static void
pdf_clean_stream_object(fz_context *ctx, pdf_document *doc, pdf_obj *obj, pdf_obj *orig_res, fz_cookie *cookie, int own_res)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *res = NULL;
pdf_obj *ref = NULL;
fz_buffer *buffer;
if (!obj)
return;
fz_var(res);
fz_var(ref);
fz_var(proc_buffer);
fz_var(proc_filter);
buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
if (own_res)
{
pdf_obj *r = pdf_dict_get(ctx, obj, PDF_NAME_Resources);
if (r)
orig_res = r;
}
res = pdf_new_dict(ctx, doc, 1);
proc_buffer = pdf_new_buffer_processor(ctx, buffer);
proc_filter = pdf_new_filter_processor(ctx, proc_buffer, doc, orig_res, res);
pdf_process_contents(ctx, proc_filter, doc, orig_res, obj, cookie);
pdf_update_stream(ctx, doc, obj, buffer, 0);
if (own_res)
{
ref = pdf_new_ref(ctx, doc, res);
pdf_dict_put(ctx, obj, PDF_NAME_Resources, ref);
}
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
pdf_drop_obj(ctx, res);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow_message(ctx, "Failed while cleaning xobject");
}
}
void
pdf_insert_page(pdf_document *doc, pdf_page *page, int at)
{
fz_context *ctx = doc->ctx;
int count = pdf_count_pages(doc);
pdf_obj *parent, *kids;
pdf_obj *page_ref;
int i;
page_ref = pdf_new_ref(doc, page->me);
fz_try(ctx)
{
if (count == 0)
{
/* TODO: create new page tree? */
fz_throw(ctx, FZ_ERROR_GENERIC, "empty page tree, cannot insert page");
}
else if (at >= count)
{
if (at > count)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot insert page beyond end of page tree");
/* append after last page */
pdf_lookup_page_loc(doc, count - 1, &parent, &i);
kids = pdf_dict_gets(parent, "Kids");
pdf_array_insert(kids, page_ref, i + 1);
}
else
{
/* insert before found page */
pdf_lookup_page_loc(doc, at, &parent, &i);
kids = pdf_dict_gets(parent, "Kids");
pdf_array_insert(kids, page_ref, i);
}
pdf_dict_puts(page->me, "Parent", parent);
/* Adjust page counts */
while (parent)
{
int count = pdf_to_int(pdf_dict_gets(parent, "Count"));
pdf_dict_puts_drop(parent, "Count", pdf_new_int(doc, count + 1));
parent = pdf_dict_gets(parent, "Parent");
}
}
fz_always(ctx)
{
pdf_drop_obj(page_ref);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
void
pdf_insert_page(pdf_document *doc, pdf_page *page, int at)
{
fz_context *ctx = doc->ctx;
int count = pdf_count_pages(doc);
pdf_obj *parent, *kids;
pdf_obj *page_ref;
int i;
page_ref = pdf_new_ref(doc, page->me);
fz_try(ctx)
{
if (count == 0)
{
pdf_obj *root = pdf_dict_gets(pdf_trailer(doc), "Root");
parent = pdf_dict_gets(root, "Pages");
if (!parent)
fz_throw(doc->ctx, FZ_ERROR_GENERIC, "cannot find page tree");
kids = pdf_dict_gets(parent, "Kids");
if (!kids)
fz_throw(doc->ctx, FZ_ERROR_GENERIC, "malformed page tree");
pdf_array_insert(kids, page_ref, 0);
}
else if (at >= count)
{
if (at == INT_MAX)
at = count;
if (at > count)
fz_throw(ctx, FZ_ERROR_GENERIC, "cannot insert page beyond end of page tree");
/* append after last page */
pdf_lookup_page_loc(doc, count - 1, &parent, &i);
kids = pdf_dict_gets(parent, "Kids");
pdf_array_insert(kids, page_ref, i + 1);
}
else
{
/* insert before found page */
pdf_lookup_page_loc(doc, at, &parent, &i);
kids = pdf_dict_gets(parent, "Kids");
pdf_array_insert(kids, page_ref, i);
}
pdf_dict_puts(page->me, "Parent", parent);
/* Adjust page counts */
while (parent)
{
int count = pdf_to_int(pdf_dict_gets(parent, "Count"));
pdf_dict_puts_drop(parent, "Count", pdf_new_int(doc, count + 1));
parent = pdf_dict_gets(parent, "Parent");
}
}
fz_always(ctx)
{
pdf_drop_obj(page_ref);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
doc->page_count = 0; /* invalidate cached value */
}
pdf_page *
pdf_create_page(pdf_document *doc, fz_rect mediabox, int res, int rotate)
