Differences due to different locale settings

[adangel]

Depending on the current locale, the messages that PMD produces for violations, might be different.
This might indeed be a problem of PMD itself (we don't have localized messages yet...).

Example:
http://chunk.io/pmd/7c0c68cd0bfa4aa781f718d56328afbd/diff/index.html

"Found 'UR'-anomaly for variable 'ast' (lines '1.086'-'1.089'). " vs. "
"Found 'UR'-anomaly for variable 'ast' (lines '1,086'-'1,089')."

The correct/reproducible setting can be achieved, by setting export LANG=en_US.UTF-8 before executing PMD.

We should at least add the locale info to the summary.
To prevent such issues, we can set the env variable accordingly.

There is a similar difference, when the baseline is built with a different JDK version. Then the line numbers in stacktrace tend to change and this is detected as add/removed errors.
So, we should also add the JDK version to the summary table of both base and patch columns.