Re: [git 17e68b85] make error in documentation

[Patrick McCarty]

On 2009-08-20, Graham Percival wrote:
> On Tue, Aug 18, 2009 at 10:33:11AM -0700, Patrick McCarty wrote:
> > On 2009-08-18, John Mandereau wrote:
> > > Le mardi 18 août 2009 à 03:55 -0700, Graham Percival a écrit :
> > > > If necessary we could use inkscape or rsvg, but I'd rather not
> > > > increase the build requirements.  (IIRC inkscape isn't required
> > > > for master)
> > > 
> > > Inkscape is required by input/regression/backend-svg.ly, which
> > > had been disabled when some version of Inkscape (provided by
> > > Fedora 8 or 9 IIRC) could no longer be run without user
> > > interaction; Fedora 11 provides 0.47-0.8.20090508svn, which can
> > > run in batch mode again, are there objections to reenable that
> > > regression test again?
> > 
> > It is a pretty large build requirement for just one regest, but I
> > wouldn't mind if we required Inkscape for "make doc".  But I think
> > Graham's opinion should count more in this case.  :-)
> >
> > I hope that the current "convert" command in the
> > Documentation/pictures makefile works for everyone, BTW.
> 
> As far as the docs go, I think that convert works fine (pending
> any more information from Werner).  As a general rule, if our
> build requirements are covered in debian stable, then we know that
> we're not pushing the bleeding edge.  :)
> 
> If inkscape would be useful for the regtests, then by all means
> use it.  You're the SVG expert, so I'm not going to argue with you
> about that.

Looking at that regtest again, I don't think we really need to
uncomment it yet.  Plus, since Inkscape 0.46 is broken wrt the
"currentColor" value, the regtest output does not look very pleasant
right now.  :-)

I eventually want to create a regtest for the "embedded-svg" stencil
expression; when that time comes, we could then uncomment
backend-svg.ly and use Inkscape for both these regtests (and maybe
others).  By then, the Inkscape problem will be fixed too.


Thanks,
Patrick