tech@mandoc.bsd.lv
 help / color / mirror / Atom feed
* eqn(7) "delim" is undocumented (maybe)
@ 2019-12-08  8:42 Anthony J. Bentley
  2020-01-07 18:32 ` Ingo Schwarze
  2020-01-08 12:21 ` Ingo Schwarze
  0 siblings, 2 replies; 4+ messages in thread
From: Anthony J. Bentley @ 2019-12-08  8:42 UTC (permalink / raw)
  To: tech

Hi,

I can't find anyplace that documents eqn(7)'s "delim" statement (even
groff's eqn(1) doesn't really). This is widely used in the X11 manuals
and others for inline math. For example, glTexGenfv(3):

 .EQ
 delim $$
 .EN
 ...
 This function can be used, for example, to texture-map terrain using sea level
 as a reference plane
 (defined by $p sub 1$, $p sub 2$, $p sub 3$, and $p sub 4$).

I gave a shot at documenting it in the diff below... of course, open to
improvements in both correctness and clarity.

As an aside, I expected these two examples to render the same, but they
don't. Should they look the same, or am I misunderstanding how eqn(7)
works?

.\" "delim off sin(<pi>)"
.EQ
delim $$
delim off
.EN
$sin { pi }$

.\" "$sin { pi }$"
.EQ
delim $$
.EN
.EQ
delim off
.EN
$sin { pi }$


Index: eqn.7
===================================================================
RCS file: /cvs/src/share/man/man7/eqn.7,v
retrieving revision 1.11
diff -u -p -r1.11 eqn.7
--- eqn.7	23 Apr 2019 17:52:12 -0000	1.11
+++ eqn.7	8 Dec 2019 08:41:25 -0000
@@ -67,6 +67,10 @@ these are not the same as
 macros, and may only be invoked as
 .Sq \&.EQ .
 .Pp
+Inline equations are delimited by characters set by a
+.Cm delim
+statement.
+.Pp
 The equation grammar is as follows, where quoted strings are
 case-sensitive literals in the input:
 .Bd -literal -offset indent
@@ -178,6 +182,23 @@ statement is a synonym for
 while
 .Cm tdefine
 is discarded.
+.It Cm delim
+Set the two characters used to delimit inline equations.
+Its syntax is as follows:
+.Pp
+.D1 Cm delim Ar param
+.Pp
+.Ar param
+is a two-character string,
+the first and second character of which delimit the start and end of inline
+equations throughout the document.
+Alternately,
+.Ar param
+may be the special string
+.Cm off ,
+which disables any previously declared delimiters, or
+.Cm on ,
+which reenables previously defined delimiters that have been turned off.
 .It Cm gfont
 Set the default font of subsequent output.
 Its syntax is as follows:
--
 To unsubscribe send an email to tech+unsubscribe@mandoc.bsd.lv

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: eqn(7) "delim" is undocumented (maybe)
  2019-12-08  8:42 eqn(7) "delim" is undocumented (maybe) Anthony J. Bentley
@ 2020-01-07 18:32 ` Ingo Schwarze
  2020-01-10 10:53   ` Anthony J. Bentley
  2020-01-08 12:21 ` Ingo Schwarze
  1 sibling, 1 reply; 4+ messages in thread
From: Ingo Schwarze @ 2020-01-07 18:32 UTC (permalink / raw)
  To: Anthony J. Bentley; +Cc: tech

Hi Anthony,

Anthony J. Bentley wrote on Sun, Dec 08, 2019 at 01:42:37AM -0700:

> I can't find anyplace that documents eqn(7)'s "delim" statement

Brian W. Kernighan and Lorinda L. Cherry:
Typesetting Mathematics - User's Guide (Second Edition)
August 15, 1978
page 8, chapter 19: Shorthand for In-line Equations

> (even groff's eqn(1) doesn't really).

I think you have a point.

> This is widely used in the X11 manuals
> and others for inline math.

Definitely.

> I gave a shot at documenting it in the diff below... of course, open to
> improvements in both correctness and clarity.
> 
> As an aside, I expected these two examples to render the same, but they
> don't. Should they look the same, or am I misunderstanding how eqn(7)
> works?
> 
> .\" "delim off sin(<pi>)"
> .EQ
> delim $$
> delim off
> .EN
> $sin { pi }$

On first sight, that looks like a bug in mandoc.
I need to have a closer look.


Regarding your documentation patch, i'd like to do it slightly
differently.

 - Let's avoid the term "inline equation".  Conventionally, an
   equation is called "inline" when it appears in the middle of
   an *output* line.  The opposite is a displayed equation.
   But here, we are talking about an equation in the middle of
   a (text) *input* line.
   In mdoc(7) and man(7), .EQ/.EN vs. $$ has nothing to do with
   inline vs. displayed.  Both input forms create inline equations
   by default.  Both forms can be wrapped in display macros, for
   example .Bd -unfilled, to achieve a display.
   Admittelly, Kernighan/Cherry also talked about "in-line equations".
   But that happened only because they focussed on the context of
   the -ms macros, and for -ms, .EQ/.EN does indeed generate a
   display by default.  However, even in that context, they were
   careful enough to not call $$ an "in-line equation" outright,
   but rather a "shorthand for in-line equations".

 - The beginning of the eqn(7) manual page is a mess.  The second
   and third paragraphs say about the same, and the non-standard
   .Sh macro between them is completely useless.  So let's delete
   the section header, combine the two paragraphs, and add the
   new information to it.

 - I don't see the point in ".D1 Cm delim Ar param".
   It says nothing new: the fact the "delim" needs to take an
   argument is already obvious from the verb "set".

 - But i would like to stress the conventional form "delim $$"
   by giving an explicit example.

OK?
  Ingo


Index: eqn.7
===================================================================
RCS file: /cvs/src/share/man/man7/eqn.7,v
retrieving revision 1.11
diff -u -p -r1.11 eqn.7
--- eqn.7	23 Apr 2019 17:52:12 -0000	1.11
+++ eqn.7	7 Jan 2020 18:27:00 -0000
@@ -44,28 +44,16 @@ specification (see
 .Sx SEE ALSO
 for references).
 .Pp
-Equations within
-.Xr mdoc 7
-or
-.Xr man 7
-documents are enclosed by the standalone
-.Sq \&.EQ
-and
-.Sq \&.EN
-tags.
-Equations are multi-line blocks consisting of formulas and control
-statements.
-.Sh EQUATION STRUCTURE
-Each equation is bracketed by
-.Sq \&.EQ
-and
-.Sq \&.EN
-strings.
-.Em Note :
-these are not the same as
-.Xr roff 7
-macros, and may only be invoked as
-.Sq \&.EQ .
+An equation starts with an input line containing exactly the characters
+.Sq \&.EQ ,
+may contain multiple input lines, and ends with an input line
+containing exactly the characters
+.Sq \&.EN .
+Equivalently, an equation can be given in the middle of a single
+text input line by surrounding it with the equation delimiters
+defined with the
+.Cm delim
+statement.
 .Pp
 The equation grammar is as follows, where quoted strings are
 case-sensitive literals in the input:
@@ -178,6 +166,25 @@ statement is a synonym for
 while
 .Cm tdefine
 is discarded.
+.It Cm delim
+This statement takes a string argument consisting of two bytes,
+to be used as the opening and closing delimiters for equations
+in the middle of text input lines.
+Conventionally, the dollar sign is used for both delimiters,
+as follows:
+.Bd -literal -offset indent
+\&.EQ
+delim $$
+\&.EN
+An equation like $sin pi = 0$ can now be entered
+in the middle of a text input line.
+.Ed
+.Pp
+The special statement
+.Cm delim off
+temporarily disables previously declared delimiters and
+.Cm delim on
+reenables them.
 .It Cm gfont
 Set the default font of subsequent output.
 Its syntax is as follows:
--
 To unsubscribe send an email to tech+unsubscribe@mandoc.bsd.lv

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: eqn(7) "delim" is undocumented (maybe)
  2019-12-08  8:42 eqn(7) "delim" is undocumented (maybe) Anthony J. Bentley
  2020-01-07 18:32 ` Ingo Schwarze
@ 2020-01-08 12:21 ` Ingo Schwarze
  1 sibling, 0 replies; 4+ messages in thread
From: Ingo Schwarze @ 2020-01-08 12:21 UTC (permalink / raw)
  To: Anthony J. Bentley; +Cc: tech

Hi Anthony,

Anthony J. Bentley wrote on Sun, Dec 08, 2019 at 01:42:37AM -0700:

> As an aside, I expected these two examples to render the same, but they
> don't. Should they look the same, or am I misunderstanding how eqn(7)
> works?
> 
> .\" "delim off sin(<pi>)"
> .EQ
> delim $$
> delim off
> .EN
> $sin { pi }$

This was is indeed misparsed and consequently misrendered
by mandoc(1); fixed by the following commit.

