popplerload hacking

metadata, relative filenames

TODO

@@ -1,10 +1,17 @@
-- pooperload URI generation needs fixing, it currently does not support 
-  relative paths
-
 - popplerload generates ARGB? need to do our own byteswap
 
 - add tests for popplerload, check docs, update c++
 
+- add load PDF from buffer
+
+- add file type sniffing .. "%PDF" should be enough
+
+- metadata: xres/yres
+
+- trim page edges? we often have black now
+
+- does poppler support other formats? SVG / PS / EPS
+
 - new vips_reduce:
 
 	affine

doc/gtk-doc.make

@@ -25,7 +25,6 @@ TARGET_DIR=$(HTML_DIR)/$(DOC_MODULE)
 
 SETUP_FILES = \
 	$(content_files)		\
-	$(expand_content_files)		\
 	$(DOC_MAIN_SGML_FILE)		\
 	$(DOC_MODULE)-sections.txt	\
 	$(DOC_MODULE)-overrides.txt
@@ -87,7 +86,7 @@ GTK_DOC_V_SETUP_0=@echo "  DOC   Preparing build";
 
 setup-build.stamp:
 	-$(GTK_DOC_V_SETUP)if test "$(abs_srcdir)" != "$(abs_builddir)" ; then \
-	    files=`echo $(SETUP_FILES) $(DOC_MODULE).types`; \
+	    files=`echo $(SETUP_FILES) $(expand_content_files) $(DOC_MODULE).types`; \
 	    if test "x$$files" != "x" ; then \
 	        for file in $$files ; do \
 	            destdir=`dirname $(abs_builddir)/$$file`; \
@@ -119,7 +118,7 @@ scan-build.stamp: setup-build.stamp $(HFILE_GLOB) $(CFILE_GLOB)
 	$(GTK_DOC_V_INTROSPECT)if grep -l '^..*$$' $(DOC_MODULE).types > /dev/null 2>&1 ; then \
 	    scanobj_options=""; \
 	    gtkdoc-scangobj 2>&1 --help | grep  >/dev/null "\-\-verbose"; \
-	    if test "$$?" = "0"; then \
+	    if test "$(?)" = "0"; then \
 	        if test "x$(V)" = "x1"; then \
 	            scanobj_options="--verbose"; \
 	        fi; \
@@ -163,17 +162,17 @@ GTK_DOC_V_XREF=$(GTK_DOC_V_XREF_$(V))
 GTK_DOC_V_XREF_=$(GTK_DOC_V_XREF_$(AM_DEFAULT_VERBOSITY))
 GTK_DOC_V_XREF_0=@echo "  DOC   Fixing cross-references";
 
-html-build.stamp: sgml.stamp $(DOC_MAIN_SGML_FILE) $(content_files) $(expand_content_files)
+html-build.stamp: sgml.stamp $(DOC_MAIN_SGML_FILE) $(content_files)
 	$(GTK_DOC_V_HTML)rm -rf html && mkdir html && \
 	mkhtml_options=""; \
 	gtkdoc-mkhtml 2>&1 --help | grep  >/dev/null "\-\-verbose"; \
-	if test "$$?" = "0"; then \
+	if test "$(?)" = "0"; then \
 	  if test "x$(V)" = "x1"; then \
 	    mkhtml_options="$$mkhtml_options --verbose"; \
 	  fi; \
 	fi; \
 	gtkdoc-mkhtml 2>&1 --help | grep  >/dev/null "\-\-path"; \
-	if test "$$?" = "0"; then \
+	if test "$(?)" = "0"; then \
 	  mkhtml_options="$$mkhtml_options --path=\"$(abs_srcdir)\""; \
 	fi; \
 	cd html && gtkdoc-mkhtml $$mkhtml_options $(MKHTML_OPTIONS) $(DOC_MODULE) ../$(DOC_MAIN_SGML_FILE)
@@ -195,11 +194,11 @@ GTK_DOC_V_PDF=$(GTK_DOC_V_PDF_$(V))
 GTK_DOC_V_PDF_=$(GTK_DOC_V_PDF_$(AM_DEFAULT_VERBOSITY))
 GTK_DOC_V_PDF_0=@echo "  DOC   Building PDF";
 
-pdf-build.stamp: sgml.stamp $(DOC_MAIN_SGML_FILE) $(content_files) $(expand_content_files)
+pdf-build.stamp: sgml.stamp $(DOC_MAIN_SGML_FILE) $(content_files)
 	$(GTK_DOC_V_PDF)rm -f $(DOC_MODULE).pdf && \
 	mkpdf_options=""; \
 	gtkdoc-mkpdf 2>&1 --help | grep  >/dev/null "\-\-verbose"; \
-	if test "$$?" = "0"; then \
+	if test "$(?)" = "0"; then \
 	  if test "x$(V)" = "x1"; then \
 	    mkpdf_options="$$mkpdf_options --verbose"; \
 	  fi; \
@@ -224,15 +223,12 @@ clean-local:
 	@if echo $(SCAN_OPTIONS) | grep -q "\-\-rebuild-types" ; then \
 	  rm -f $(DOC_MODULE).types; \
 	fi
-	@if echo $(SCAN_OPTIONS) | grep -q "\-\-rebuild-sections" ; then \
-	  rm -f $(DOC_MODULE)-sections.txt; \
-	fi
 
 distclean-local:
 	@rm -rf xml html $(REPORT_FILES) $(DOC_MODULE).pdf \
 	    $(DOC_MODULE)-decl-list.txt $(DOC_MODULE)-decl.txt
 	@if test "$(abs_srcdir)" != "$(abs_builddir)" ; then \
-	    rm -f $(SETUP_FILES) $(DOC_MODULE).types; \
+	    rm -f $(SETUP_FILES) $(expand_content_files) $(DOC_MODULE).types; \
 	fi
 
 maintainer-clean-local:

libvips/foreign/popplerload.c

@@ -45,6 +45,7 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
+#include <errno.h>
 
 #include <vips/vips.h>
 #include <vips/buf.h>
@@ -110,15 +111,38 @@ vips_foreign_load_poppler_get_flags( VipsForeignLoad *load )
 	return( VIPS_FOREIGN_PARTIAL );
 }
 
+/* String-based metadatra fields we extract.
+ */
+typedef struct _VipsForeignLoadPopperMetadata {
+	char *(*poppler_fetch)( PopplerDocument *doc );
+	char *field;
+} VipsForeignLoadPopperMetadata;
+
+static VipsForeignLoadPopperMetadata vips_foreign_load_poppler_metadata[] = {
+	{ poppler_document_get_title, "poppler-title" },
+	{ poppler_document_get_author, "poppler-author" },
+	{ poppler_document_get_subject, "poppler-subject" },
+	{ poppler_document_get_keywords, "poppler-keywords" },
+	{ poppler_document_get_creator, "poppler-creator" },
+	{ poppler_document_get_producer, "poppler-producer" },
+	{ poppler_document_get_metadata, "poppler-metadata" },
+};
+static int n_metadata = VIPS_NUMBER( vips_foreign_load_poppler_metadata );
+
 static void
 vips_foreign_load_poppler_parse( VipsForeignLoadPoppler *poppler, 
 	VipsImage *out )
 {
+	PopplerRectangle crop_box;
 	double width;
 	double height;
+	int i;
+	char *str;
 
 	poppler_page_get_size( poppler->page, &width, &height ); 
 
+	poppler_page_get_crop_box( poppler->page, &crop_box ); 
+
 	vips_image_init_fields( out, 
 		width * poppler->scale, height * poppler->scale, 
 		4, VIPS_FORMAT_UCHAR,
@@ -129,6 +153,21 @@ vips_foreign_load_poppler_parse( VipsForeignLoadPoppler *poppler,
 	/* We render to a linecache, so fat strips work well.
 	 */
         vips_image_pipelinev( out, VIPS_DEMAND_STYLE_FATSTRIP, NULL );
+
+	/* Extract and attach metadata.
+	 */
+	vips_image_set_int( out, "poppler-n_pages", 
+		poppler_document_get_n_pages( poppler->doc ) );
+
+	for( i = 0; i < n_metadata; i++ ) {
+		VipsForeignLoadPopperMetadata *metadata = 
+			&vips_foreign_load_poppler_metadata[i];
+
+		if( (str = metadata->poppler_fetch( poppler->doc )) ) { 
+			vips_image_set_string( out, metadata->field, str ); 
+			g_free( str );
+		}
+	}
 }
 
 static int
@@ -137,11 +176,24 @@ vips_foreign_load_poppler_header( VipsForeignLoad *load )
 	VipsObjectClass *class = VIPS_OBJECT_GET_CLASS( load );
 	VipsForeignLoadPoppler *poppler = (VipsForeignLoadPoppler *) load;
 
+	char *path;
 	GError *error = NULL;
 
 	poppler->scale = poppler->dpi / 72.0;
 
-	poppler->uri = g_strdup_printf( "file://%s", poppler->filename ); 
+	/* We need an absolute path for a URI.
+	 */
+	if( !(path = realpath( poppler->filename, NULL )) ) {
+		vips_error_system( errno, class->nickname,
+			"%s", _( "unable to form filename" ) ); 
+		return( -1 );
+	}
+	if( !(poppler->uri = g_filename_to_uri( path, NULL, &error )) ) { 
+		free( path );
+		vips_g_error( &error );
+		return( -1 ); 
+	}
+	free( path );
 
 	if( !(poppler->doc = poppler_document_new_from_file( 
 		poppler->uri, NULL, &error )) ) {