Re: eqn formatting issues with grops and gropdf

[Deri]

I have managed to rescue my second message which was truncated!!

On Tuesday, 26 July 2022 17:19:04 BST Deri wrote:
> On Tuesday, 26 July 2022 09:00:25 BST joerg van den hoff wrote:
> > me again with an update/correction to the previous description of the
> > issue
> > (the described problem remains, though):
> > 
> > 1.
> > regarding the symobl fonts used by grops and gropdf I previously stated
> > the
> > former were using SS (symbols slanted) and the latter S (symbols) which I
> > presumed according to the looks of the greek letters in the ps output
> > (upright) and pdf output (slanted to the right like italics). this was
> > *wrong*. looking into the font information in the formatted files it was
> > the other way around (grops was using SS and gropdf using S).
> > 
> > looking into the DESC files, I do find indeed entries
> > 
> > grops:  fonts 9 0 0 0 0 0 SS S ZD ZDR
> > gropdf: fonts 9 0 0 0 0 0 0 S ZD ZDR
> > 
> > which explains the font selection that occurred. I do not understand,
> > however, while this ultimately lead to _slanted_ glyphs with gropdf and
> > _upright_ glyphs with grops (exactly the other way around as I would have
> > expected for S vs SS).
> > 
> > 2.
> > forcing grops to also use S (by editing the DESC file and removing SS from
> > the entry) leads to sane ps and pdf output with both devices (no
> > misalignment and strange irregular widths of the greek letters). so this
> > would be the quick patch to "repair" grops: change the DESC file.
> > 
> > 3.
> > using now the same font S, the glyphs produced by grops are upright
> > (expected) and those produced by gropdf are slanted (unexpected). why is
> > that??
> > 
> > the main observation remains unaltered: in standard setup grops uses SS
> > for
> > typesetting greek letters since SS is found before S according to DESC and
> > this leads to rather massive typesetting errors in equations using
> > possibly
> > many greek letters: cumulative mispositioning of stuff later on the same
> > line.
> > 
> > what do to about this?
> > 
> > thank you
> > joerg
> 
> Hi Joerg,
> 
> You are correct that gropdf does not include the SS font. The reason is
> because it is not a proper font, it is instead a postscript program, which,
> when run by a postscript interpreter such as ghostscript or a postscript
> printer, generates a slanted version of the symbol font. This is not valid
> as a pdf font.
> 
> The SS font and the S font both define *a but only S defines *A so when they
> are both loaded with .special SS S the lower case is found in SS but
> uppercase in S. Since gropdf does not have SS *a is found in S and a
> special command is sent to gropdf "x Slant 16" which tells it to slant the
> glyph by 16 degrees.
> 
> If you type:-
> 
> echo "\[*a]" | groff -Z
> 
> You will see:-
> 
> x T ps
> x res 72000 1 1
> x init
> p1
> x font 11 S
> f11
> s10000
> V12000
> H72000
> md
> DFd
> C*a
> h6310
> n12000 0
> x trailer
> V792000
> x stop
> 
> But if you type:-
> 
> echo "\[*a]" | groff -Tpdf -Z
> 
> It changes to:-
> 
> x T pdf
> x res 72000 1 1
> x init
> p1
> x font 11 S
> f11
> s10000
> x Slant 16
> V12000
> H72000
> md
> DFd
> C*a
> h6310


This is a bit of a disappointment, 80% of my email is missing, my fault I'm 
sure. So this will be the short version!

The reason the pdf version slants the S characters when outputting lowercase 
greek is because of this code in pdf.tmac:-

.de pdf:SS
.    char \\$1 \\S'16'\\$1\\S'0'
..
.pdf:SS \[+h]
.pdf:SS \[ts]
.pdf:SS \[*a]
.pdf:SS \[*b]

Which slants each character when you use it. The reason grops outputs 
lowercase greek is because it searches in SS before S which has definitions 
for lowercase but not uppercase. The S font has definitions for both cases, so 
when you prevent grops from using SS you end up with non-italicised glyphs.

Now to look at why the character spacing is different. Italic fonts have extra 
metrics which determine how much space to add for italic correction (space to 
add when italic and normal characters are directly adjacent). SS has these 
extra metrics, S does not. You tell groff to add this space by placing \, or 
\/ between the two characters. When eqn is processing your equation it 
encapsulates each greek character with these escapes. Which causes the extra 
space to appear. I suspect it does this because usually these greek characters 
appear on their own in formula, rather than a group together.

So the solution, if you want the slanted characters you see from gropdf, but 
using grops. Is to prevent grops using SS, as you have done, and include the 
code you find in pdf.tmac, which slants the characters, into your own file. 
The "x Slant" command is understood by grops as well as gropdf so it should 
work.

Here's your test file.


.\"====================cut=================
.\"groff -e -ms tt.trf > tt1.ps         -- broken output (output from ps2pdf 
tt1.ps, too, of course)
.\"groff -e -ms -Tpdf tt.trf > tt2.pdf  -- looks good.
.
\"----------------------------------------------------------------------------------------------
.
.\" Use S only not SS (don't have to modify DESC)
.special S
.de pdf:SS
.    char \\$1 \\S'16'\\$1\\S'0'\^
..
.pdf:SS \[+h]
.pdf:SS \[ts]
.pdf:SS \[*a]
.pdf:SS \[*b]
.pdf:SS \[*x]
.pdf:SS \[*d]
.pdf:SS \[*e]
.pdf:SS \[*f]
.pdf:SS \[*g]
.pdf:SS \[*y]
.pdf:SS \[*i]
.pdf:SS \[+f]
.pdf:SS \[*k]
.pdf:SS \[*l]
.pdf:SS \[*m]
.pdf:SS \[*n]
.pdf:SS \[*o]
.pdf:SS \[*p]
.pdf:SS \[*h]
.pdf:SS \[*r]
.pdf:SS \[*s]
.pdf:SS \[*t]
.pdf:SS \[*u]
.pdf:SS \[+p]
.pdf:SS \[*w]
.pdf:SS \[*c]
.pdf:SS \[*q]
.pdf:SS \[*z]
.LP
.EQ (1)
alpha beta gamma delta rho sigma 1 over 2
.EN
.LP
.EQ (2)
a b c d r s 1 over 2
.EN
.\"====================cut=================

I have added a tiny amount of space between the characters. If you don't like, 
remove the \^ from line 9.


Cheers 

Deri