Hi all I'm struggling with font embedding in fop's PostScript output - it's embedding multiple copies of fonts, just like pdf output with fop-pdf-image, and this is causing some big problems with our RIP.
As a workaround for the issues I've been having with fop-pdf-image where output PDFs land up with tens or hundreds of different embedded subsets of the same font, I've been trying to see if chucking PostScript output from fop through Adobe Acrobat Distiller would work better. In our current (non-fop-based) workflow Distiller seems to coalesce fonts so the documents that're produced contain only one copy of the font despite the input EPSs each having their own embedded subset. I was hoping it'd do the same for PostScript from fop. Unexpectedly, I get the same kind of font duplication when I process fop's PostScript through distiller as I get from fop-pdf-image and the pdf renderer. A side-by-side comparison of the PostScript from the old system and from fop shows that the embedded EPS files are copied verbatim in to the PostScript stream by both fop and the old system, so it's not doing any kind of clever transformation to the EPS data its self. Somehow, though, the old system manages to merge fonts, and fop's PostScript output doesn't. The old system uses similar-looking pre- and post- eps inclusion procedures, but they're a little simpler. They don't try to compensate for leaked stack entries like fop's seem to; they also don't mess with `userdict', strokeadjust or overprint. There's certainly no font magic in them. It's not that, though, as I get the same results if I replace fop's pre- and post-eps procedures with those used in the old system. I don't speak PostScript particularly well, and I'm beginning to really bash my head against a wall here, so I thought I'd look for some advice. Is merging fonts the norm with Distiller, or is it something magic that the old system's PostScript file is doing? If it's the norm, what might Fop's PostScript files be doing to block it? If it's not the norm, any ideas how the old files might be doing the merging magic? I can happily provide PostScript files from the old system to anyone interested. The ones I'm working on are about 60MB each, but I can produce cut-down one page ones that should be no more than 5MB compressed. I don't know if I can legally provide excerpts of just the PostScript procedures so I'm going to err on the side of providing a whole file as a complete document. If I/we can figure out what's different it'll hopefully allow font merging in fop's PostScript output, so I'd be very happy to hear from anyone else interested in the same problem. I'm still working on PDF font merging in fop-pdf-image, btw, I was just hoping this would be a faster track where I could let Distiller do the merging for me. Merging fonts in fop-pdf-image will require me to extend fop's core to allow image handlers etc to hook on events in the renderer system so it's going to take patches to fop core and therefore time. Ideas on what might be going on? Let me know if you want a copy of the PostScript files involved. -- Craig Ringer --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
