update mupdf

mupdf/docs/src/mutool-object-pdf-document.rst

@@ -104,6 +104,35 @@ With :title:`MuPDF` it is also possible to create, edit and manipulate :title:`P
         var lang = pdfDocument.getLanguage();
 
 
+
+.. method:: rearrangePages(pages)
+
+    |mutool_tag|
+
+    Rearrange (re-order and/or delete) pages in the `PDFDocument`.
+
+    The pages in the document will be rearranged according to the input list.
+    Any pages not listed will be removed, and pages may be duplicated by
+    listing them multiple times.
+
+    The PDF objects describing removed pages will remain in the file and take
+    up space (and can be recovered by forensic tools) unless you save with the
+    `garbage` option.
+
+    N.B. the `PDFDocument` should not be used for anything except saving after rearranging the pages (FIXME).
+
+    :arg pages: An array of page numbers (0-based).
+
+    |example_tag|
+
+    .. code-block:: javascript
+
+        var document = new Document.openDocument("my_pdf.pdf");
+        pdfDocument.rearrangePages([3,2]);
+        pdfDocument.save("fewer_pages.pdf", "garbage");
+
+
+
 .. method:: save(fileName, options)
 
     |mutool_tag|

mupdf/include/mupdf/pdf/clean.h

@@ -37,4 +37,9 @@ typedef struct
 */
 void pdf_clean_file(fz_context *ctx, char *infile, char *outfile, char *password, pdf_clean_options *opts, int retainlen, char *retainlist[]);
 
+/*
+	Recreate page tree to include only the pages listed in the array, in the order listed.
+*/
+void pdf_rearrange_pages(fz_context *ctx, pdf_document *doc, int count, int *pages);
+
 #endif

mupdf/platform/java/jni/pdfdocument.c

@@ -1775,3 +1775,34 @@ FUN(PDFDocument_appendExplicitDestToURI)(JNIEnv *env, jclass cls, jstring jurl,
 	fz_free(ctx, uri);
 	return juri;
 }
+
+JNIEXPORT void JNICALL
+FUN(PDFDocument_rearrangePages)(JNIEnv *env, jobject self, jobject jpages)
+{
+	fz_context *ctx = get_context(env);
+	pdf_document *pdf = from_PDFDocument(env, self);
+	jsize len = 0;
+	int *pages = NULL;
+
+	if (!ctx || !pdf) return;
+
+	len = (*env)->GetArrayLength(env, jpages);
+	fz_try(ctx)
+		pages = fz_malloc_array(ctx, len, int);
+	fz_catch(ctx)
+		jni_rethrow_void(env, ctx);
+
+	(*env)->GetIntArrayRegion(env, jpages, 0, len, pages);
+	if ((*env)->ExceptionCheck(env))
+	{
+		fz_free(ctx, pages);
+		return;
+	}
+
+	fz_try(ctx)
+		pdf_rearrange_pages(ctx, pdf, len, pages);
+	fz_always(ctx)
+		fz_free(ctx, pages);
+	fz_catch(ctx)
+		jni_rethrow_void(env, ctx);
+}

mupdf/platform/java/src/com/artifex/mupdf/fitz/PDFDocument.java

@@ -149,6 +149,8 @@ public PDFObject addPage(Rect mediabox, int rotate, PDFObject resources, String
 	public native boolean canBeSavedIncrementally();
 	public native boolean isRedacted();
 
+	public native void rearrangePages(int[] pages);
+
 	public native void save(String filename, String options);
 
 	protected native void nativeSaveWithStream(SeekableInputOutputStream stream, String options);

mupdf/source/fitz/xml.c

@@ -28,6 +28,8 @@
 
 #include <gumbo.h>
 
+#define FZ_XML_MAX_DEPTH 4096
+
 /* #define FZ_XML_SEQ */
 
 static const struct { const char *name; int c; } html_entities[] = {
@@ -496,6 +498,8 @@ static void xml_emit_open_tag(fz_context *ctx, struct parser *parser, const char
 
 	parser->head = head;
 	parser->depth++;
+	if (parser->depth >= FZ_XML_MAX_DEPTH)
+		fz_throw(ctx, FZ_ERROR_SYNTAX, "too deep xml element nesting");
 }
 
 static void xml_emit_att_name(fz_context *ctx, struct parser *parser, const char *a, const char *b)

mupdf/source/html/html-layout.c

@@ -999,6 +999,17 @@ static float largest_max_width(fz_context *ctx, fz_html_box *box)
 	return r_max;
 }
 
+static void squish_block(fz_context *ctx, fz_html_box *box)
+{
+	fz_html_box *child;
+
+	box->s.layout.b = box->s.layout.y;
+	if (box->type == BOX_FLOW)
+		return;
+	for (child = box->down; child; child = child->next)
+		squish_block(ctx, child);
+}
+
 static void layout_table_row(fz_context *ctx, layout_data *ld, fz_html_box *row, int ncol, struct column_width *colw, float spacing)
 {
 	fz_html_box *cell, *child;
@@ -1100,7 +1111,11 @@ static void layout_table(fz_context *ctx, layout_data *ld, fz_html_box *box, fz_
 	}
 
 	/* TODO: remove 'vertical' margin adjustments across automatic page breaks */
-	if (layout_block_page_break(ctx, ld, &top->s.layout.b, box->style->page_break_before))
+	if (restart && restart->start != NULL)
+	{
+		/* We're still skipping, don't check for pagebreak before! */
+	}
+	else if (layout_block_page_break(ctx, ld, &top->s.layout.b, box->style->page_break_before))
 		eop = 1;
 
 	/* Position table in box flow, and add margins and padding */
@@ -1226,19 +1241,23 @@ static void layout_table(fz_context *ctx, layout_data *ld, fz_html_box *box, fz_
 			row->s.layout.x = box->s.layout.x;
 			row->s.layout.w = box->s.layout.w;
 			row->s.layout.y = row->s.layout.b = box->s.layout.b;
+			row->s.layout.b = row->s.layout.y;
 
 			if (restart && restart->start != NULL)
 			{
 				if (restart->start == row)
 					restart->start = NULL;
 				else
+				{
+					squish_block(ctx, row);
 					continue; /* still skipping */
 			}
+			}
 
 			layout_table_row(ctx, ld, row, ncol, colw, spacing);
 
 			/* If the row doesn't fit on the current page, break here and put the row on the next page.
-			 * Unless the row was at the very start of the page, in which case it'll overfloaw instead.
+			 * Unless the row was at the very start of the page, in which case it'll overflow instead.
 			 * FIXME: Don't overflow, draw twice with offset to break it abruptly at the page border!
 			 */
 			if (ld->page_h > 0)
@@ -1268,6 +1287,22 @@ static void layout_table(fz_context *ctx, layout_data *ld, fz_html_box *box, fz_
 		fz_free(ctx, colw);
 	fz_catch(ctx)
 		fz_rethrow(ctx);
+
+	if (restart && restart->start != NULL)
+	{
+		/* We're still skipping, don't check for pagebreak after! */
+	}
+	else if (layout_block_page_break(ctx, ld, &top->s.layout.b, box->style->page_break_after))
+	{
+		if (restart && restart->end == NULL)
+		{
+			if (restart->potential)
+				restart->end = restart->potential;
+			else
+				restart->end = box;
+			return;
+		}
+	}
 }
 
 /* === LAYOUT BLOCKS === */
@@ -1315,7 +1350,11 @@ static void layout_block(fz_context *ctx, layout_data *ld, fz_html_box *box, fz_
 	}
 
 	/* TODO: remove 'vertical' margin adjustments across automatic page breaks */
-	if (layout_block_page_break(ctx, ld, &top->s.layout.b, style->page_break_before))
+	if (restart && restart->start != NULL)
+	{
+		/* We're still skipping, don't check for pagebreak before! */
+	}
+	else if (layout_block_page_break(ctx, ld, &top->s.layout.b, style->page_break_before))
 		eop = 1;
 
 	/* Important to remember that box->{x,y,w,b} are the coordinates of the content. The
@@ -1416,7 +1455,21 @@ static void layout_block(fz_context *ctx, layout_data *ld, fz_html_box *box, fz_
 		box->s.layout.b += fz_from_css_number_scale(style->line_height, em);
 	}
 
-	(void) layout_block_page_break(ctx, ld, &box->s.layout.b, style->page_break_after);
+	if (restart && restart->start != NULL)
+	{
+		/* We're still skipping, don't check for pagebreak after! */
+	}
+	else if (layout_block_page_break(ctx, ld, &box->s.layout.b, style->page_break_after))
+	{
+		if (restart && restart->end == NULL)
+		{
+			if (restart->potential)
+				restart->end = restart->potential;
+			else
+				restart->end = box;
+			return;
+		}
+	}
 }
 
 /* === LAYOUT === */