Russ Allbery wrote: > Raphael Geissert writes: > >> [1] As a matter of fact, the huge-usr-share test fails on my machine for >> some time now. I haven't been able to determine what's causing this >> behaviour. I briefly talked about it with Adam once but he was unable to >> reproduce the failure. See: >> -I: huge-usr-share-percent: arch-dep-package-has-big-usr-share 2076kB 99% >> +I: huge-usr-share-percent: arch-dep-package-has-big-usr-share 2064kB >> 100% > > I suspect some sort of file system artifact, probably on my system when I > created the tags file originally (I think that was me). du is notoriously > imprecise since the values it returns can depend on the way blocks are > allocated in the file system. In this case, I suspect that the > directories outside of usr/share are taking up some tiny amount of space, > just enough to cause the results to be weird. >
Isn't that exactly the reason of the existence of --apparent-size? > We should really just post-process the results of this test to round the > numbers off, since it's going to change slightly depending on various > things outside of our control and none of those changes really matter for > the test. Does modifying checks/huge-usr-share to call du(1) with --apparent-size help at all to get results closer to mine? Otherwise yes, we should probably just s/\d+kB/xkB/ and expect this to be the output: I: huge-usr-share-percent: arch-dep-package-has-big-usr-share xkB 100% Cheers, -- Raphael Geissert - Debian Developer www.debian.org - get.debian.net -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/[email protected]
