On Fri, Jan 15, 2016 at 11:41:12AM -0500, I wrote:
I'm not convinced by your argument that it just doesn't matter. In
general we try really hard to perform install operations atomically
and just waving the problem away because you think something else is a
more important goal isn't
On Mon, Jan 04, 2016 at 05:54:39AM +0100, Marco d'Itri wrote:
I have not heard from you in many months: do you still have concerns
about how to fix this? Do you need help?
I'm not convinced by your argument that it just doesn't matter. In
general we try really hard to perform install
On Jan 04, Marco d'Itri wrote:
> > That's great, because locally written scripts will probably not be run
> > while installing coreutils in d-i or upgrading from Debian 8 to Debian
> > 9, so this does not look like a significant concern. The symlink will
> > always exist
On Apr 26, Marco d'Itri wrote:
> That's great, because locally written scripts will probably not be run
> while installing coreutils in d-i or upgrading from Debian 8 to Debian
> 9, so this does not look like a significant concern. The symlink will
> always exist during
On Mar 14, Michael Stone mst...@debian.org wrote:
The attached patch solves this problem by creating the link in postinst
and only if it is needed.
So this opens a window during which a path people may be using for an
essential package won't exist, right?
Right, but does it matter in
On Sun, Apr 26, 2015 at 06:27:21PM +0200, Marco d'Itri wrote:
On Mar 14, Michael Stone mst...@debian.org wrote:
The attached patch solves this problem by creating the link in postinst
and only if it is needed.
So this opens a window during which a path people may be using for an
essential
On Apr 26, Michael Stone mst...@debian.org wrote:
Right, but does it matter in practice? Debian packages are not supposed
to access touch using the full path anyway.
I'm less concerned about debian packages than locally written scripts.
Debian packages are easy to fix.
That's great, because
On Mar 14, Michael Stone mst...@debian.org wrote:
So this opens a window during which a path people may be using for an
essential package won't exist, right?
Technically, yes.
In practice, who is going to notice?
/usr/bin/touch is a compatibility symlink for a path that was in use at
least
On Sun, Nov 02, 2014 at 02:48:54AM +0100, you wrote:
The package installs a symbolic link to a file with the same name in
both /bin/ and /usr/bin/, so it makes impossible to convert a system
to the everything-in-usr directories scheme.
The attached patch solves this problem by creating the link
Package: coreutils
Version: 8.23-3
Severity: normal
User: m...@linux.it
Usertags: usrmerge
Tags: patch
The package installs a symbolic link to a file with the same name in
both /bin/ and /usr/bin/, so it makes impossible to convert a system
to the everything-in-usr directories scheme.
The
10 matches
Mail list logo