static void renumberobjs(void)
{
pdf_xref_entry *oldxref;
int newlen;
int num;
/* Apply renumber map to indirect references in all objects in xref */
renumberobj(xref->trailer);
for (num = 0; num < xref->len; num++)
{
fz_obj *obj = xref->table[num].obj;
if (fz_is_indirect(obj))
{
obj = fz_new_indirect(ctx, renumbermap[fz_to_num(obj)], 0, xref);
pdf_update_object(xref, num, 0, obj);
fz_drop_obj(obj);
}
else
{
renumberobj(obj);
}
}
/* Create new table for the reordered, compacted xref */
oldxref = xref->table;
xref->table = fz_malloc_array(xref->ctx, xref->len, sizeof(pdf_xref_entry));
xref->table[0] = oldxref[0];
/* Move used objects into the new compacted xref */
newlen = 0;
for (num = 1; num < xref->len; num++)
{
if (uselist[num])
{
if (newlen < renumbermap[num])
newlen = renumbermap[num];
xref->table[renumbermap[num]] = oldxref[num];
}
else
{
if (oldxref[num].obj)
fz_drop_obj(oldxref[num].obj);
}
}
fz_free(xref->ctx, oldxref);
/* Update the used objects count in compacted xref */
xref->len = newlen + 1;
/* Update list of used objects to fit with compacted xref */
for (num = 1; num < xref->len; num++)
uselist[num] = 1;
}
/*
* Load uncompressed contents of a stream into buf.
*/
fz_error
pdf_load_stream(fz_buffer **bufp, pdf_xref *xref, int num, int gen)
{
fz_error error;
fz_stream *stm;
fz_obj *dict, *obj;
int i, len;
error = pdf_open_stream(&stm, xref, num, gen);
if (error)
return fz_rethrow(error, "cannot open stream (%d %d R)", num, gen);
error = pdf_load_object(&dict, xref, num, gen);
if (error)
return fz_rethrow(error, "cannot load stream dictionary (%d %d R)", num, gen);
len = fz_to_int(fz_dict_gets(dict, "Length"));
obj = fz_dict_gets(dict, "Filter");
len = pdf_guess_filter_length(len, fz_to_name(obj));
for (i = 0; i < fz_array_len(obj); i++)
len = pdf_guess_filter_length(len, fz_to_name(fz_array_get(obj, i)));
fz_drop_obj(dict);
error = fz_read_all(bufp, stm, len);
if (error)
{
fz_close(stm);
return fz_rethrow(error, "cannot read raw stream (%d %d R)", num, gen);
}
fz_close(stm);
return fz_okay;
}
static void expandstream(fz_obj *obj, int num, int gen)
{
fz_buffer *buf, *tmp;
fz_obj *newlen;
buf = pdf_load_stream(xref, num, gen);
fz_dict_dels(obj, "Filter");
fz_dict_dels(obj, "DecodeParms");
if (doascii && isbinarystream(buf))
{
tmp = hexbuf(buf->data, buf->len);
fz_drop_buffer(ctx, buf);
buf = tmp;
addhexfilter(obj);
}
newlen = fz_new_int(ctx, buf->len);
fz_dict_puts(obj, "Length", newlen);
fz_drop_obj(newlen);
fprintf(out, "%d %d obj\n", num, gen);
fz_fprint_obj(out, obj, doexpand == 0);
fprintf(out, "stream\n");
fwrite(buf->data, 1, buf->len, out);
fprintf(out, "endstream\nendobj\n\n");
fz_drop_buffer(ctx, buf);
}
/*
* Load raw (compressed but decrypted) contents of a stream into buf.
