Bruno Haible <bruno <at> clisp.org> writes: > > James Youngman wrote: > > this version of the patch was not tested on NetBSD-3.1 with UFS2. Sigh. > > But now at least you get a ready-to-use package for testing by doing > > ./gnulib-tool --create-testdir --dir=/tmp/testdir --with-tests stat-time
On CVS cygwin: $ gnulib-tool --with-tests --test stat-time ... checking for struct stat.st_atim.tv_nsec... yes checking whether struct stat.st_atim is of type struct timespec... yes checking for struct stat.st_birthtimespec.tv_sec... no checking for struct stat.st_birthtime... yes checking for struct stat.st_birthtimensec... no checking for struct stat.st_spare4... no checking for struct stat.st_birthtimespec.tv_nsec... no checking for struct stat.st_birthtime... (cached) yes checking for struct stat.st_birthtimensec... (cached) no checking for struct stat.st_spare4... (cached) no configure: creating ./config.status ... make[4]: Entering directory `/home/eblake/gnulib/testdir14664/build/gltests' PASS: test-stat-time.exe PASS: test-time.exe ================== All 2 tests passed ================== ... I need to repeat the test with cygwin 1.5.24, which does not have st_birthtime. Meanwhile, I would recommend getting rid of any test for st_spare1 or st_spare4 in m4/stat-time.m4. st_spare1 has not existed in cygwin for years now, and that was all the comment could tie it to, so it is worthless bloat unless someone can point out another platform where it is in use. And cygwin 1.5.24 has st_spare4 instead of st_birthtime, but if you are compiling on a cygwin machine where st_spare4 is visible, then that field will not be populated with birthtime information. So checking for it is pointless; the check for st_birthtime is adequate. On the other hand, you should add a check for st_birthtim.tv_nsec, so that you can access the subsecond resolution of cygwin's birthtime. -- Eric Blake
