With the above snapshot and --disable-acl, all of the default tests, including those enabled via RUN_VERY_EXPENSIVE_TESTS pass. as for the check-root tests: this one now passes: special-bits these fa
Thanks for testing it. It looks like they're all due to the fact that you used NON_ROOT_USERNAME=nobody and "nobody" lacks access to required files. Can you run it again, as recommended in README i.e
D'oh -- should've read that. So, with 6.10 and --disable-acl, all of the non-root tests pass and a single root-only test fails: % tar xf coreutils-6.10.tar.gz % cd coreutils-6.10 % ./configure --disa
... Thanks for the report and details. That's probably due to your running the root-only tests in a directory with the sticky set-GID bit set. If that's the problem, here's an untested fix: * tests/c
Unfortunately, that doesn't seem to help. % tar xf coreutils-6.10.tar.gz % cd coreutils-6.10 % cat ~/foo.patch diff --git a/tests/cp/preserve-gid b/tests/cp/preserve-gid index 5f2c050..e96a0bd 100755
Thanks for checking. That suggests there's a more fundamental problem. Please do this as root: cd coreutils-6.10/src ./id -a ./rm -rf f g echo a > f ./chown +0:+0 f ls -ld . f ./cp f g ls -l g and lo
For some reason it doesn't. uid=0(root) gid=0(wheel) groups=0(wheel),1(daemon),8(procview), 2(kmem),29(certusers),3(sys),9(procmod),4(tty), 102(com.apple.sharepoint.group.2),5(operator),80(admin),20(
That suggests that the bogus group is set by cp's open call. I suspect some sort of ACL mechanism that specifies "staff" as the default group -- or maybe some rule says "inherit group from parent dir
The darwin equivalent appears to be ktrace. Andreas. -- Andreas Schwab, SuSE Labs, address@hidden SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany PGP key fingerprint = 58CA 54C7
The latter seems to be the case. /tmp uid=0(root) gid=0(wheel) groups=0(wheel),1(daemon),8(procview), 2(kmem),29(certusers),3(sys),9(procmod),4(tty), 102(com.apple.sharepoint.group.2),5(operator),80(
It does, indeed. Please see if this makes the test pass: Avoid cp/preserve-gid test failure on Mac OS 10.5.1 (Darwin 9.1) * tests/cp/preserve-gid: Set group as well as owner on ".". Reported by Elias
The test still fails, but it gets a whole lot further :D -- Elias Attachment: preserve-gid_test_new Description: Text document Attachment: preserve-gid_test.diff Description: Text document Attachment
Good. Progress :) Here's an additional patch that should get us past the latest failure. diff --git a/tests/cp/preserve-gid b/tests/cp/preserve-gid index 9537dc0..15bb51c 100755 -- a/tests/cp/preserv
Unfortunately, it doesn't. -- Elias Attachment: preserve-gid_test_new2 Description: Text document Attachment: preserve-gid_test2.diff Description: Text document Attachment: pgpNgJouIFK6n.pgp Descript
... But it came close :-) Thanks for the quick feedback. I should have matched the whole command, from "setuidgid -g...cp", not just "cp". Here's a better one: * tests/cp/preserve-gid: accommodate Da
More progress! (still fails) -- Elias Attachment: preserve-gid_test_new3 Description: Text document Attachment: preserve-gid_test3.diff Description: Text document Attachment: pgpZOfRRHjzkV.pgp Descri
Thanks for testing that. In case you're still inclined, here's a coreutils development snapshot: It includes the gnulib fix and tests. coreutils snapshot: http://meyering.net/cu/coreutils-ss.tar.gz 8
My big test factorization changes broke at least two tests. Thanks to Peter Fales for spotting and reporting the new failures. Here is a new snapshot that fixes them: coreutils snapshot: http://meyer
There have been over 50 change-sets since the last one, so... coreutils snapshot: http://meyering.net/cu/coreutils-ss.tar.gz 8.9 MB http://meyering.net/cu/coreutils-ss.tar.lzma 3.6 MB http://meyering