*/
fz_error
pdf_load_raw_stream(fz_buffer **bufp, pdf_xref *xref, int num, int gen)
{
fz_error error;
fz_stream *stm;
fz_obj *dict;
int len;
error = pdf_load_object(&dict, xref, num, gen);
if (error)
return fz_rethrow(error, "cannot load stream dictionary (%d %d R)", num, gen);
len = fz_to_int(fz_dict_gets(dict, "Length"));
fz_drop_obj(dict);
error = pdf_open_raw_stream(&stm, xref, num, gen);
if (error)
return fz_rethrow(error, "cannot open raw stream (%d %d R)", num, gen);
error = fz_read_all(bufp, stm, len);
if (error)
{
fz_close(stm);
return fz_rethrow(error, "cannot read raw stream (%d %d R)", num, gen);
}
fz_close(stm);
return fz_okay;
}
static void copystream(fz_obj *obj, int num, int gen)
{
fz_error error;
fz_buffer *buf, *tmp;
fz_obj *newlen;
error = pdf_load_raw_stream(&buf, xref, num, gen);
if (error)
die(error);
if (doascii && isbinarystream(buf))
{
tmp = hexbuf(buf->data, buf->len);
fz_drop_buffer(ctx, buf);
buf = tmp;
addhexfilter(obj);
newlen = fz_new_int(ctx, buf->len);
fz_dict_puts(ctx, obj, "Length", newlen);
fz_drop_obj(ctx, newlen);
}
fprintf(out, "%d %d obj\n", num, gen);
fz_fprint_obj(ctx, out, obj, !doexpand);
fprintf(out, "stream\n");
fwrite(buf->data, 1, buf->len, out);
fprintf(out, "endstream\nendobj\n\n");
fz_drop_buffer(ctx, buf);
}
void
pdf_free_xref(pdf_xref *xref)
{
int i;
if (xref->store)
pdf_free_store(xref->store);
if (xref->table)
{
for (i = 0; i < xref->len; i++)
{
if (xref->table[i].obj)
{
fz_drop_obj(xref->table[i].obj);
xref->table[i].obj = NULL;
}
}
fz_free(xref->table);
}
if (xref->page_objs)
{
for (i = 0; i < xref->page_len; i++)
fz_drop_obj(xref->page_objs[i]);
fz_free(xref->page_objs);
}
if (xref->page_refs)
{
for (i = 0; i < xref->page_len; i++)
fz_drop_obj(xref->page_refs[i]);
fz_free(xref->page_refs);
}
if (xref->file)
fz_close(xref->file);
if (xref->trailer)
fz_drop_obj(xref->trailer);
if (xref->crypt)
pdf_free_crypt(xref->crypt);
pdf_free_ocg(xref->ocg);
fz_free(xref);
}
static void addhexfilter(fz_obj *dict)
{
fz_obj *f, *dp, *newf, *newdp;
fz_obj *ahx, *nullobj;
ahx = fz_new_name(ctx, "ASCIIHexDecode");
nullobj = fz_new_null(ctx);
newf = newdp = NULL;
f = fz_dict_gets(dict, "Filter");
dp = fz_dict_gets(dict, "DecodeParms");
if (fz_is_name(f))
{
newf = fz_new_array(ctx, 2);
fz_array_push(newf, ahx);
fz_array_push(newf, f);
f = newf;
if (fz_is_dict(dp))
{
newdp = fz_new_array(ctx, 2);
fz_array_push(newdp, nullobj);
fz_array_push(newdp, dp);
dp = newdp;
}
}
else if (fz_is_array(f))
{
fz_array_insert(f, ahx);
if (fz_is_array(dp))
fz_array_insert(dp, nullobj);
}
else
f = ahx;
fz_dict_puts(dict, "Filter", f);
if (dp)
fz_dict_puts(dict, "DecodeParms", dp);
fz_drop_obj(ahx);
fz_drop_obj(nullobj);
if (newf)
fz_drop_obj(newf);
if (newdp)
fz_drop_obj(newdp);
}
static void writexref(void)
{
fz_obj *trailer;
fz_obj *obj;
int startxref;
int num;
startxref = ftell(out);
fprintf(out, "xref\n0 %d\n", xref->len);
for (num = 0; num < xref->len; num++)
{
if (uselist[num])
fprintf(out, "%010d %05d n \n", ofslist[num], genlist[num]);
else
fprintf(out, "%010d %05d f \n", ofslist[num], genlist[num]);
}
fprintf(out, "\n");
trailer = fz_new_dict(ctx, 5);
obj = fz_new_int(ctx, xref->len);
fz_dict_puts(trailer, "Size", obj);
