Add a comment explaning necessity of page removal during #relayout

Autoported commit.
Original commit hash: [cc0f55342]

Author: Snipx

itext/itext.layout/itext/layout/Document.cs

@@ -212,6 +212,10 @@ public virtual void Relayout() {
             if (nextRelayoutRenderer == null || !(nextRelayoutRenderer is RootRenderer)) {
                 nextRelayoutRenderer = new DocumentRenderer(this, immediateFlush);
             }
+            // Even though #relayout() only makes sense when immediateFlush=false and therefore no elements
+            // should have been written to document, still empty pages are created during layout process
+            // because we need to know the effective page size which may differ from page to page.
+            // Therefore, we remove all the pages that might have been created before proceeding to relayout elements.
             while (pdfDocument.GetNumberOfPages() > 0) {
                 pdfDocument.RemovePage(pdfDocument.GetNumberOfPages());
             }

itext/itext.layout/itext/layout/renderer/DocumentRenderer.cs

@@ -181,7 +181,7 @@ protected internal virtual PageSize AddNewPage(PageSize customPageSize) {
 
         /// <summary>Ensures that PDF document has n pages.</summary>
         /// <remarks>
-        /// Ensures that PDF document has n pages. If document has less pages,
+        /// Ensures that PDF document has n pages. If document has fewer pages,
         /// adds new pages by calling
         /// <see cref="AddNewPage(iText.Kernel.Geom.PageSize)"/>
         /// method.

port-hash

@@ -1 +1 @@
-bbabbfd61feab7ef47a538e9f0f4248255ff138a
+cc0f553429cb077f8d86c93ec0c56a0b9f0d19ce