Aristotle Pagaltzis wrote: >Is there a chance that that was also on the Linux subsystem for Windows? It's possible. Neither the ultimate host system nor the filesystem was specified. >Does File::Find (or something else) need a patch to detect the situation >and adjust itself accordingly? Preferably yes, File::Find should handle it automatically. But it's not really a per-platform question, it's per-filesystem, like case sensitivity of filenames. Anything we do statically at build time will suffer both false positives and false negatives. Really the reliability of the nlink==2 trick ought to be advertised through pathconf(2), but it isn't. In the absence of that, there are still some ways we could autodetect it. statfs(2) on Linux advertises a `filesystem type', and we could use that to determine whether nlink is reliable for known types. -zeframThread Previous | Thread Next