[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: dpkg semi-hijack - an announcement (also, triggers)



On Sun March 9 2008 16:07:58 Marc 'HE' Brockschmidt wrote:
> Mike Bird <mgb@yosemite.net> writes:
> > Ian hijacked his own program back from the people who had been blocking
> > updates for six months - including the triggers enhancement which is
> > needed for boot time improvements
>
> dpkg triggers are nice to have, but they are not the reason why we
> haven't switched to a dependency-based init system yet. Please try to
> only refer to issues you have fully understood.

Marc,

A necessary condition is not the same as a sufficient condition.

Triggers are "needed" - a necessary condition, or at least a highly
desirable condition - for efficient installation of packages in order
to avoid unnecessary repetitive global reorderings of the initscript
dependency DAG as each package is installed.

"The reason why we haven't switched yet" does not exist, as there
are several reasons not one, but if a single reason did exist it
would have been a sufficient condition, not a necessary condition.

In simple terms: "dpkg triggers" is a highly desirable precondition for
dependency-based initscripts, but so are several other preconditions,
not least of which is a substantial test period.  The issues are
well addressed in Frans' posting, the URL of which I have already
posted once in this thread and will now post again[0].

I believe it was you who wrote "Please try to only refer to issues
you have fully understood."

--Mike Bird

[0] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=465587#35


Reply to: