Kamil Dudka wrote: ... >> I saw similar numbers when timing rm -rf and chmod -R. >> Thus, I'm reluctant to impose such a penalty without >> first exhausting all other possibilities. > > Yes, it makes sense. However what other possibilities do we have actually?
Change the kernel, or find a way to make fts do what we want without the added cost. > Preferring higher performance in some rear cases over correct behavior > in more common cases is of course not a possibility for me. If you are implying that I might make such a trade-off, you have not been paying attention ;-) BTW, in my experience, this is *not* a common case at all. In fact, the patch we're considering impacts each and every directory access in *all* hierarchy traversals, when only a very small number of them actually need the work-around.
