Re: segfault with exceedingly long path

[Bart Schaefer <schaefer@brasslantern.com>]

On Jan 19,  7:10pm, Peter Stephenson wrote:
}
} I'm still in favour of reducing dependency on PATH_MAX wherever
} possible, since any arbitrary limit we don't actually need we don't
} actually want

In this particular case even if the shell were able to deal with an
"unlimited" directory nesting depth, system calls like chdir(2) would
begin to fail.  fchdir() sort of works around it by not passing the
path name, but you still must somehow obtain a file descriptor to the
directory, so it's just pushing the problem up a level.