On Aug 21, 2007, at 18:53, Bryan Blackburn wrote:
On Aug 21, 2007, at 3:48 AM, Ryan Schmidt wrote:
ImageMagick +gs requires the "gslib" component of Ghostscript. But the ghostscript port doesn't build gslib by default. So, to properly install ImageMagick +gs, you need to first install ghostscript +gslib [1]. This is cumbersome.
Is there a compelling reason why the ghostscript port shouldn't just build gslib all the time? It would seem to simplify things. It's not like it requires any additional dependencies. This goes back to the ideas in the other thread on this list recently, that ports should build the most feature-complete software they can, so long as it's not terribly inconvenient to do so. And in this case I don't see any inconvenience in doing so.
[1] That is, I assume this required. A comment that was already in the portfile when I took it over says ghostscript with gslib is required. I have not yet gotten ImageMagick +gs to work, or rather, I'm not sure how to determine if it is working.
That's why I originally set it up to depend on gplghostscript (or gnughostscript at the time, before the naming change from GNU to GPL), which does build (or at least did, last time I built it) libgs without any extra effort. Wow, that's quite a bit of parenthetical parts, hopefully it makes sense.
Yeah... you see how it's confusing!
Also, a quick look at the ghostscript Portfile suggests that +gslib actually builds a framework, not libgs, so I have no idea if IM is going to be able to use it for its ghostscript use.
I noticed that too, that it builds a framework. Again, I have no idea how to test that ImageMagick is actually attempting to use gslib. I have read, though, that ImageMagick will happily use the gs binary if it cannot find gslib. gslib is just a performance enhancement. Given that, I'm wondering if I should just remove the gs variant from ImageMagick entirely. Users who want Ghostscript support in ImageMagick can then simply install the ghostscript port.