Bruno Haible wrote: > Hi Jim, > >> However, I'm not sure it's portable enough. >> Do you know when the f_frsize member began to be useful? >> Without knowing that, I cannot judge whether this introduces >> a portability problem on older glibc, kernels or file systems. > > There are also other platforms than glibc/Linux. > > Upon inspection, it appears that the fu_cv_sys_stat_statfs2_bsize > test in m4/fsusage.m4 (labelled "AIX") would succeed on > glibc, MacOS X, FreeBSD, NetBSD, OpenBSD, AIX, HP-UX, Solaris, Cygwin > and out of these, only > glibc/Linux, glibc/Hurd, Solaris > do have an f_frsize member in 'struct statfs'. > > But that test is preceded by the test for fu_cv_sys_stat_statvfs (labelled > "SVR4", which I would expect to succeed on > glibc/Hurd, FreeBSD, NetBSD, AIX, HP-UX, IRIX, Solaris, Cygwin, Interix, > BeOS > > So it may be that actually only glibc/Linux platforms are affected by Miklos' > patch, but without a clear overview of which platform ends up using which > code, > I can't assert this.
Hi Bruno, I realize that there is some small risk for other platforms, but am willing to discover that if/when people actually test the new code on those systems. If necessary, we add additional guards later.
