Re: Raster3D (and gff2aplot too)
On Fri, April 22, 2005 13:36, Andreas Tille wrote:
> OK, this is a reasonable fallback because the user is informaed what
> happens. The question is whether it might really make sense to patch
> upstream software only for the sake of using Recommends / Suggests instead
> of Depends. It might be error prone and you win nearly nothing (as I said:
> ImageMagick and ghostscript
> are not so extraordinary packages on image processing machines that it
> would make any sense to leave them out.
That's indeed true - I was reading your statement as a sort of general
rule, but surely in specific situations it can be better to depend on
rather common software rather than to jump through hoops to support it as
something optional. I think patching for a graceful error (which involves
about 3 lines in the PHP-PEAR example) can be worthwile when it prevents
you from dragging in the whole PEAR circus when you are not even going to
use it.
Regards,
Thijs
Reply to: