diff --git a/pmd-core/src/main/java/net/sourceforge/pmd/cpd/Mark.java b/pmd-core/src/main/java/net/sourceforge/pmd/cpd/Mark.java
index 088da59d39..ab47816341 100644
--- a/pmd-core/src/main/java/net/sourceforge/pmd/cpd/Mark.java
+++ b/pmd-core/src/main/java/net/sourceforge/pmd/cpd/Mark.java
@@ -1,3 +1,6 @@
+/**
+ * BSD-style license; for more info see http://pmd.sourceforge.net/license.html
+ */
 package net.sourceforge.pmd.cpd;
 
 public class Mark implements Comparable<Mark> {
diff --git a/src/site/markdown/overview/changelog.md b/src/site/markdown/overview/changelog.md
index 638ca4b34e..5ca8342881 100644
--- a/src/site/markdown/overview/changelog.md
+++ b/src/site/markdown/overview/changelog.md
@@ -12,6 +12,7 @@
 **Feature Requests and Improvements:**
 
 * XML: Line numbers for XML documents are more accurate. This is a further improvement of [#1054](https://sourceforge.net/p/pmd/bugs/1054/).
+* CPD: New output format 'csv_with_linecount_per_file'
 
 **New/Modified Rules:**
 
@@ -24,6 +25,7 @@
 * [#15](https://github.com/adangel/pmd/pull/15): (pmd-cs) Fixed incorrect line numbers after mutiline comments and verbatim strings.
 * [#16](https://github.com/adangel/pmd/pull/16): Fixed several C++ lexical / tokenize errors.
 * [#17](https://github.com/adangel/pmd/pull/17): Fixed '--files' command line option of CPD, so it also works for files and not only for directories.
+* [#18](https://github.com/adangel/pmd/pull/18): Created extra CSV output format `csv_with_linecount_per_file` which outputs the correct line count per file.
 
 **Bugfixes:**
 
@@ -33,3 +35,11 @@
 * [#1299](https://sourceforge.net/p/pmd/bugs/1299/): MethodReturnsInternalArray false positive
 * [#1306](https://sourceforge.net/p/pmd/bugs/1306/): False positive on duplicate when using static imports
 * [#1308](https://sourceforge.net/p/pmd/bugs/1308/): PMD runs endlessly on some generated files
+
+**API Changes:**
+
+* `net.sourceforge.pmd.cpd.Match.iterator()` now returns an iterator of the new type `net.sourceforge.pmd.cpd.Mark` instead
+  of TokenEntry. A `Mark` contains all the informations about each single duplication, including the TokenEntry via `Mark.getToken()`.
+  This Mark is useful for reporting the correct line count for each duplication. Previously only one line count was available.
+  As for some languages CPD can be instructed to ignore comments, the line count could be different in the different files
+  for the same duplication.