Bruno Haible [EMAIL PROTECTED] wrote:
Still about POSIX-draft like ACLs. Now that the ACL triviality test has been
moved into a function 'acl_access_nontrivial', the code that uses it can
be simplified: coalesce two 'if' branches together. Then exploit the fact
that chmod_or_fchmod has already
Bruno Haible [EMAIL PROTECTED] wrote:
In the MacOS X case, we need to create an empty acl_t object. This simplifies
the code for doing that, so that we don't need to know about the (very
platform dependent) textual representation.
2008-06-07 Bruno Haible [EMAIL PROTECTED]
*
Bruno Haible [EMAIL PROTECTED] wrote:
This fixes a memory leak in the MacOS X specific code that I introduced on
2008-05-22.
2008-06-07 Bruno Haible [EMAIL PROTECTED]
Fix memory leak introduced on 2008-05-22.
* lib/set-mode-acl.c (qset_acl) [!MODE_INSIDE_ACL]: Free ACLs after
Bruno Haible [EMAIL PROTECTED] wrote:
On OSF/1, trying to retrieve the AC_TYPE_DEFAULT ACL always returns NULL
with unset errno (errno = 0 in my case). Since our code is not prepared
for this, it's best to just disable this call.
2008-06-07 Bruno Haible [EMAIL PROTECTED]
Improve
Bruno Haible [EMAIL PROTECTED] wrote:
On MacOS X the testsuite still fails:
...
Please review (although I commit it immediately, because more patches are
coming.)
...
+ if (ret != 0)
+{
+ int saved_errno = errno;
+
+ if (ACL_NOT_WELL_SUPPORTED (errno) !(acl_entries (acl)
exploit the fact
that chmod_or_fchmod has already the right return value convention.
More simplifications of the same kind:
2008-06-08 Bruno Haible [EMAIL PROTECTED]
* lib/set-mode-acl.c (qset_acl): Trivial code simplifications.
*** lib/set-mode-acl.c.orig 2008-06-08
Jim Meyering wrote:
+ if (ret != 0)
+{
+ int saved_errno = errno;
+
+ if (ACL_NOT_WELL_SUPPORTED (errno) !(acl_entries (acl) 0))
+{
Why use the negation?
That's for analogy with the Linux/FreeBSD/... code a couple of lines above.
The acl_entries (acl) 0
So far, the test results on Solaris and Cygwin are:
Solaris 10:
PASS: test-file-has-acl.sh
PASS: test-set-mode-acl.sh
files tmpfile0 and tmpfile2 have different number of ACLs: 5 and 4
FAIL: test-copy-acl.sh
Cygwin:
PASS: test-file-has-acl.sh
PASS: test-set-mode-acl.sh
files
I said, I don't touch Paul Eggert's code for Solaris 10+/11 systems. But
here it is buggy: The POSIX-draft like code contains special handling for
setuid/setgid modes (because the ACL contains only the 0777 part of the mode).
But it is missing in the Solaris 10+/11 code.
I'm committing this.
On Cygwin, two files with the same ACL can have different mode:
$ ls -l tmpfile1
-rwxr-x--- 1 haible None 14 Jun 2 22:06 tmpfile1
$ getfacl.exe tmpfile1
# file: tmpfile1
# owner: haible
# group: None
user::rw-
group::r-x
mask:rwx
other:---
$ echo foo tmpfile2
$ chmod 650 tmpfile2
$ getfacl.exe
So far, the tests fail on HP-UX 11:
file_has_acl(tmpfile0) returned no, expected yes
FAIL: test-file-has-acl.sh
PASS: test-set-mode-acl.sh
files tmpfile0 and tmpfile2 have different number of ACLs: 4 and 3
FAIL: test-copy-acl.sh
The HP-UX ACL API is similar to the one of Solaris 7-9.
The acl tests still fail on AIX 5.1:
file_has_acl(tmpfile0) returned no, expected yes
FAIL: test-file-has-acl.sh
PASS: test-set-mode-acl.sh
files tmpfile0 and tmpfile2 have different access modes: 200500600 and 500600
FAIL: test-copy-acl.sh
There are two ACL APIs: one in AIX = 5.3, one
Hi all,
So far, I'm done with the 'acl' module changes. Jim, for coreutils an
appropriate NEWS entry is:
Improved support for access control lists (ACLs): On MacOS X, Solaris 7..10,
HP-UX 11, Tru64, AIX, IRIX 6.5, and Cygwin, ls -l now displays the presence
of an ACL on a file through a
Hi,
I received a report that the gnulib checking whether strcasestr works
in linear time configure test on MacOS hangs for 3 to 4 hours
before completing. Presumably this means the alarm() call in the test
doesn't work.
The version of gnuit with the problem is at:
Ian Beckwith wrote:
I received a report that the gnulib checking whether strcasestr works
in linear time configure test on MacOS hangs for 3 to 4 hours
before completing. Presumably this means the alarm() call in the test
doesn't work.
Presumably, yes...
The version of gnuit with the
On Mon, Jun 09, 2008 at 12:42:18AM +0200, Bruno Haible wrote:
On which MacOS X version was this? It configures fine and quickly for me
on MacOS X 10.3.9 and 10.5.2.
Erm, don't know, I'll ask.
In case you can map kernel versions to MacOS versions, config.log
says:
uname -v = Darwin Kernel
16 matches
Mail list logo