Santiago vs. common sense (was Re: Debian Weekly News - August 10th, 1999)
Santiago Vila <sanvila@unex.es> writes:
> If I'm not mistaken, debmake 3.6 was uploaded to follow current policy,
> before the Project Leader's request.
>
> Even after the request, I don't think it is a good idea to change debmake
> again without a corresponding change in policy. There may be people who
> feel uncomfortable by breaking policy and there may also be people who
> feel uncomfortable by not following the Project Leader's request. For this
> reason, I consider a good thing that debmake users, not debmake
> maintainer, have the choice about this, by using either debmake
> from slink, or debmake from potato.
That's absurd. Consider that any debmake binary arch-dependant
package to be uploaded will be compiled 5 additional times by robots.
Your suggestion to use slink debmake is simply not feasible.
> In summary: Regardless of the Project Leader's request, I would not
> like to *force* people to *violate* current policy.
That would be all very respectible, except that you are asking folks
to sacrifice common sense in the name of blind aherence to Policy.
Come'on, Santiago. Are you trying to add more and move evidence to
the folks who say that debmake should never be used?
> I think policy deserves a great degree of respect also, so please don't
> talk about disrespect.
If that is the case, please explain the following quote from
/usr/bin/debstd:
# etc files that could need some tweaking
if [ -f $3$i ]; then
FILEX="$i"
echo ===========================================================
echo WARNING: This package tries to modify /etc/$FILEX.
echo This might be not ok according to policy.
echo ===========================================================
cat ${debmake_dir}/etc.first.postinst $3$i \
${debmake_dir}/etc.last.postinst | addscript postinst
And from /usr/share/debmake/etc.first.postinst:
# etc file modification generated by debmake #DATE#
if ! grep -q '#-- #PACKAGE# begin' /etc/#FILE#; then
cat >>/etc/#FILE# <<EOF
#-- #PACKAGE# begin
Sure, you emit a warning, but this a brutal, blantant violation of
policy. Policy doesn't mean "should not modify other conffiles", it
means "must not".
So I find it absurd that you should cling to your fidelity to Policy,
even when mistakes are made (IMHO) in Policy (namely, telling folks to
trasnition to /usr/share/doc without really working out how this would
affect users).
How can you adopt the "sacred cloak" of policy yet you've obviously
never even reviewed your flagship packge, debmake, for Policy
compliance?
--
.....Adam Di Carlo....adam@onShore.com.....<URL:http://www.onShore.com/>
Reply to: