Bruno Haible [EMAIL PROTECTED] wrote:
I'm applying this.
2006-10-02 Bruno Haible [EMAIL PROTECTED]
* gnulib-tool (func_emit_lib_Makefile_am): Don't add no-dependencies
to the AUTOMAKE_OPTIONS.
Thanks.
With that, I've just reverted the work-around in coreutils' bootstrap.
URL:
http://savannah.gnu.org/bugs/?17900
Summary: rm -f should ignore ENOTDIR
Project: GNU Core Utilities
Submitted by: schwab
Submitted on: Montag 02.10.2006 um 16:35
Category: None
Severity: 3 - Normal
Hello,
When root is copying a non-privileged user file with the bit suid on,
the file gets chowned to root and bit suid is not removed.
That could mean a security compromise.
Example:
[EMAIL PROTECTED] ~]$ cat c.c
/* Suid shell || uid=0(root) gid=0(root) || Lo compilas y ocultas
* en la
Hmm i was checking the source code ... and seems is the kernel syscall
who remove the suid, when you make a chown ...
2006/10/2, Pedro Andujar [EMAIL PROTECTED]:
Hello,
When root is copying a non-privileged user file with the bit suid on,
the file gets chowned to root and bit suid is not
It would be nice if gnu chmod had an option to set two sets of
permissions -- one for files and one for directories. Or perhaps one
for directories and one for {everything else}?
For example, if you want to standardize the set of permissions in a
large directory tree, you could do this --
On Mon, 2 Oct 2006, Doug McLaren wrote:
Right now, the best way to do what I'm referring to is something like
this --
find /directory -type d -print0 | xargs --no-run-if-empty -0 chmod 755
find /directory '!' -type d '!' -type l -print0 | \
xargs --no-run-if-empty -0 chmod 644
If
URL:
http://savannah.gnu.org/bugs/?17903
Summary: cp/mv only read/write 512 byte blocks when
filesystem blksize 4MB
Project: GNU Core Utilities
Submitted by: tonyernst
Submitted on: Monday 10/02/2006 at 20:29
Category:
Follow-up Comment #1, bug #17903 (project coreutils):
I am attaching a patch against CVS that reverses the change (described above)
that caused this problem.
If you don't want to totally remove it, perhaps it could be special cased
somehow to only use the 4MB hack on HP-UX builds?
coreutils-6.3 includes a patch, c99-to-c89.diff, to allow it to be
compiled on platforms that don't support declarations after statements.
However, this patch is not sufficent -- I had to fix two other areas to
get coreutils to compile under GCC 2.95.3.
I've appended my fixes. They are to be
When testing coreutils-6.3, I observed a make check failure which I have
judged to be spurious.
The test was ls/stat_dtype, which verifies that `ls' is using the
dirent.d_type feature -- which is not provided by all kernels and
filesystems. While the test does try to skip systems that don't
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
According to Jim Meyering on 10/2/2006 11:19 AM:
This error should be ignored. After all, if a file name component is not a
directory the file cannot exist, so rm -f should handle it the same as
ENOENT.
Thanks for reporting that.
Here's the
Eric Blake [EMAIL PROTECTED] wrote:
According to Jim Meyering on 10/2/2006 11:19 AM:
This error should be ignored. After all, if a file name component is not a
directory the file cannot exist, so rm -f should handle it the same as
ENOENT.
Thanks for reporting that.
Here's the patch I'll
12 matches
Mail list logo
