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

Re: slink dpkg failed assertion/potato dpkg in slink

On Jul 11, Josip Rodin wrote:
> On Tue, Jul 11, 2000 at 10:41:58AM -0400, Nick Cabatoff wrote:
> > > > dpkg will most probably work fine... I doubt that dselect will compile
> > > > out-of-the-box, though :)
> > > 
> > > dselect should compile just fine on slink I think.
> > 
> > I don't use it, so I'm not concerned.  What is bothering me is that I
> > can't build dpkg just because of problems with dpkg-doc.  
> You can, you just have to skip the dpkg-doc build somehow. :)

Uh-huh.  Tell me something I don't know...

> > - In general, how do I build just a single component from a dsc that
> >   is the source to multiple packages?
> In general, there's no way. 

This strikes me as unfortunate.  Sadly, I don't see any (technical)
way to change it.  Policy is another matter, but I don't wish to
become embroiled in such matters.

> Individual packages may implement it in their debian/rules somehow.
> Try 'fakeroot debian/rules binary-dpkg' in dpkg tree, that should
> build just the dpkg package.

It should, I agree.  However, it doesn't.  The binary-* rules simply
determine which .deb files get created, but the whole shebang is built

> > - In this case, looking at dpkg's Makefile.in, debiandoc2html is
> >   called with flags '-D. -h'; I don't know what -h is supposed to do
> >   here, but clearly -D. means to work in the current directory.
> >   Unfortunately, the version in potato and woody (1.1.41) doesn't seem
> >   to support that flag, and -h produces a usage message.  Help?
> Actually, it should be "-n -D." (that's how it is in my CVS checkout),
> meaning it's an option to be passed on to nsgmls.

That makes a great deal of sense.  Wichert, should I file a bug for
this one-byte fix, or (sense I imagine you're following this thread
anyway) will you make the change without it?  

Incidently, dpkg-source unpacks the 'debian' directory with mode 2755,
which prevents dpkg-deb from executing properly.  Anyone care to
hazard a guess why it's done that way?

p.s. I got the thing built finally, and it works beautifully - thanks
     to both of you for your help.

Reply to: