Hi,
Gabor Gombas wrote:
And that means the
package must be built with the headers from the latest kernel version it
is designed to support.
In fact, it should be built against all versions of the kernel headers
it is supposed to support. If that causes difficulties, upstream
deserves to be
On Tue, Oct 25, 2005 at 12:02:16PM -0700, Thomas Bushnell BSG wrote:
> How does the ping package know which kernel I have installed?
That's the important thing: building of the package _must not_ depend on
the version of the installed kernel. If the kernel interface changes
from one kernel vers
On Oct 25, Adam McKenna <[EMAIL PROTECTED]> wrote:
> > > >> Even if you only want to support Linux, it's *STILL* wrong to include
> > > >> the kernel headers right in the package.
> > > > It's not.
> > > What if they are *wrong* then?
> > This is not supposed to happen.
> Famous last words...
Do y
On Oct 25, Thomas Bushnell BSG <[EMAIL PROTECTED]> wrote:
> >> >> Even if you only want to support Linux, it's *STILL* wrong to include
> >> >> the kernel headers right in the package.
> >> > It's not.
> >> What if they are *wrong* then?
> > This is not supposed to happen.
> Why do you think we ha
On Tue, Oct 25, 2005 at 09:44:40PM +0200, Marco d'Itri wrote:
> On Oct 25, Thomas Bushnell BSG <[EMAIL PROTECTED]> wrote:
>
> > >> Even if you only want to support Linux, it's *STILL* wrong to include
> > >> the kernel headers right in the package.
> > > It's not.
> > What if they are *wrong* then
[EMAIL PROTECTED] (Marco d'Itri) writes:
> On Oct 25, Thomas Bushnell BSG <[EMAIL PROTECTED]> wrote:
>
>> >> Even if you only want to support Linux, it's *STILL* wrong to include
>> >> the kernel headers right in the package.
>> > It's not.
>> What if they are *wrong* then?
> This is not supposed
On Oct 25, Thomas Bushnell BSG <[EMAIL PROTECTED]> wrote:
> >> Even if you only want to support Linux, it's *STILL* wrong to include
> >> the kernel headers right in the package.
> > It's not.
> What if they are *wrong* then?
This is not supposed to happen.
> How does the ping package know which
[EMAIL PROTECTED] (Marco d'Itri) writes:
> On Oct 25, Thomas Bushnell BSG <[EMAIL PROTECTED]> wrote:
>
>> Even if you only want to support Linux, it's *STILL* wrong to include
>> the kernel headers right in the package.
>
> It's not.
What if they are *wrong* then?
How does the ping package know
On Oct 25, Thomas Bushnell BSG <[EMAIL PROTECTED]> wrote:
> Even if you only want to support Linux, it's *STILL* wrong to include
> the kernel headers right in the package.
It's not.
--
ciao,
Marco
signature.asc
Description: Digital signature
Simon Richter <[EMAIL PROTECTED]> writes:
> Thomas Bushnell BSG wrote:
>
>> Of course. So let's have one ping package that provides the feature
>> on systems where the headers say "yes, we have this feature."
>
> Unfortunately, exactly that is not going to happen because the iputils
> upstream d
On 24-Oct-05, 11:11 (CDT), Marco d'Itri <[EMAIL PROTECTED]> wrote:
> We already have one:
>
> Package: iputils-ping
> Provides: ping
What is the defined interface for the virtual package?
This is a intended as a real question, not a snipe. There was a problem
several months ago with the Nagios
On Oct 24, Simon Richter <[EMAIL PROTECTED]> wrote:
> In my opinion, it is time to make a virtual package and specify which
We already have one:
Package: iputils-ping
Provides: ping
> interfaces a ping command must support in order to provide this virtual
> package.
Feel free to work on it, if
Hi,
Thomas Bushnell BSG wrote:
Of course. So let's have one ping package that provides the feature
on systems where the headers say "yes, we have this feature."
Unfortunately, exactly that is not going to happen because the iputils
upstream does not care about compatibility and cannot be bo
Anthony Towns writes:
>> No, the point is that the Hurd developers should write that feature in
>> the standard tar so that it can be turned on and off with a normal
>> configure test,
>
> Well, why not say "the Hurd developers should write that feature in the
> standard Linux kernel so that it
On Sun, Oct 23, 2005 at 12:51:15PM -0700, Thomas Bushnell BSG wrote:
> Anthony Towns writes:
> > Going too far on the "consistency" side of things makes working on
> > alternative OSes pointless: if Debian GNU/Linux and Debian GNU/Hurd do
> > the exact same thing, why bother putting in the effort
Anthony Towns writes:
> Going too far on the "consistency" side of things makes working on
> alternative OSes pointless: if Debian GNU/Linux and Debian GNU/Hurd do
> the exact same thing, why bother putting in the effort to have both?
> Would you say that "tar" shouldn't have special options on t
On Sun, Oct 23, 2005 at 11:33:47AM -0500, Steve Greenland wrote:
> > > > > Making one of the portable versions the default ping for Debian seems
> > > > > like the
> > > > > right thing to do.
> > > > Please explain why.
> > > Consistancy.
> > Losing important features to be consistent with unrel
On 23-Oct-05, 09:42 (CDT), Marco d'Itri <[EMAIL PROTECTED]> wrote:
> On Oct 23, Stephen Frost <[EMAIL PROTECTED]> wrote:
>
> > * Marco d'Itri ([EMAIL PROTECTED]) wrote:
> > > On Oct 23, Nathanael Nerode <[EMAIL PROTECTED]> wrote:
> > >
> > > > Making one of the portable versions the default ping
On Oct 23, Stephen Frost <[EMAIL PROTECTED]> wrote:
> * Marco d'Itri ([EMAIL PROTECTED]) wrote:
> > On Oct 23, Nathanael Nerode <[EMAIL PROTECTED]> wrote:
> >
> > > Making one of the portable versions the default ping for Debian seems
> > > like the
> > > right thing to do.
> > Please explain w
* Marco d'Itri ([EMAIL PROTECTED]) wrote:
> On Oct 23, Nathanael Nerode <[EMAIL PROTECTED]> wrote:
>
> > Making one of the portable versions the default ping for Debian seems like
> > the
> > right thing to do.
> Please explain why.
Consistancy. The alternatives system could be used if someone
On Oct 23, Nathanael Nerode <[EMAIL PROTECTED]> wrote:
> Making one of the portable versions the default ping for Debian seems like
> the
> right thing to do.
Please explain why.
--
ciao,
Marco
signature.asc
Description: Digital signature
[EMAIL PROTECTED] wrote:
>How about naming the packages netkit-ping and iputils-ping,
>respectively? Oh, wait, we have that already...
>
>iputils-arping also has a more portable, alternative implementation
>(cf. package arping). That leaves us with tracepath that indeed appears
>to be Linux-only at
22 matches
Mail list logo