fz_drop_obj(obj);
obj = fz_dict_gets(xref->trailer, "Info");
if (obj)
fz_dict_puts(trailer, "Info", obj);
obj = fz_dict_gets(xref->trailer, "Root");
if (obj)
fz_dict_puts(trailer, "Root", obj);
obj = fz_dict_gets(xref->trailer, "ID");
if (obj)
fz_dict_puts(trailer, "ID", obj);
fprintf(out, "trailer\n");
fz_fprint_obj(out, trailer, doexpand == 0);
fprintf(out, "\n");
fz_drop_obj(trailer);
fprintf(out, "startxref\n%d\n%%%%EOF\n", startxref);
}
void
pdf_free_link(fz_context *ctx, pdf_link *link)
{
if (link->next)
pdf_free_link(ctx, link->next);
if (link->dest)
fz_drop_obj(ctx, link->dest);
fz_free(ctx, link);
}
static void renumberobj(fz_obj *obj)
{
int i;
fz_context *ctx = xref->ctx;
if (fz_is_dict(obj))
{
int n = fz_dict_len(obj);
for (i = 0; i < n; i++)
{
fz_obj *key = fz_dict_get_key(obj, i);
fz_obj *val = fz_dict_get_val(obj, i);
if (fz_is_indirect(val))
{
val = fz_new_indirect(ctx, renumbermap[fz_to_num(val)], 0, xref);
fz_dict_put(obj, key, val);
fz_drop_obj(val);
}
else
{
renumberobj(val);
}
}
}
else if (fz_is_array(obj))
{
int n = fz_array_len(obj);
for (i = 0; i < n; i++)
{
fz_obj *val = fz_array_get(obj, i);
if (fz_is_indirect(val))
{
val = fz_new_indirect(ctx, renumbermap[fz_to_num(val)], 0, xref);
fz_array_put(obj, i, val);
fz_drop_obj(val);
}
else
{
renumberobj(val);
}
}
}
}
static void writeobject(int num, int gen)
{
fz_error error;
fz_obj *obj;
fz_obj *type;
error = pdf_load_object(&obj, xref, num, gen);
if (error)
die(error);
/* skip ObjStm and XRef objects */
if (fz_is_dict(ctx, obj))
{
type = fz_dict_gets(ctx, obj, "Type");
if (fz_is_name(ctx, type) && !strcmp(fz_to_name(ctx, type), "ObjStm"))
{
uselist[num] = 0;
fz_drop_obj(ctx, obj);
return;
}
if (fz_is_name(ctx, type) && !strcmp(fz_to_name(ctx, type), "XRef"))
{
uselist[num] = 0;
fz_drop_obj(ctx, obj);
return;
}
}
if (!pdf_is_stream(xref, num, gen))
{
fprintf(out, "%d %d obj\n", num, gen);
fz_fprint_obj(ctx, out, obj, !doexpand);
fprintf(out, "endobj\n\n");
}
else
{
if (doexpand && !pdf_is_jpx_image(ctx, obj))
expandstream(obj, num, gen);
else
copystream(obj, num, gen);
}
fz_drop_obj(ctx, obj);
}
static void
pdf_free_ocg(pdf_ocg_descriptor *desc)
{
if (desc == NULL)
return;
if (desc->intent)
fz_drop_obj(desc->intent);
fz_free(desc->ocgs);
fz_free(desc);
}
static fz_obj *
pdf_clone_for_view_only(pdf_xref *xref, fz_obj *obj)
{
fz_obj *ocgs = pdf_dict_from_string(xref, ANNOT_OC_VIEW_ONLY);
obj = fz_copy_dict(xref->ctx, pdf_resolve_indirect(obj));
fz_dict_puts(xref->ctx, obj, "OC", ocgs);
fz_drop_obj(xref->ctx, ocgs);
return obj;
}
void
pdf_free_annot(fz_context *ctx, pdf_annot *annot)
{
if (annot->next)
pdf_free_annot(ctx, annot->next);
if (annot->ap)
pdf_drop_xobject(ctx, annot->ap);
if (annot->obj)
fz_drop_obj(ctx, annot->obj);
fz_free(ctx, annot);
}
void
pdf_drop_pattern(pdf_pattern *pat)
{
if (pat && --pat->refs == 0)
{
if (pat->resources)
fz_drop_obj(pat->resources);
if (pat->contents)
fz_drop_buffer(pat->contents);
fz_free(pat);
}
}
static fz_error
pdf_read_xref_sections(pdf_xref *xref, int ofs, char *buf, int cap)
{
fz_error error;
fz_obj *trailer;
fz_obj *prev;
fz_obj *xrefstm;
error = pdf_read_xref(&trailer, xref, ofs, buf, cap);
if (error)
return fz_rethrow(error, "cannot read xref section");
/* FIXME: do we overwrite free entries properly? */
xrefstm = fz_dict_gets(trailer, "XRefStm");
if (xrefstm)
{
error = pdf_read_xref_sections(xref, fz_to_int(xrefstm), buf, cap);
if (error)
{
fz_drop_obj(trailer);
return fz_rethrow(error, "cannot read /XRefStm xref section");
}
}
prev = fz_dict_gets(trailer, "Prev");
if (prev)
{
error = pdf_read_xref_sections(xref, fz_to_int(prev), buf, cap);
if (error)
{
fz_drop_obj(trailer);
return fz_rethrow(error, "cannot read /Prev xref section");
}
}
fz_drop_obj(trailer);
return fz_okay;
}
static void preloadobjstms(void)
{
fz_obj *obj;
int num;
for (num = 0; num < xref->len; num++)
{
if (xref->table[num].type == 'o')
{
obj = pdf_load_object(xref, num, 0);
fz_drop_obj(obj);
}
}
}
static void preloadobjstms(void)
{
fz_error error;
fz_obj *obj;
int num;
for (num = 0; num < xref->len; num++)
{
if (xref->table[num].type == 'o')
{
error = pdf_load_object(&obj, xref, num, 0);
if (error)
die(error);
fz_drop_obj(ctx, obj);
}
}
}
static void showobject(int num)
{
fz_error error;
fz_obj *obj;
if (!xref)
die(fz_error_make(ctx, "no file specified"));
error = pdf_load_object(&obj, xref, num, 0);
if (error)
die(error);
if (isimage(obj))
saveimage(num);
else if (isfontdesc(obj))
savefont(obj, num);
fz_drop_obj(ctx, obj);
}
void
fz_array_put(fz_obj *obj, int i, fz_obj *item)
{
RESOLVE(obj);
if (!obj)
return; /* Can't warn :( */
if (obj->kind != FZ_ARRAY)
fz_warn(obj->ctx, "assert: not an array (%s)", fz_objkindstr(obj));
else if (i < 0)
fz_warn(obj->ctx, "assert: index %d < 0", i);
else if (i >= obj->u.a.len)
fz_warn(obj->ctx, "assert: index %d > length %d", i, obj->u.a.len);
else
{
if (obj->u.a.items[i])
fz_drop_obj(obj->u.a.items[i]);
obj->u.a.items[i] = fz_keep_obj(item);
}
}
/* Replace numbered object -- for use by pdfclean and similar tools */
void
pdf_update_object(pdf_xref *xref, int num, int gen, fz_obj *newobj)
{
pdf_xref_entry *x;
if (num < 0 || num >= xref->len)
{
fz_warn("object out of range (%d %d R); xref size %d", num, gen, xref->len);
return;
}
x = &xref->table[num];
if (x->obj)
fz_drop_obj(x->obj);
x->obj = fz_keep_obj(newobj);
x->type = 'n';
x->ofs = 0;
}
static void
xps_hacky_get_prop(fz_context *ctx, char *data, fz_obj *dict, char *name, char *tag_name)
{
char *start, *end;
fz_obj *value;
start = strstr(data, tag_name);
if (!start || start == data || start[-1] != '<')
return;
end = strstr(start + 1, tag_name);
start = strchr(start, '>');
if (!start || !end || start >= end || end[-2] != '<' || end[-1] != '/')
return;
for (start++; iswhite(*start); start++);
for (end -= 3; iswhite(*end) && end > start; end--);
value = fz_new_string(ctx, start, end - start + 1);
fz_dict_puts(ctx, dict, name, value);
fz_drop_obj(ctx, value);
}
void
fz_drop_font(fz_context *ctx, fz_font *font)
{
int fterr;
int i, drop;
fz_lock(ctx, FZ_LOCK_ALLOC);
drop = (font && --font->refs == 0);
fz_unlock(ctx, FZ_LOCK_ALLOC);
if (!drop)
return;
if (font->t3procs)
{
if (font->t3resources)
fz_drop_obj(font->t3resources);
for (i = 0; i < 256; i++)
if (font->t3procs[i])
fz_drop_buffer(ctx, font->t3procs[i]);
fz_free(ctx, font->t3procs);
fz_free(ctx, font->t3widths);
fz_free(ctx, font->t3flags);
}
if (font->ft_face)
{
fz_lock(ctx, FZ_LOCK_FREETYPE);
fterr = FT_Done_Face((FT_Face)font->ft_face);
fz_unlock(ctx, FZ_LOCK_FREETYPE);
if (fterr)
fz_warn(ctx, "freetype finalizing face: %s", ft_error_string(fterr));
fz_drop_freetype(ctx);
}
fz_free(ctx, font->ft_file);
fz_free(ctx, font->ft_data);
fz_free(ctx, font->bbox_table);
fz_free(ctx, font->width_table);
fz_free(ctx, font);
}
static void saveimage(int num)
{
fz_error error;
fz_pixmap *img;
fz_obj *ref;
char name[1024];
ref = fz_new_indirect(ctx, num, 0, xref);
/* TODO: detect DCTD and save as jpeg */
error = pdf_load_image(&img, xref, ref);
if (error)
die(error);
if (dorgb && img->colorspace && img->colorspace != fz_device_rgb)
{
fz_pixmap *temp;
temp = fz_new_pixmap_with_rect(ctx, fz_device_rgb, fz_bound_pixmap(img));
fz_convert_pixmap(ctx, img, temp);
fz_drop_pixmap(ctx, img);
img = temp;
}
if (img->n <= 4)
{
sprintf(name, "img-%04d.png", num);
printf("extracting image %s\n", name);
fz_write_png(ctx, img, name, 0);
}
else
{
sprintf(name, "img-%04d.pam", num);
printf("extracting image %s\n", name);
fz_write_pam(ctx, img, name, 0);
}
fz_drop_pixmap(ctx, img);
fz_drop_obj(ctx, ref);
}
static int
xps_decode_tiff_fax(struct tiff *tiff, int comp, fz_stream *chain, byte *wp, int wlen)
{
fz_stream *stm;
fz_obj *params;
fz_obj *columns, *rows, *black_is_1, *k, *encoded_byte_align;
int n;
fz_context *ctx = tiff->ctx;
columns = fz_new_int(ctx, tiff->imagewidth);
rows = fz_new_int(ctx, tiff->imagelength);
black_is_1 = fz_new_bool(ctx, tiff->photometric == 0);
k = fz_new_int(ctx, comp == 4 ? -1 : 0);
encoded_byte_align = fz_new_bool(ctx, comp == 2);
params = fz_new_dict(ctx, 5);
fz_dict_puts(ctx, params, "Columns", columns);
fz_dict_puts(ctx, params, "Rows", rows);
fz_dict_puts(ctx, params, "BlackIs1", black_is_1);
fz_dict_puts(ctx, params, "K", k);
fz_dict_puts(ctx, params, "EncodedByteAlign", encoded_byte_align);
fz_drop_obj(ctx, columns);
fz_drop_obj(ctx, rows);
fz_drop_obj(ctx, black_is_1);
fz_drop_obj(ctx, k);
fz_drop_obj(ctx, encoded_byte_align);
stm = fz_open_faxd(chain, params);
n = fz_read(stm, wp, wlen);
fz_close(stm);
fz_drop_obj(ctx, params);
if (n < 0)
return fz_error_note(ctx, n, "cannot read fax strip");
return fz_okay;
}
static void writeobject(int num, int gen)
{
fz_obj *obj;
fz_obj *type;
obj = pdf_load_object(xref, num, gen);
/* skip ObjStm and XRef objects */
if (fz_is_dict(obj))
{
type = fz_dict_gets(obj, "Type");
if (fz_is_name(type) && !strcmp(fz_to_name(type), "ObjStm"))
{
uselist[num] = 0;
fz_drop_obj(obj);
return;
}
if (fz_is_name(type) && !strcmp(fz_to_name(type), "XRef"))
{
uselist[num] = 0;
fz_drop_obj(obj);
return;
}
}
if (!pdf_is_stream(xref, num, gen))
{
fprintf(out, "%d %d obj\n", num, gen);
fz_fprint_obj(out, obj, doexpand == 0);
fprintf(out, "endobj\n\n");
}
else
{
int dontexpand = 0;
if (doexpand != 0 && doexpand != expand_all)
{
fz_obj *o;
if ((o = fz_dict_gets(obj, "Type"), !strcmp(fz_to_name(o), "XObject")) &&
(o = fz_dict_gets(obj, "Subtype"), !strcmp(fz_to_name(o), "Image")))
dontexpand = !(doexpand & expand_images);
if (o = fz_dict_gets(obj, "Type"), !strcmp(fz_to_name(o), "Font"))
dontexpand = !(doexpand & expand_fonts);
if (o = fz_dict_gets(obj, "Type"), !strcmp(fz_to_name(o), "FontDescriptor"))
dontexpand = !(doexpand & expand_fonts);
if ((o = fz_dict_gets(obj, "Length1")) != NULL)
dontexpand = !(doexpand & expand_fonts);
if ((o = fz_dict_gets(obj, "Length2")) != NULL)
dontexpand = !(doexpand & expand_fonts);
if ((o = fz_dict_gets(obj, "Length3")) != NULL)
dontexpand = !(doexpand & expand_fonts);
if (o = fz_dict_gets(obj, "Subtype"), !strcmp(fz_to_name(o), "Type1C"))
dontexpand = !(doexpand & expand_fonts);
if (o = fz_dict_gets(obj, "Subtype"), !strcmp(fz_to_name(o), "CIDFontType0C"))
dontexpand = !(doexpand & expand_fonts);
}
if (doexpand && !dontexpand && !pdf_is_jpx_image(ctx, obj))
expandstream(obj, num, gen);
else
copystream(obj, num, gen);
}
fz_drop_obj(obj);
}
static void retainpages(int argc, char **argv)
{
fz_obj *oldroot, *root, *pages, *kids, *countobj, *parent, *olddests;
/* Keep only pages/type and (reduced) dest entries to avoid
* references to unretained pages */
oldroot = fz_dict_gets(xref->trailer, "Root");
pages = fz_dict_gets(oldroot, "Pages");
olddests = pdf_load_name_tree(xref, "Dests");
root = fz_new_dict(ctx, 2);
fz_dict_puts(root, "Type", fz_dict_gets(oldroot, "Type"));
fz_dict_puts(root, "Pages", fz_dict_gets(oldroot, "Pages"));
pdf_update_object(xref, fz_to_num(oldroot), fz_to_gen(oldroot), root);
fz_drop_obj(root);
/* Create a new kids array with only the pages we want to keep */
parent = fz_new_indirect(ctx, fz_to_num(pages), fz_to_gen(pages), xref);
kids = fz_new_array(ctx, 1);
/* Retain pages specified */
while (argc - fz_optind)
{
int page, spage, epage;
char *spec, *dash;
char *pagelist = argv[fz_optind];
spec = fz_strsep(&pagelist, ",");
while (spec)
{
dash = strchr(spec, '-');
if (dash == spec)
spage = epage = pdf_count_pages(xref);
else
spage = epage = atoi(spec);
if (dash)
{
if (strlen(dash) > 1)
epage = atoi(dash + 1);
else
epage = pdf_count_pages(xref);
}
if (spage > epage)
page = spage, spage = epage, epage = page;
if (spage < 1)
spage = 1;
if (epage > pdf_count_pages(xref))
epage = pdf_count_pages(xref);
for (page = spage; page <= epage; page++)
{
fz_obj *pageobj = xref->page_objs[page-1];
fz_obj *pageref = xref->page_refs[page-1];
fz_dict_puts(pageobj, "Parent", parent);
/* Store page object in new kids array */
fz_array_push(kids, pageref);
}
spec = fz_strsep(&pagelist, ",");
}
fz_optind++;
}
fz_drop_obj(parent);
/* Update page count and kids array */
countobj = fz_new_int(ctx, fz_array_len(kids));
fz_dict_puts(pages, "Count", countobj);
fz_drop_obj(countobj);
fz_dict_puts(pages, "Kids", kids);
fz_drop_obj(kids);
/* Also preserve the (partial) Dests name tree */
if (olddests)
{
int i;
fz_obj *names = fz_new_dict(ctx, 1);
fz_obj *dests = fz_new_dict(ctx, 1);
fz_obj *names_list = fz_new_array(ctx, 32);
for (i = 0; i < fz_dict_len(olddests); i++)
{
fz_obj *key = fz_dict_get_key(olddests, i);
fz_obj *val = fz_dict_get_val(olddests, i);
fz_obj *key_str = fz_new_string(ctx, fz_to_name(key), strlen(fz_to_name(key)));
fz_obj *dest = fz_dict_gets(val, "D");
dest = fz_array_get(dest ? dest : val, 0);
if (fz_array_contains(fz_dict_gets(pages, "Kids"), dest))
{
fz_array_push(names_list, key_str);
fz_array_push(names_list, val);
}
fz_drop_obj(key_str);
}
root = fz_dict_gets(xref->trailer, "Root");
fz_dict_puts(dests, "Names", names_list);
fz_dict_puts(names, "Dests", dests);
fz_dict_puts(root, "Names", names);
fz_drop_obj(names);
fz_drop_obj(dests);
fz_drop_obj(names_list);
fz_drop_obj(olddests);
}
}
fz_error
pdf_open_xref_with_stream(pdf_xref **xrefp, fz_stream *file, char *password)
{
pdf_xref *xref;
fz_error error;
fz_obj *encrypt, *id;
fz_obj *dict, *obj;
int i, repaired = 0;
/* install pdf specific callback */
fz_resolve_indirect = pdf_resolve_indirect;
xref = fz_malloc(sizeof(pdf_xref));
memset(xref, 0, sizeof(pdf_xref));
xref->file = fz_keep_stream(file);
error = pdf_load_xref(xref, xref->scratch, sizeof xref->scratch);
if (error)
{
fz_catch(error, "trying to repair");
if (xref->table)
{
fz_free(xref->table);
xref->table = NULL;
xref->len = 0;
}
if (xref->trailer)
{
fz_drop_obj(xref->trailer);
xref->trailer = NULL;
}
error = pdf_repair_xref(xref, xref->scratch, sizeof xref->scratch);
if (error)
{
pdf_free_xref(xref);
return fz_rethrow(error, "cannot repair document");
}
repaired = 1;
}
encrypt = fz_dict_gets(xref->trailer, "Encrypt");
id = fz_dict_gets(xref->trailer, "ID");
if (fz_is_dict(encrypt))
{
error = pdf_new_crypt(&xref->crypt, encrypt, id);
if (error)
{
pdf_free_xref(xref);
return fz_rethrow(error, "cannot decrypt document");
}
}
if (pdf_needs_password(xref))
{
/* Only care if we have a password */
if (password)
{
int okay = pdf_authenticate_password(xref, password);
if (!okay)
{
pdf_free_xref(xref);
return fz_throw("invalid password");
}
}
}
if (repaired)
{
int hasroot, hasinfo;
error = pdf_repair_obj_stms(xref);
if (error)
{
pdf_free_xref(xref);
return fz_rethrow(error, "cannot repair document");
}
hasroot = fz_dict_gets(xref->trailer, "Root") != NULL;
hasinfo = fz_dict_gets(xref->trailer, "Info") != NULL;
for (i = 1; i < xref->len; i++)
{
if (xref->table[i].type == 0 || xref->table[i].type == 'f')
continue;
error = pdf_load_object(&dict, xref, i, 0);
if (error)
{
fz_catch(error, "ignoring broken object (%d 0 R)", i);
continue;
}
if (!hasroot)
{
obj = fz_dict_gets(dict, "Type");
if (fz_is_name(obj) && !strcmp(fz_to_name(obj), "Catalog"))
{
obj = fz_new_indirect(i, 0, xref);
fz_dict_puts(xref->trailer, "Root", obj);
fz_drop_obj(obj);
}
}
if (!hasinfo)
{
if (fz_dict_gets(dict, "Creator") || fz_dict_gets(dict, "Producer"))
{
obj = fz_new_indirect(i, 0, xref);
fz_dict_puts(xref->trailer, "Info", obj);
fz_drop_obj(obj);
}
}
fz_drop_obj(dict);
}
}
error = pdf_read_ocg(xref);
if (error)
{
pdf_free_xref(xref);
return fz_rethrow(error, "Broken Optional Content");
}
*xrefp = xref;
return fz_okay;
}
static fz_error
pdf_load_obj_stm(pdf_xref *xref, int num, int gen, char *buf, int cap)
{
fz_error error;
fz_stream *stm;
fz_obj *objstm;
int *numbuf;
int *ofsbuf;
fz_obj *obj;
int first;
int count;
int i, n;
int tok;
error = pdf_load_object(&objstm, xref, num, gen);
if (error)
return fz_rethrow(error, "cannot load object stream object (%d %d R)", num, gen);
count = fz_to_int(fz_dict_gets(objstm, "N"));
first = fz_to_int(fz_dict_gets(objstm, "First"));
numbuf = fz_calloc(count, sizeof(int));
ofsbuf = fz_calloc(count, sizeof(int));
error = pdf_open_stream(&stm, xref, num, gen);
if (error)
{
error = fz_rethrow(error, "cannot open object stream (%d %d R)", num, gen);
goto cleanupbuf;
}
for (i = 0; i < count; i++)
{
error = pdf_lex(&tok, stm, buf, cap, &n);
if (error || tok != PDF_TOK_INT)
{
error = fz_rethrow(error, "corrupt object stream (%d %d R)", num, gen);
goto cleanupstm;
}
numbuf[i] = atoi(buf);
error = pdf_lex(&tok, stm, buf, cap, &n);
if (error || tok != PDF_TOK_INT)
{
error = fz_rethrow(error, "corrupt object stream (%d %d R)", num, gen);
goto cleanupstm;
}
ofsbuf[i] = atoi(buf);
}
fz_seek(stm, first, 0);
for (i = 0; i < count; i++)
{
fz_seek(stm, first + ofsbuf[i], 0);
error = pdf_parse_stm_obj(&obj, xref, stm, buf, cap);
if (error)
{
error = fz_rethrow(error, "cannot parse object %d in stream (%d %d R)", i, num, gen);
goto cleanupstm;
}
if (numbuf[i] < 1 || numbuf[i] >= xref->len)
{
fz_drop_obj(obj);
error = fz_throw("object id (%d 0 R) out of range (0..%d)", numbuf[i], xref->len - 1);
goto cleanupstm;
}
if (xref->table[numbuf[i]].type == 'o' && xref->table[numbuf[i]].ofs == num)
{
if (xref->table[numbuf[i]].obj)
fz_drop_obj(xref->table[numbuf[i]].obj);
xref->table[numbuf[i]].obj = obj;
}
else
{
fz_drop_obj(obj);
}
}
fz_close(stm);
fz_free(ofsbuf);
fz_free(numbuf);
fz_drop_obj(objstm);
return fz_okay;
cleanupstm:
fz_close(stm);
cleanupbuf:
fz_free(ofsbuf);
fz_free(numbuf);
fz_drop_obj(objstm);
return error; /* already rethrown */
}
static fz_outline *
pdf_load_outline_imp(pdf_xref *xref, fz_obj *dict)
{
pdf_link *link;
fz_outline *node;
fz_obj *obj;
/* SumatraPDF: prevent potential stack overflow */
fz_outline *prev, *root = NULL;
fz_obj *origDict = dict;
fz_context *ctx = xref->ctx;
if (fz_is_null(ctx, dict))
return NULL;
/* SumatraPDF: prevent cyclic outlines */
do
{
if (fz_dict_gets(ctx, dict, ".seen"))
break;
obj = fz_new_null(ctx);
fz_dict_puts(ctx, dict, ".seen", obj);
fz_drop_obj(ctx, obj);
node = fz_malloc(ctx, sizeof(fz_outline));
node->title = NULL;
node->page = -1;
node->down = NULL;
node->next = NULL;
obj = fz_dict_gets(ctx, dict, "Title");
if (obj)
node->title = pdf_to_utf8(ctx, obj);
/* SumatraPDF: support expansion states */
node->is_open = fz_to_int(ctx, fz_dict_gets(ctx, dict, "Count")) >= 0;
/* SumatraPDF: extended outline actions */
node->data = node->free_data = NULL;
if (fz_dict_gets(ctx, dict, "Dest") || fz_dict_gets(ctx, dict, "A"))
{
link = pdf_load_link(xref, dict);
if (link) /* SumatraPDF: don't crash if it's no link after all */
{
if (link->kind == PDF_LINK_GOTO)
node->page = pdf_find_page_number(xref, fz_array_get(ctx, link->dest, 0));
/* SumatraPDF: extended outline actions */
node->data = link;
node->free_data = pdf_free_link;
}
}
obj = fz_dict_gets(ctx, dict, "First");
if (obj)
node->down = pdf_load_outline_imp(xref, obj);
/* SumatraPDF: prevent potential stack overflow */
if (!root)
prev = root = node;
else
prev = prev->next = node;
dict = fz_dict_gets(ctx, dict, "Next");
} while (dict && !fz_is_null(ctx, dict));
node = root;
/* SumatraPDF: prevent cyclic outlines */
for (dict = origDict; dict && fz_dict_gets(ctx, dict, ".seen"); dict = fz_dict_gets(ctx, dict, "Next"))
fz_dict_dels(ctx, dict, ".seen");
return node;
}
