Re: [PATCH] fchmodat, lchmod: port to buggy Linux filesystems

[Paul Eggert]

On 3/11/20 2:04 AM, Kamil Dudka wrote:
> Would not it make sense to fix this in mknod by turning the EOPNOTSUPP
> failure into a warning only?

No, because that would be a regression. mknod used to work in this case, and now 
it doesn't.

Formerly, Gnulib-using programs like mknod worked around the problem that glibc 
lchmod always failed, by silently substituting chmod for lchmod in all cases and 
not bothering to call lchmod. Now that glibc lchmod merely *sometimes* fails, 
Gnulib-using problems need a fancier workaround.

Most likely this workaround will consist of changing the 'configure'-time test 
for lchmod. Currently, the 'configure'-time test checks whether lchmod fails on 
a non-symlink. The simplest fix is to change the 'configure'-time test so that 
it always fails if the Linux kernel is being used. (Presumably we'll do the same 
for Cygwin and for Android.)