On Thursday 28 October 2010 19:04:46 Larry Kollar wrote:
> As a follow-up, I think I may have finally identified the problem.
> The syntax for my XREF macro is:
>
> .XREF type dest [trailing]
>
> And (depending on the type) the resulting command looks something
> like:
>
> .pdfhref L -D \\$2 -A \\$3 {stored xref text}
>
> All the dropped links have no third argument.
This means that all content of {stored xref text}, up to the first
unquoted space, will be interpreted as affixed text, (the mandatory
argument to "-A"), which is to be placed after the hot-spot end marker;
if there is no unquoted space within that stored text, then there will
be nothing to separate the beginning and ending markers for the
hot-spot. I did wonder, when I read your original message...
> > In my continued attempts to get PDF links working properly, I'm
> > now running into one of the most frustrating aspects of any
> > troubleshooting job: the intermittent error. Cross-reference links
> > that work in one place don't work in the other, and I haven't yet
> > found a pattern. Looking at the reference file, the non-working
> > links correspond to "pdfhref Z" pairs like this:
> >
> > .pdfhref Z 252 475260 541000
> > .pdfhref Z 252 475260 541000
> >
> > In other words, they begin and end at the same point.
...if you may have been invoking ".pdfhref L ..." with no content
defined for output in the hot-spot region.
> Setting the trailing text to a \& (zero-width space) at least works
> around this issue.
I think I would code the xref macro itself to guard against the
possibility of an empty \$3, e.g. such that it does:
.pdfhref L -D \\$1 -A \&\\$3 {stored xref text}
> All that's left to do is get the ToC/Index page references working.
I'd be interested in how you get on with this. I made a tentative, but
unsuccessful, attempt to get ToC references working in conjunction with
the ms .XS, .XA, .XE, .TC mechanism. It wasn't a priority requirement
for me, at the time, and I've never found time to get back to it; I
suspect that it may be more successful with a more proactive approach
to ToC generation, (which doesn't rely on delayed collation).
--
Regards,
Keith.
