else if (!strcmp(collection, "Adobe-Identity") && fontdesc->font->ft_file)
fontdesc->font->ft_substitute = 0;
}
}
/* cf. http://code.google.com/p/sumatrapdf/issues/detail?id=1961 */
fz_try(ctx)
{
pdf_load_to_unicode(doc, fontdesc, NULL, collection, to_unicode);
}
fz_catch(ctx)
{
fz_warn(ctx, "cannot load ToUnicode CMap");
}
/* If we have an identity encoding, we're supposed to use the glyph ids directly.
* If we only have a substitute font, that won't work.
* Make a last ditch attempt by using
* the ToUnicode table if it exists to map via the substitute font's cmap. */
if (strstr(fontdesc->encoding->cmap_name, "Identity-") && fontdesc->font->ft_substitute)
{
fz_warn(ctx, "non-embedded font using identity encoding: %s", basefont);
if (fontdesc->to_unicode && !fontdesc->to_ttf_cmap)
fontdesc->to_ttf_cmap = pdf_keep_cmap(ctx, fontdesc->to_unicode);
}
/* Horizontal */
dw = 1000;
obj = pdf_dict_gets(dict, "DW");
if (obj)
dw = pdf_to_int(obj);
pdf_set_default_hmtx(ctx, fontdesc, dw);
widths = pdf_dict_gets(dict, "W");
if (widths)
{
int c0, c1, w, n, m;
n = pdf_array_len(widths);
for (i = 0; i < n; )
{
c0 = pdf_to_int(pdf_array_get(widths, i));
obj = pdf_array_get(widths, i + 1);
if (pdf_is_array(obj))
{
m = pdf_array_len(obj);
for (k = 0; k < m; k++)
{
w = pdf_to_int(pdf_array_get(obj, k));
pdf_add_hmtx(ctx, fontdesc, c0 + k, c0 + k, w);
}
i += 2;
}
else
{
c1 = pdf_to_int(obj);
w = pdf_to_int(pdf_array_get(widths, i + 2));
pdf_add_hmtx(ctx, fontdesc, c0, c1, w);
i += 3;
}
}
}
pdf_end_hmtx(ctx, fontdesc);
/* Vertical */
if (pdf_cmap_wmode(ctx, fontdesc->encoding) == 1)
{
int dw2y = 880;
int dw2w = -1000;
obj = pdf_dict_gets(dict, "DW2");
if (obj)
{
dw2y = pdf_to_int(pdf_array_get(obj, 0));
dw2w = pdf_to_int(pdf_array_get(obj, 1));
}
pdf_set_default_vmtx(ctx, fontdesc, dw2y, dw2w);
widths = pdf_dict_gets(dict, "W2");
if (widths)
{
int c0, c1, w, x, y, n;
n = pdf_array_len(widths);
for (i = 0; i < n; )
{
c0 = pdf_to_int(pdf_array_get(widths, i));
obj = pdf_array_get(widths, i + 1);
if (pdf_is_array(obj))
{
int m = pdf_array_len(obj);
for (k = 0; k * 3 < m; k ++)
{
w = pdf_to_int(pdf_array_get(obj, k * 3 + 0));
x = pdf_to_int(pdf_array_get(obj, k * 3 + 1));
y = pdf_to_int(pdf_array_get(obj, k * 3 + 2));
pdf_add_vmtx(ctx, fontdesc, c0 + k, c0 + k, x, y, w);
}
i += 2;
}
else
{
c1 = pdf_to_int(obj);
w = pdf_to_int(pdf_array_get(widths, i + 2));
x = pdf_to_int(pdf_array_get(widths, i + 3));
y = pdf_to_int(pdf_array_get(widths, i + 4));
pdf_add_vmtx(ctx, fontdesc, c0, c1, x, y, w);
i += 5;
}
}
}
pdf_end_vmtx(ctx, fontdesc);
}
}
fz_catch(ctx)
{
pdf_drop_font(ctx, fontdesc);
fz_rethrow_message(ctx, "cannot load cid font (%d %d R)", pdf_to_num(dict), pdf_to_gen(dict));
}
return fontdesc;
}
