On Sun 22 Nov 2015 23:53:11, SKINGTON wrote:
Show quoted text> On Sun Nov 22 17:53:06 2015, rwfranks@acm.org wrote:
> > On Fri Nov 20 21:02:22 2015, TODDR wrote:
> > > On Thu Nov 19 14:24:47 2015, SKINGTON wrote:
> > > > Net::DNS::Resolver::Programmable uses Net::DNS::Resolver::Base-
> > > > > make_query_packet - e.g line 194
> > >
> > > I opened RT
https://rt.cpan.org/Ticket/Display.html?id=109266
> > >
> > > It would be really good if you could recommend in the ticket what
> > > they
> > > should be doing instead?
I have done that.
Show quoted text> > >
> > > Thanks!
> >
> > The recommendation is to refrain from diving into the code and
> > (ab)using internal (undocumented) subroutines which may be changed or
> > removed without notice.
>
> Net::DNS::Resolver::Programmable has been using make_query_packet
> since at the latest 2007 (and hasn't been updated since, although that
> might be because the author reckoned they'd done all they needed to
> do.) Ray from Nominet asked for this method to be explicitly
> documented and supported, because it was useful to him, back in 2008.
> While the RT ticket
https://rt.cpan.org/Ticket/Display.html?id=37104
> ended at an impasse, there is no indication given that this
> functionality should be withdrawn - the most recent Changes for
> Net::DNS don't mention this deprecation, for instance.
>
> I appreciate the frustration in finding out that other people have
> been using what you thought was a private method. It's perfectly
> reasonable to say "Hey, you guys! Stop doing this!" in your code.
> Unfortunately (a curse in disguise?) Net::DNS is sufficiently-used
> that you probably can't get away with arbitrary stuff like this.
Understood.
That is why we warn about it first with the carp (which this time caused an unintentional bug). Note that the warning is emitted only once and not every time you use the method.
Show quoted text> I don't understand DNS very well, but at $WORK we use
> Net::DNS::Programmable for testing, and we'd prefer not to be locked
> into an old version of Net::DNS as a consequence. If we can work out a
> reasonable way for Net::DNS::Programmable to do its thing without
> meddling with things it should not wot, I'd be more than happy to
> apply to get co-maint on a useful but clearly neglected module.
That would be wonderfull!
We have another bugfix for the Net::DNS::Resolver::Programmable module:
https://rt.cpan.org/Ticket/Display.html?id=95901 , It would be very nice if you could apply that too.