{
pdf_page *page = NULL;
pdf_obj *pageobj, *obj;
float userunit = 1;
fz_context *ctx = doc->ctx;
fz_matrix ctm, tmp;
fz_rect realbox;
page = fz_malloc_struct(ctx, pdf_page);
obj = NULL;
fz_var(obj);
fz_try(ctx)
{
page->resources = NULL;
page->contents = NULL;
page->transparency = 0;
page->links = NULL;
page->annots = NULL;
page->me = pageobj = pdf_new_dict(doc, 4);
pdf_dict_puts_drop(pageobj, "Type", pdf_new_name(doc, "Page"));
page->mediabox.x0 = fz_min(mediabox.x0, mediabox.x1) * userunit;
page->mediabox.y0 = fz_min(mediabox.y0, mediabox.y1) * userunit;
page->mediabox.x1 = fz_max(mediabox.x0, mediabox.x1) * userunit;
page->mediabox.y1 = fz_max(mediabox.y0, mediabox.y1) * userunit;
pdf_dict_puts_drop(pageobj, "MediaBox", pdf_new_rect(doc, &page->mediabox));
/* Snap page->rotate to 0, 90, 180 or 270 */
if (page->rotate < 0)
page->rotate = 360 - ((-page->rotate) % 360);
if (page->rotate >= 360)
page->rotate = page->rotate % 360;
page->rotate = 90*((page->rotate + 45)/90);
if (page->rotate > 360)
page->rotate = 0;
pdf_dict_puts_drop(pageobj, "Rotate", pdf_new_int(doc, page->rotate));
fz_pre_rotate(fz_scale(&ctm, 1, -1), -page->rotate);
realbox = page->mediabox;
fz_transform_rect(&realbox, &ctm);
fz_pre_scale(fz_translate(&tmp, -realbox.x0, -realbox.y0), userunit, userunit);
fz_concat(&ctm, &ctm, &tmp);
page->ctm = ctm;
obj = pdf_new_dict(doc, 4);
page->contents = pdf_new_ref(doc, obj);
pdf_drop_obj(obj);
obj = NULL;
pdf_dict_puts(pageobj, "Contents", page->contents);
}
fz_catch(ctx)
{
pdf_drop_obj(page->me);
pdf_drop_obj(obj);
fz_free(ctx, page);
fz_rethrow_message(ctx, "Failed to create page");
}
return page;
}
void pdf_set_free_text_details(fz_context *ctx, pdf_document *doc, pdf_annot *annot, fz_point *pos, char *text, char *font_name, float font_size, float color[3])
{
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_get(ctx, annot->page->me, PDF_NAME_Resources);
if (!dr)
{
dr = pdf_new_dict(ctx, doc, 1);
pdf_dict_put_drop(ctx, annot->page->me, 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;
int da_len;
fz_rect bounds;
find_free_font_name(ctx, form_fonts, nbuf, sizeof(nbuf));
font = pdf_new_dict(ctx, doc, 5);
ref = pdf_new_ref(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, &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));
update_rect(ctx, annot);
}
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);
}
}
static void
pdf_clean_type3(fz_context *ctx, pdf_document *doc, pdf_obj *obj, pdf_obj *orig_res, fz_cookie *cookie)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *res = NULL;
pdf_obj *ref = NULL;
pdf_obj *charprocs;
int i, l;
fz_var(res);
fz_var(ref);
fz_var(proc_buffer);
fz_var(proc_filter);
fz_try(ctx)
{
res = pdf_dict_get(ctx, obj, PDF_NAME_Resources);
if (res)
orig_res = res;
res = NULL;
res = pdf_new_dict(ctx, doc, 1);
charprocs = pdf_dict_get(ctx, obj, PDF_NAME_CharProcs);
l = pdf_dict_len(ctx, charprocs);
for (i = 0; i < l; i++)
{
pdf_obj *val = pdf_dict_get_val(ctx, charprocs, i);
fz_buffer *buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
proc_buffer = pdf_new_buffer_processor(ctx, buffer);
proc_filter = pdf_new_filter_processor(ctx, proc_buffer, doc, orig_res, res);
pdf_process_contents(ctx, proc_filter, doc, orig_res, val, cookie);
pdf_update_stream(ctx, doc, val, buffer, 0);
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
}
fz_catch(ctx)
{
fz_rethrow(ctx);
}
}
/* ProcSet - no cleaning possible. Inherit this from the old dict. */
pdf_dict_put(ctx, res, PDF_NAME_ProcSet, pdf_dict_get(ctx, orig_res, PDF_NAME_ProcSet));
ref = pdf_new_ref(ctx, doc, res);
pdf_dict_put(ctx, obj, PDF_NAME_Resources, ref);
}
fz_always(ctx)
{
pdf_drop_obj(ctx, res);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow_message(ctx, "Failed while cleaning xobject");
}
}
void pdf_clean_page_contents(fz_context *ctx, pdf_document *doc, pdf_page *page, fz_cookie *cookie, pdf_page_contents_process_fn *proc_fn, void *proc_arg)
{
pdf_processor *proc_buffer = NULL;
pdf_processor *proc_filter = NULL;
pdf_obj *new_obj = NULL;
pdf_obj *new_ref = NULL;
pdf_obj *res = NULL;
pdf_obj *ref = NULL;
pdf_obj *obj;
pdf_obj *contents;
fz_buffer *buffer;
fz_var(new_obj);
fz_var(new_ref);
fz_var(res);
fz_var(ref);
fz_var(proc_buffer);
fz_var(proc_filter);
buffer = fz_new_buffer(ctx, 1024);
fz_try(ctx)
{
res = pdf_new_dict(ctx, doc, 1);
proc_buffer = pdf_new_buffer_processor(ctx, buffer);
proc_filter = pdf_new_filter_processor(ctx, proc_buffer, doc, page->resources, res);
pdf_process_contents(ctx, proc_filter, doc, page->resources, page->contents, cookie);
contents = page->contents;
if (pdf_is_array(ctx, contents))
{
/* create a new object to replace the array */
new_obj = pdf_new_dict(ctx, doc, 1);
new_ref = pdf_new_ref(ctx, doc, new_obj);
page->contents = contents = new_ref;
}
else
{
pdf_dict_del(ctx, contents, PDF_NAME_Filter);
pdf_dict_del(ctx, contents, PDF_NAME_DecodeParms);
}
/* Now deal with resources. The spec allows for Type3 fonts and form
* XObjects to omit a resource dictionary and look in the parent.
* Avoid that by flattening here as part of the cleaning. This could
* conceivably cause changes in rendering, but we don't care. */
/* ExtGState */
obj = pdf_dict_get(ctx, res, PDF_NAME_ExtGState);
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *o = pdf_dict_get(ctx, pdf_dict_get_val(ctx, obj, i), PDF_NAME_SMask);
if (!o)
continue;
o = pdf_dict_get(ctx, o, PDF_NAME_G);
if (!o)
continue;
/* Transparency group XObject */
pdf_clean_stream_object(ctx, doc, o, page->resources, cookie, 1);
}
}
/* ColorSpace - no cleaning possible */
/* Pattern */
obj = pdf_dict_get(ctx, res, PDF_NAME_Pattern);
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *pat = pdf_dict_get_val(ctx, obj, i);
if (!pat)
continue;
if (pdf_to_int(ctx, pdf_dict_get(ctx, pat, PDF_NAME_PatternType)) == 1)
pdf_clean_stream_object(ctx, doc, pat, page->resources, cookie, 0);
}
}
/* Shading - no cleaning possible */
/* XObject */
obj = pdf_dict_get(ctx, res, PDF_NAME_XObject);
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *xobj = pdf_dict_get_val(ctx, obj, i);
if (!pdf_name_eq(ctx, PDF_NAME_Form, pdf_dict_get(ctx, xobj, PDF_NAME_Subtype)))
continue;
pdf_clean_stream_object(ctx, doc, xobj, page->resources, cookie, 1);
}
}
/* Font */
obj = pdf_dict_get(ctx, res, PDF_NAME_Font);
if (obj)
{
int i, l;
l = pdf_dict_len(ctx, obj);
for (i = 0; i < l; i++)
{
pdf_obj *o = pdf_dict_get_val(ctx, obj, i);
if (pdf_name_eq(ctx, PDF_NAME_Type3, pdf_dict_get(ctx, o, PDF_NAME_Subtype)))
{
pdf_clean_type3(ctx, doc, o, page->resources, cookie);
}
}
}
/* ProcSet - no cleaning possible. Inherit this from the old dict. */
obj = pdf_dict_get(ctx, page->resources, PDF_NAME_ProcSet);
if (obj)
pdf_dict_put(ctx, res, PDF_NAME_ProcSet, obj);
/* Properties - no cleaning possible. */
if (proc_fn)
(*proc_fn)(ctx, buffer, res, proc_arg);
pdf_update_stream(ctx, doc, contents, buffer, 0);
pdf_drop_obj(ctx, page->resources);
ref = pdf_new_ref(ctx, doc, res);
page->resources = pdf_keep_obj(ctx, ref);
pdf_dict_put(ctx, page->me, PDF_NAME_Resources, ref);
}
fz_always(ctx)
{
pdf_drop_processor(ctx, proc_filter);
pdf_drop_processor(ctx, proc_buffer);
fz_drop_buffer(ctx, buffer);
pdf_drop_obj(ctx, new_obj);
pdf_drop_obj(ctx, new_ref);
pdf_drop_obj(ctx, res);
pdf_drop_obj(ctx, ref);
}
fz_catch(ctx)
{
fz_rethrow_message(ctx, "Failed while cleaning page");
}
}
