What's all this about changing xmlto to use Saxon instead of xsltproc? How will that fix PDF output?
Tim. */
Uttered Tim Waugh twaugh@redhat.com, spake thus:
What's all this about changing xmlto to use Saxon instead of xsltproc? How will that fix PDF output?
Won't fix it in itself, but were we to use FOP, which itself uses saxon, that would be one less dependancy to keep current.
Cheers
On Thu, 2005-06-16 at 17:06 +0100, Tim Waugh wrote:
What's all this about changing xmlto to use Saxon instead of xsltproc? How will that fix PDF output?
The DocBook developers mainly use a Java toolchain of Saxon and FOP, aiui.
By following the same or closely related toolchain, we benefit in many ways.
It makes sense to keep our work within the bounds of the formal DocBook and XML specifications, for all the usual reasons.
If the problem is that Saxon and FOP are not in Fedora Extras, we will get them packaged and maintained.
- Karsten
Uttered Tim Waugh twaugh@redhat.com, spake thus:
Maybe we should be looking at xmlroff, which doesn't have a dependency on *yet another* XML parser
Xmlroff just doesn't work at all yet. Try converting the IG, for instance.
At least these saxon/fop partially function.
Cheers
On Thu, Jun 16, 2005 at 05:05:43PM -0500, Tommy Reynolds wrote:
Uttered Tim Waugh twaugh@redhat.com, spake thus:
Maybe we should be looking at xmlroff, which doesn't have a dependency on *yet another* XML parser
Xmlroff just doesn't work at all yet. Try converting the IG, for instance.
At least these saxon/fop partially function.
It's unfair to say it doesn't work at all, but it certainly isn't something that will work for everything right now, that's true.
But I do really question the benefit of switching xmlto to using Saxon. We should be concentrating effort on the XSL-FO -> PDF step, not the XML -> XSL-FO step which already works perfectly well.
Tim. */
On Fri, 2005-06-17 at 09:41 +0100, Tim Waugh wrote:
On Thu, Jun 16, 2005 at 05:05:43PM -0500, Tommy Reynolds wrote:
But I do really question the benefit of switching xmlto to using Saxon. We should be concentrating effort on the XSL-FO -> PDF step, not the XML -> XSL-FO step which already works perfectly well.
Agreed in principle. However, see previous comments about keeping in step with the leaders of the DocBook community. For that reason alone, we behoove ourselves to consider Saxon. It's like using Emacs for DocBook -- it is not required, it is not essential, but it sure can make writing a -ton- easier.
- Karsten
On Thu, Jun 16, 2005 at 05:05:43PM -0500, Tommy Reynolds wrote:
Uttered Tim Waugh twaugh@redhat.com, spake thus:
Maybe we should be looking at xmlroff, which doesn't have a dependency on *yet another* XML parser
Xmlroff just doesn't work at all yet. Try converting the IG, for instance.
FWIW, I took a look at this. There is a bug concerning bare indexterms at the beginning of para blocks, and working around that I ended up with some PDF output.
The fo:external-graphic support is being re-written currently, and so there are no images. Also, there is incorrect significant white-space in a lot of places because of this sort of style:
<para> <!-- this white space is significant and *will* be rendered! --> Blah blah blah...
There are a few rendering glitches too, of course, but I am confident things would improve fairly rapidly if the Fedora Project would lend xmlroff some development support. I think it's at the point now where it is possible to pick an unimplemented property that you care about and implement it. The upstream maintainer is helpful when you have questions about the code.
I've put the PDF I got here:
http://cyberelk.net/tim/data/tmp/fedora-install-guide-en.pdf
..and now I'll go and file the bug that indexterms trigger.
Tim. */
On Fri, Jun 17, 2005 at 10:38:31AM +0100, Tim Waugh wrote:
I've put the PDF I got here:
http://cyberelk.net/tim/data/tmp/fedora-install-guide-en.pdf
Hmm, seems like no-one even looked at this.
Guess we can't be all *that* interested in getting PDF output working..
Tim. */
Uttered Tim Waugh twaugh@redhat.com, spake thus:
On Fri, Jun 17, 2005 at 10:38:31AM +0100, Tim Waugh wrote:
I've put the PDF I got here: http://cyberelk.net/tim/data/tmp/fedora-install-guide-en.pdf
Hmm, seems like no-one even looked at this.
Sorry, I've been attending some classes, meetings, travelling and the like. I lost my pencil and the dog ate my homework. (I just missed it, sorry)
Overall this looks promising. There appear to be some font metric problems and the no-pictures issue mentioned earlier.
I apologize if I sounded overly harsh about xmlroff but I still believe that it isn't ready for prime time yet, just like the FOP tool isn't.
Please continue to dog this tool so we can look at it again when these issues are resolved. I'm not married to FOP, it's just the closest thing to a working tool I've seen that isn't a commercial project.
Cheers!
-
Karsten Wade -
Tim Waugh -
Tommy Reynolds