Thanks for all the work so far! I am also curious if we are making more copies of strings (just a general concern) and if that has any perf impact for effected tables — perhaps there are one or two that are string copy heavy that are now heavier?

In fileops.cpp, there will be... but I haven’t touched that one yet. There was a comment on #6190 that this PR didn’t fix fileops.cpp case and I checked and those were all *A calls right now.