On 04.02.2004 16:22:03 Alex wrote: > I've been looking at this for a couple of days, and figured I'd do best > to post and see what other folks think about this.
Good idea. :-) > I have to implement PPD when generating PS for the project (FOP embedded > app) that I'm working on - I need to be able to force printer tray > selection/stapling etc. > > As such, I have "butchered" the version of FOP (0.20.5) that were using > in development here and have managed to get a rough and ready version of > what we need. If anyone is interested or has had a play with area, > please feel free to drop me a line. Being the PS renderer's original author I still have PPD support on my list as a long-term goal. When I was still actively using FOP at work I took another approach however: I've patched the generated PostScript files using home-written Java classes. I also had to add things like OMR marks or embedded fonts so doing the tray selection in the same way was easy. Although I'm not really active in the development right now, I'm certainly keeping an eye on this issue. > I'd be really interested if the group as a whole had any thoughts/ideas > which I could look at implementing and/or contributing back to the > project (if you'll take 'em of course!). Contributions are always welcome. Just keep in mind that patches for FOP 0.20.5 (or the maintenance branch) will have a very high probability of not making it into our codebase since that branch is frozen. We're concentrating on the new codebase (HEAD, redesign). It shouldn't be a big problem to avoid writing all of the code twice if you need PPD support for 0.20.5 (as the 1.0 release is still a few months in the future). It's important to have a good separate PPD parser and then all you need is to include the right sequences in the right places. Well, the latter won't be so easy in the end since you have to have some kind of communication from the XSL-FO file (using custom tags or attributes) to the renderer to map page masters to a tray etc. There's probably an easier way if you provide the mapping information directly to the renderer in Java code but in the long term I believe the mapping should be done in XSL-FO. Anyway, as a long-term investment it would be cool if you wrote the code for the HEAD in parallel with the one for 0.20.5. I'll assist you with (hopefully) good advice and maybe more, as time permits. Jeremias Maerki