Migration Steps

The issue fixed in this version (PFIMCORE-653) occurred only at Unix systems and routes using:

<to uri=”pfx-csv:…:/>

<to uri="file:…fileExists=Append"/>

As recordSeparator was not specified in pfx-csv, default \r\n was used to distinguish rows. Since Camel 3.0, the Camel file component with the option fileExists=Append did not recognize \r\n on Unix systems, resulting in duplicating the end-of-file chars, which led to ‘\r\n\n’ as a row separator.

This was fixed by introducing a platform dependent recordSeparator in the pfx-csv component. This separator is always recognized by the file component, thus the duplicity does not happen anymore.

From now on, the platform-dependent recordSeparator is used (\r\n on Windows, \n on Unix) in the pfx-csv component to separate rows, unless specified otherwise. If you counted on \r\n at Unix systems by default, this has changed. However, we recommend to enforce recordSeparator by setting the respective property in the pfx-csv component.

Bugfixes