class should use python's standard cvs reader/writer modules to provide proper csv files. Otherwise like currently it produces csv files that can't be parsed due to incorrect quoting/delimiters mismatches.
At, http://lists.openembedded.org/pipermail/openembedded-core/2015-June/105579.html
Merged at, http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=20a71d63fc590e8d5f3fb537a0afd5895661385a
Can you backport also same bug fix to fido branch ? Thanks.