[tex-k] dviljk changes, committed to trunk

[Tom Rokicki]

Heh, this goes back to 1985, I suspect.  :-)

llx/lly/urx/ury are the bounding box of the included image just as it
is in the BoundingBox DSC comment, so it's PostScript points
(bp or big points in TeX nomenclature).

These *can* be floating point values, I believe, in any dvips from
the past so many years, but originally they were restricted to
integer values.

rwi and rhi are the desired width and height of the resulting figure,
in *tenths* of big points (so 720 of these to the inch).  The reason I
used the rather arbitrary factor of ten scaling is to provide some
additional precision over plain old big points, which are pretty large
and thus if we restrict ourselves to integral values (and I think
the values calculated by epsf.tex/epsf.sty are indeed so restricted)
we end up with significant scaling/offset issues.  If I was doing it
today, I would probably use a larger factor than 10 since even
1/720th of an inch is visible.

If both rwi/rhi are nonzero, the picture is scaled, possibly losing
aspect ratio.  If only one of rwi/rhi is nonzero, the picture is
scaled, preserving aspect ratio.  If both rwi/rhi are zero, I'm not
sure what happens.

I'll try to find documentation for these.  I'm kind of surprised they
are not documented.

On 7/5/07, Karl Berry <karl at freefriends.org> wrote:
>
>     I committed release 2.6p4 of dviljk to the trunk of TL's SVN
>
> Great!
>
>     PS: dviljk has a psfile special with the following syntax:
>      psfile=file llx=x0 lly=y0 urx=x1 ury=y1 rwi=s
>
> I assume it's the same as in dvips(k).  Although I've used that special
> and even messed with those parameters, I don't know of any comprehensive
> description (I just changed the numbers until my picture came out in the
> right place).
>
> In dvipsk/special.lpro, I find the following, FWIW:
>
> %   Here we handle bounding box calculations, if necessary.
> %
> /@rwi { 10 div /rwi X /rwiSeen true N } B % rwi will be real width after
> scaling
> /@rhi { 10 div /rhi X /rhiSeen true N } B % rhi will be real height after
> scaling
> /@llx { /llx X } B
> /@lly { /lly X } B
> /@urx { /urx X } B
> /@ury { /ury X } B
>
> Although dvips.texi has a `psfile special' node, it doesn't describe any
> of those keys :(.
>
> Tom ... ?
>
> Thanks,
> k
>



-- 
Check out golly at http://golly.sf.net/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://tug.org/pipermail/tex-k/attachments/20070705/86d1a3f0/attachment.html