Re: [9fans] eqn and unicode

Re: [9fans] eqn and unicode

[Russ Cox]

On Thu, May 28, 2009 at 4:08 AM, Gregory Pavelcak
<g.pavelcak@comcast.net> wrote:
> If you write the eqn-word for a greek letter, "GAMMA" for
> example; eqn passes the unicode character (the output of
> Alt-*G) to troff. If, on the other hand, you type Alt-*G in eqn,
> it passes `"\f2Γ\fP' to troff, thus producing, by my lights anyway,
> a nicer looking character. I was just wondering if this was
> intended as a way to give people both a roman-greek letter
> and an italic one, or if it was intended to discourage the use
> of eqn's letter names in favor of unicode, or if it just sorta
> happened. Perhaps none of the above. Anyone know?

Eqn should not generate different output for GAMMA vs Γ.
Feel free to fix it.

Russ

Re: [9fans] eqn and unicode

[erik quanstrom]

On Thu May 28 19:07:48 EDT 2009, rsc@swtch.com wrote:
> On Thu, May 28, 2009 at 4:08 AM, Gregory Pavelcak
> <g.pavelcak@comcast.net> wrote:
> > If you write the eqn-word for a greek letter, "GAMMA" for
> > example; eqn passes the unicode character (the output of
> > Alt-*G) to troff. If, on the other hand, you type Alt-*G in eqn,
> > it passes `"\f2Γ\fP' to troff, thus producing, by my lights anyway,
> > a nicer looking character. I was just wondering if this was
> > intended as a way to give people both a roman-greek letter
> > and an italic one, or if it was intended to discourage the use
> > of eqn's letter names in favor of unicode, or if it just sorta
> > happened. Perhaps none of the above. Anyone know?
>
> Eqn should not generate different output for GAMMA vs Γ.
> Feel free to fix it.

this is an interesting case.  fonts are not applied to symbols
in the resword table.  the following fix does solve the problem
with historical correctness.  if the font is ugly, perhaps a better
font is in order. :-)

- erik

diffy -c lookup.c
/n/dump/2009/0601/sys/src/cmd/eqn/lookup.c:145,150 - lookup.c:145,204
  	"PSI",		"Ψ",
  	"chi",		"χ",
  	"CHI",		"Χ",
+
+ 	/*
+ 	 * greek letters translate to themselves; this prevents
+ 	 * fonts from being applied to Γ but not GAMMA
+ 	 */
+ 	"α",	"α",
+ 	"Α",	"Α",
+ 	"β",	"β",
+ 	"Β",	"Β",
+ 	"γ",	"γ",
+ 	"Γ",	"Γ",
+ 	"δ",	"δ",
+ 	"Δ",	"Δ",
+ 	"ε",	"ε",
+ 	"Ε",	"Ε",
+ 	"ω",	"ω",
+ 	"Ω",	"Ω",
+ 	"λ",	"λ",
+ 	"Λ",	"Λ",
+ 	"μ",	"μ",
+ 	"Μ",	"Μ",
+ 	"ν",	"ν",
+ 	"Ν",	"Ν",
+ 	"θ",	"θ",
+ 	"Θ",	"Θ",
+ 	"φ",	"φ",
+ 	"Φ",	"Φ",
+ 	"π",	"π",
+ 	"Π",	"Π",
+ 	"σ",	"σ",
+ 	"Σ",	"Σ",
+ 	"ξ",	"ξ",
+ 	"Ξ",	"Ξ",
+ 	"ζ",	"ζ",
+ 	"Ζ",	"Ζ",
+ 	"ι",	"ι",
+ 	"Ι",	"Ι",
+ 	"η",	"η",
+ 	"Η",	"Η",
+ 	"κ",	"κ",
+ 	"Κ",	"Κ",
+ 	"ρ",	"ρ",
+ 	"Ρ",	"Ρ",
+ 	"τ",	"τ",
+ 	"Τ",	"Τ",
+ 	"ο",	"ο",
+ 	"Ο",	"Ο",
+ 	"υ",	"υ",
+ 	"Υ",	"Υ",
+ 	"ψ",	"ψ",
+ 	"Ψ",	"Ψ",
+ 	"χ",	"χ",
+ 	"Χ",	"Χ",
+
  	"and",		"\\f1and\\fP",
  	"for",		"\\f1for\\fP",
  	"if",		"\\f1if\\fP",

Re: [9fans] eqn and unicode

[Gregory Pavelcak]

> On Thu May 28 19:07:48 EDT 2009, rsc@swtch.com wrote:
>> On Thu, May 28, 2009 at 4:08 AM, Gregory Pavelcak
>> <g.pavelcak@comcast.net> wrote:
>> > If you write the eqn-word for a greek letter, "GAMMA" for
>> > example; eqn passes the unicode character (the output of
>> > Alt-*G) to troff. If, on the other hand, you type Alt-*G in eqn,
>> > it passes `"\f2Γ\fP' to troff, thus producing, by my lights anyway,
>> > a nicer looking character. I was just wondering if this was
>> > intended as a way to give people both a roman-greek letter
>> > and an italic one, or if it was intended to discourage the use
>> > of eqn's letter names in favor of unicode, or if it just sorta
>> > happened. Perhaps none of the above. Anyone know?
>>
>> Eqn should not generate different output for GAMMA vs Γ.
>> Feel free to fix it.
>
> this is an interesting case.  fonts are not applied to symbols
> in the resword table.  the following fix does solve the problem
> with historical correctness.  if the font is ugly, perhaps a better
> font is in order. :-)
>
> - erik

