Re: Breaking programs because a not yet implemented solution exists in theory (Was: Bug#658139: evince: missing mime entry)
Josselin Mouette writes ("Re: Breaking programs because a not yet implemented solution exists in theory (Was: Bug#658139: evince: missing mime entry)"):
> Le jeudi 02 février 2012 à 15:51 +0100, Jonas Smedegaard a écrit :
> > The "issue" here (or at least at the dawn of this thread) is that some
> > package maintainers have chosen to break something that used to work.
>
> Claiming that the mime-support system actually works is stretching
> reality as much as a bone who just met Chuck Norris’ fist. The policy
> might claim it is the recommended way, but only a handful of programs,
> such as mutt and lynx, actually make use of this information.
The correct approach if you don't like the existing machinery, and
think policy is mandating an obsolete mechanism, is to get consensus
to change the approach for Debian as a whole.
The correct approach it is not to unilaterally decide to do switch to
some other half-implemented system, remove support for the previously
working machinery, and demand that bug submitters write the
compatibility code.
Ian.
Reply to: