summaryrefslogtreecommitdiff
path: root/source/pdf/pdf-resources.c
diff options
context:
space:
mode:
Diffstat (limited to 'source/pdf/pdf-resources.c')
-rw-r--r--source/pdf/pdf-resources.c295
1 files changed, 295 insertions, 0 deletions
diff --git a/source/pdf/pdf-resources.c b/source/pdf/pdf-resources.c
new file mode 100644
index 00000000..31812b3c
--- /dev/null
+++ b/source/pdf/pdf-resources.c
@@ -0,0 +1,295 @@
+#include "mupdf/pdf.h"
+
+static void
+res_table_free(fz_context *ctx, pdf_res_table *table)
+{
+ int i, n;
+ pdf_res *res;
+
+ if (table == NULL)
+ return;
+ if (table->hash != NULL)
+ {
+ n = fz_hash_len(ctx, table->hash);
+ for (i = 0; i < n; i++)
+ {
+ void *v = fz_hash_get_val(ctx, table->hash, i);
+ if (v)
+ {
+ res = (pdf_res*)v;
+ pdf_drop_obj(ctx, res->obj);
+ fz_free(ctx, res);
+ }
+ }
+ fz_drop_hash(ctx, table->hash);
+ }
+ fz_free(ctx, table);
+}
+
+static void
+res_image_get_md5(fz_context *ctx, fz_image *image, unsigned char *digest)
+{
+ fz_pixmap *pixmap = NULL;
+ int n, size;
+ fz_buffer *buffer = NULL;
+ fz_md5 state;
+
+ fz_var(pixmap);
+ fz_var(buffer);
+
+ fz_try(ctx)
+ {
+ pixmap = fz_get_pixmap_from_image(ctx, image, 0, 0);
+ n = (pixmap->n == 1 ? 1 : pixmap->n - 1);
+ size = image->w * image->h * n;
+ buffer = fz_new_buffer(ctx, size);
+ buffer->len = size;
+ if (pixmap->n == 1)
+ {
+ memcpy(buffer->data, pixmap->samples, size);
+ }
+ else
+ {
+ /* Need to remove the alpha plane */
+ unsigned char *d = buffer->data;
+ unsigned char *s = pixmap->samples;
+ int mod = n;
+ while (size--)
+ {
+ *d++ = *s++;
+ mod--;
+ if (mod == 0)
+ s++, mod = n;
+ }
+ }
+ fz_md5_init(&state);
+ fz_md5_update(&state, buffer->data, buffer->len);
+ fz_md5_final(&state, digest);
+ }
+ fz_always(ctx)
+ {
+ fz_drop_pixmap(ctx, pixmap);
+ fz_drop_buffer(ctx, buffer);
+ }
+ fz_catch(ctx)
+ {
+ fz_rethrow_message(ctx, "image md5 calculation failed");
+ }
+}
+
+/* Image specific methods */
+static void
+res_image_init(fz_context *ctx, pdf_document *doc, pdf_res_table *table)
+{
+ int len, k;
+ pdf_obj *obj;
+ pdf_obj *type;
+ pdf_res *res = NULL;
+ fz_image *image = NULL;
+ unsigned char digest[16];
+ int num = 0;
+
+ fz_var(obj);
+ fz_var(image);
+ fz_var(res);
+
+ fz_try(ctx)
+ {
+ table->hash = fz_new_hash_table(ctx, 4096, 16, -1);
+ len = pdf_count_objects(ctx, doc);
+ for (k = 1; k < len; k++)
+ {
+ obj = pdf_load_object(ctx, doc, k, 0);
+ type = pdf_dict_get(ctx, obj, PDF_NAME_Subtype);
+ if (pdf_name_eq(ctx, type, PDF_NAME_Image))
+ {
+ image = pdf_load_image(ctx, doc, obj);
+ res_image_get_md5(ctx, image, digest);
+ fz_drop_image(ctx, image);
+ image = NULL;
+
+ /* Don't allow overwrites. Number the resources for pdfwrite */
+ if (fz_hash_find(ctx, table->hash, (void *)digest) == NULL)
+ {
+ res = fz_malloc(ctx, sizeof(pdf_res));
+ res->num = num;
+ res->obj = obj;
+ num = num + 1;
+ fz_hash_insert(ctx, table->hash, (void *)digest, obj);
+ }
+ }
+ else
+ {
+ pdf_drop_obj(ctx, obj);
+ }
+ obj = NULL;
+ }
+ }
+ fz_always(ctx)
+ {
+ table->count = num;
+ fz_drop_image(ctx, image);
+ pdf_drop_obj(ctx, obj);
+ }
+ fz_catch(ctx)
+ {
+ res_table_free(ctx, table);
+ fz_rethrow_message(ctx, "image resources table failed to initialize");
+ }
+}
+
+static void*
+res_image_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table, void *item,
+ void *md5)
+{
+ unsigned char digest[16];
+
+ fz_image *image = (fz_image*)item;
+ fz_hash_table *hash = table->hash;
+ pdf_res *res;
+
+ if (hash == NULL)
+ res_image_init(ctx, doc, doc->resources->image);
+ hash = doc->resources->image->hash;
+
+ /* Create md5 and see if we have the item in our table */
