On 18.02.2008 02:57:51 The Web Maestro wrote:
> On Feb 17, 2008 10:56 AM, Max Berger <[EMAIL PROTECTED]> wrote:
> > Dear FOP Devs,
> >
> > Am 16.02.2008 um 16:54 schrieb The Web Maestro:
> > > I would think the default should be to continue (warning in
> > > LOG/stdout) create an empty (blank/transparent) container the size and
> > > placement of the image.
> >
> > +1

Just to compare with Batik: Batik uses a "broken image" (painted using
Java2D) if an image cannot be found:

What should be the intrinsic size of a missing image? 1x1cm? 2x2in?

> > > It would be nifty if it could be a flag in the config or CLI args
> > > giving the user the choiice to fail on missing images.
> >
> > I think this would be a good "test case" for the new feedback
> > mechanism. IMO there should be a warning by default, but the user
> > should be able to configure it to fail if needed.

I haven't considered providing configuration settings to override
default severity levels, yet. I initially thought that would be left to
the integrator. But wiring something like that into FOP should be easy.
You could simply provide a list of event IDs and the overriding severity

> > > Clay
> >
> > Max Berger
> Even better... In addition to a config setting, another option would
> be to place it in the xsl-fo file itself (I guess in the fox:
> namespace since it's probably not in the XSL-FO spec...).

That's also a possibility but if it's better.... The config approach
lets you specify the behaviour in one place. The extension attribute
approach makes it necessary to provide a setting for every single image
even though the expected behaviour is probably always the same within a
system. And this mechanism only works for missing images. The config
approach lets you specify the behaviour for all events.

Jeremias Maerki

Reply via email to