Now that we've again enabled array-bounds warnings in S3_0 also, build on Xenial seems to fail. It complains about activity_id_exists(). I don't see anything wrong with the function, and the fact that no newer compiler (with better heuristics) gives such warning leads me to believe it's a false positive. There's no similar problem with other branches.
So I think it's enough to just rearrange the code so that gcc-5.4 no longer complains about it, and not to look for an actual bug.
Now that we've again enabled array-bounds warnings in S3_0 also, build on Xenial seems to fail. It complains about activity_id_exists(). I don't see anything wrong with the function, and the fact that no newer compiler (with better heuristics) gives such warning leads me to believe it's a false positive. There's no similar problem with other branches.
So I think it's enough to just rearrange the code so that gcc-5.4 no longer complains about it, and not to look for an actual bug.
Will treat as a build emergency.