+ res_image_get_md5(ctx, image, digest);
+ res = fz_hash_find(ctx, hash, (void*)digest);
+
+ /* Return the digest value so that we can avoid having to recompute it when
+ * we come back to add the new resource reference */
+ if (res == NULL)
+ memcpy(md5, digest, 16);
+ else
+ pdf_keep_obj(ctx, res->obj);
+ return (void*) res;
+}
+
+/* Font specific methods */
+
+/* We do need to come up with an effective way to see what is already in the
+ * file to avoid adding to what is already there. This is avoided for pdfwrite
+ * as we check as we add each font. For adding text to an existing file though
+ * it may be more problematic */
+static void
+res_font_init(fz_context *ctx, pdf_document *doc, pdf_res_table *table)
+{
+ table->hash = fz_new_hash_table(ctx, 4096, 16, -1);
+}
+
+static void
+res_font_get_md5(fz_context *ctx, fz_buffer *buffer, unsigned char *digest)
+{
+ fz_md5 state;
+
+ fz_md5_init(&state);
+ fz_md5_update(&state, buffer->data, buffer->len);
+ fz_md5_final(&state, digest);
+}
+
+static void*
+res_font_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table, void *item,
+ void *md5)
+{
+ unsigned char digest[16];
+ fz_buffer *buffer = (fz_buffer*)item;
+ fz_hash_table *hash = table->hash;
+ pdf_res *res;
+
+ if (hash == NULL)
+ res_font_init(ctx, doc, doc->resources->font);
+ hash = doc->resources->font->hash;
+
+ /* Create md5 and see if we have the item in our table */
+ res_font_get_md5(ctx, buffer, digest);
+ res = fz_hash_find(ctx, hash, (void*)digest);
+
+ /* Return the digest value so that we can avoid having to recompute it when
+ * we come back to add the new resource reference */
+ if (res == NULL)
+ memcpy(md5, digest, 16);
+ else
+ pdf_keep_obj(ctx, res->obj);
+ return (void*)res;
+}
+
+/* Accessible methods */
+void*
+pdf_resource_table_search(fz_context *ctx, pdf_document *doc, pdf_res_table *table,
+ void *item, void *md5)
+{
+ return table->search(ctx, doc, table, item, md5);
+}
+
+void*
+pdf_resource_table_put(fz_context *ctx, pdf_res_table *table, void *key, pdf_obj *obj)
+{
+ void *result;
+ pdf_res *res = NULL;
+
+ fz_var(res);
+
+ fz_try(ctx)
+ {
+ res = fz_malloc(ctx, sizeof(pdf_res));
+ res->num = table->count + 1;
+ res->obj = obj;
+ result = fz_hash_insert(ctx, table->hash, key, (void*)res);
+ if (result != NULL)
+ {
+ fz_free(ctx, res);
+ fz_warn(ctx, "warning: hash already present");
+ }
+ else
+ {
+ table->count = table->count + 1;
+ pdf_keep_obj(ctx, obj);
+ result = res;
+ }
+ }
+ fz_catch(ctx)
+ {
+ fz_free(ctx, res);
+ fz_rethrow(ctx);
+ }
+ return result;
+}
+
+void
+pdf_resource_table_free(fz_context *ctx, pdf_document *doc)
+{
+ if (doc->resources == NULL)
+ return;
+ res_table_free(ctx, doc->resources->color);
+ res_table_free(ctx, doc->resources->font);
+ res_table_free(ctx, doc->resources->image);
+ res_table_free(ctx, doc->resources->pattern);
+ res_table_free(ctx, doc->resources->shading);
+ fz_free(ctx, doc->resources);
+ doc->resources = NULL;
+}
+
+void
+pdf_resource_table_init(fz_context *ctx, pdf_document *doc)
+{
+ fz_var(doc);
+ fz_try(ctx)
+ {
+ doc->resources = fz_calloc(ctx, 1, sizeof(pdf_resource_tables));
+ doc->resources->image = fz_calloc(ctx, 1, sizeof(pdf_res_table));
+ doc->resources->image->search = res_image_search;
+ doc->resources->font = fz_calloc(ctx, 1, sizeof(pdf_res_table));
+ doc->resources->font->search = res_font_search;
+ }
+ fz_catch(ctx)
+ {
+ if (doc->resources != NULL)
+ {
+ fz_free(ctx, doc->resources->color);
+ fz_free(ctx, doc->resources->font);
+ fz_free(ctx, doc->resources);
+ doc->resources = NULL;
+ }
+ fz_rethrow_message(ctx, "resources failed to allocate");
+ }
+}