The include and exclude filters should be stored in the xml result files. The FindBugs GUI and any other program reader can then read in the filters and accurately reproduce a bug analysis.
Just to make sure I understand, this is to allow a reanalysis to be performed. For the results contained in the xml file, they aren't needed.
The question here is consistency between the inlined filters and external files. What if the specified include/exclude files are modified? Would we want to use the inlined filters, or the values in the files. How potentially confusing could this be for people?
yes, that is correct. The maven plugin already does an analysis and store the results. Through maven you can run the ide where you may like to see the filters and rerun an analysis with the filters as is or to modify them. If in the analysis file then the Eclipse plugin would have the means to reanalyze with the common filter configuration if desired. This may need to be address in the Eclipse plugin as a separate issue.