On Fri July 22 2011 15:25:13 Jim Meyering wrote: > Kamil Dudka wrote: > > On Wednesday 06 April 2011 16:49:31 Kamil Dudka wrote: > >> the attached patch allows ls(1) from coreutils to eliminate unnecessary > >> calls of getxattr() and thus prevents it from triggerring unnecessary > >> mounts while listing files. The feature is enabled automatically > >> whenever acl_extended_file_nofollow() is available at build-time. See > >> the following bug for more details: > >> > >> https://bugzilla.redhat.com/692823 > >> > >> Thanks in advance for considering the patch! > > > > Attached is a new version with a typo fixed (spotted by Ondrej Vasik). > > Thanks, Kamil. > Here's what I'm about to push. > I've tweaked the commit message, adjusted the added comment to mention > the race condition and reduced indentation of the added cpp directives > so that this file passes 'make check's cppi test.
Jim, thank you for taking care of the patch. I completely forgot about this one and then left for two weeks long vacation. I have retested the current upstream version of gnulib/coreutils and it seems to work as expected with up2date libacl installed. As a side note I needed to add fclose to gnulib_modules in bootstrap.conf in order to be able to build the latest coreutils. Not sure if it is a real bug or something stale on my box. > I also ensured that the new test does detect/use > acl_extended_file_nofollow on Fedora 15 and that all of > the existing tests still pass. It would be nice to add a > test for this (strace-based -- not automount-based) What do you actually want to test? coreutils? gnulib? libacl? Kamil