Hmm. I submitted a patch so that fonts would apply to resword
rather than making it so that they don't apply to greek letters,
and I'm not even sure I did that right.

diff -c /sys/src/cmd/eqn/text.c /usr/gp/sys/src/cmd/eqn/text.c
/sys/src/cmd/eqn/text.c:83,89 - /usr/gp/sys/src/cmd/eqn/text.c:83,90
  	else if (t == TAB)
  		p = "\\t";
  	else if ((tp = lookup(restbl, p1)) != NULL) {
- 		p = tp->cval;
+ 		sprintf(cs, "\\f%s%s\\fP", ftp->name, tp->cval);
+ 		p = cs;
  	} else {
  		lf = rf = 0;
  		lastft = 0;

Now I don't know what's right. What if, for example, you
use "gfont H"? Shouldn't that really make Γ get passed to
troff as "\fHΓ\fP"?

Greg

Re: [9fans] eqn and unicode

[erik quanstrom]

> Hmm. I submitted a patch so that fonts would apply to resword
> rather than making it so that they don't apply to greek letters,
> and I'm not even sure I did that right.
>
> diff -c /sys/src/cmd/eqn/text.c /usr/gp/sys/src/cmd/eqn/text.c
> /sys/src/cmd/eqn/text.c:83,89 - /usr/gp/sys/src/cmd/eqn/text.c:83,90
>   	else if (t == TAB)
>   		p = "\\t";
>   	else if ((tp = lookup(restbl, p1)) != NULL) {
> - 		p = tp->cval;
> + 		sprintf(cs, "\\f%s%s\\fP", ftp->name, tp->cval);
> + 		p = cs;
>   	} else {
>   		lf = rf = 0;
>   		lastft = 0;
>
> Now I don't know what's right. What if, for example, you
> use "gfont H"? Shouldn't that really make Γ get passed to
> troff as "\fHΓ\fP"?

i'm pretty sure that's wrong for reswords like "≥" or "and".
a font change would also change the appearance of existing
documents.  my change also changes the appearance of
existing documents, but only those that use the relatively-
new unicode input.

on the other hand, for Γ to be treated the same as G,
one would need a new table gktbl() that would be added
here

/sys/src/cmd/eqn/text.c:85,91
	else if ((tp = lookup(restbl, p1)) != NULL) {
		p = tp->cval;
	} else {
+ 		if (tp = lookup(gktbl, p1)) != NULL)
+ 			p = p1 = tp->cval
		lf = rf = 0;
		lastft = 0;
		nclass = NONE;	/* get started with no class == no pad */
		csp = cs;

one would also need to modifiy isalpharune to return
true for greek letters.  or just use the system isalpharune.
i'm sure one could find a number of other bugs.

the reason i didn't suggest this is because it changes old
documents.  perhaps we should just take our lumps and
fix the problem.

- erik

Re: [9fans] eqn and unicode

[Gregory Pavelcak]

erik,

You're reply and some reflection convinced me that my first
approach is just wrong. I don't understand the next-to-last
paragraph though. I tried creating gktbl and making the change
you suggest to text.c, and it seems to work. Why does
isalpharune() need changing?

Greg


> i'm pretty sure that's wrong for reswords like "≥" or "and".
> a font change would also change the appearance of existing
> documents.  my change also changes the appearance of
> existing documents, but only those that use the relatively-
> new unicode input.
>
> on the other hand, for Γ to be treated the same as G,
> one would need a new table gktbl() that would be added
> here
>
> /sys/src/cmd/eqn/text.c:85,91
> 	else if ((tp = lookup(restbl, p1)) != NULL) {
> 		p = tp->cval;
> 	} else {
> + 		if (tp = lookup(gktbl, p1)) != NULL)
> + 			p = p1 = tp->cval
> 		lf = rf = 0;
> 		lastft = 0;
> 		nclass = NONE;	/* get started with no class == no pad */
> 		csp = cs;
>
> one would also need to modifiy isalpharune to return
> true for greek letters.  or just use the system isalpharune.
> i'm sure one could find a number of other bugs.
>
> the reason i didn't suggest this is because it changes old
> documents.  perhaps we should just take our lumps and
> fix the problem.
>
> - erik

Re: [9fans] eqn and unicode

[erik quanstrom]

> erik,
>
> You're reply and some reflection convinced me that my first
> approach is just wrong. I don't understand the next-to-last
> paragraph though. I tried creating gktbl and making the change
> you suggest to text.c, and it seems to work. Why does
> isalpharune() need changing?

in trans(), there are two calls to isalpharune.  since i think
the point here is to say that greek (at least) letters are on par
with roman letters, they should get the same shimming.

there seems to be some deep magic with 'f' and 'j'.  i'd be
very intersted if anyone knows why they are exceptional.

- erik