> .\" "$sin { pi }$"
> .EQ
> delim $$
> .EN
> .EQ
> delim off
> .EN
> $sin { pi }$

Thanks for reporting,
  Ingo


Log Message:
-----------
Skip whitespace before tokens, too.
Bug found by bentley@ with input like "delim $$ delim off".

Modified Files:
--------------
    mandoc:
        eqn.c
    mandoc/regress/eqn/delim:
        Makefile
        basic.in
        basic.out_ascii

Added Files:
-----------
    mandoc/regress/eqn/delim:
        basic.out_utf8

Revision Data
-------------
Index: eqn.c
===================================================================
RCS file: /home/cvs/mandoc/mandoc/eqn.c,v
retrieving revision 1.83
retrieving revision 1.84
diff -Leqn.c -Leqn.c -u -p -r1.83 -r1.84
--- eqn.c
+++ eqn.c
@@ -1,7 +1,7 @@
 /*	$Id$ */
 /*
  * Copyright (c) 2011, 2014 Kristaps Dzonsons <kristaps@bsd.lv>
- * Copyright (c) 2014, 2015, 2017, 2018 Ingo Schwarze <schwarze@openbsd.org>
+ * Copyright (c) 2014,2015,2017,2018,2020 Ingo Schwarze <schwarze@openbsd.org>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
@@ -399,6 +399,14 @@ eqn_next(struct eqn_node *ep, enum parse
 		case '"':
 			quoted = 1;
 			break;
+		case ' ':
+		case '\t':
+		case '~':
+		case '^':
+			if (quoted)
+				break;
+			ep->start++;
+			continue;
 		default:
 			break;
 		}
@@ -669,7 +677,7 @@ eqn_parse(struct eqn_node *ep)
 	if (ep->data == NULL)
 		return;
 
-	ep->start = ep->end = ep->data + strspn(ep->data, " ^~");
+	ep->start = ep->end = ep->data;
 
 next_tok:
 	tok = eqn_next(ep, MODE_TOK);
Index: Makefile
===================================================================
RCS file: /home/cvs/mandoc/mandoc/regress/eqn/delim/Makefile,v
retrieving revision 1.2
retrieving revision 1.3
diff -Lregress/eqn/delim/Makefile -Lregress/eqn/delim/Makefile -u -p -r1.2 -r1.3
--- regress/eqn/delim/Makefile
+++ regress/eqn/delim/Makefile
@@ -1,6 +1,7 @@
-# $OpenBSD: Makefile,v 1.2 2020/01/08 10:50:54 schwarze Exp $
+# $OpenBSD: Makefile,v 1.3 2020/01/08 12:09:14 schwarze Exp $
 
 REGRESS_TARGETS	 = basic
+UTF8_TARGETS	 = basic
 GOPTS		 = -e
 SKIP_GROFF	 =
 
--- /dev/null
+++ regress/eqn/delim/basic.out_utf8
@@ -0,0 +1,10 @@
+DELIM-BASIC(1)              General Commands Manual             DELIM-BASIC(1)
+
+N\bNA\bAM\bME\bE
+     d\bde\bel\bli\bim\bm-\b-b\bba\bas\bsi\bic\bc – inline eqn delimiters
+
+D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
+     initial text α inline β γ inline [delta] ε inline ζ inline $eta$ inline θ
+     final text
+
+OpenBSD                         January 8, 2020                        OpenBSD
Index: basic.out_ascii
===================================================================
RCS file: /home/cvs/mandoc/mandoc/regress/eqn/delim/basic.out_ascii,v
retrieving revision 1.4
retrieving revision 1.5
diff -Lregress/eqn/delim/basic.out_ascii -Lregress/eqn/delim/basic.out_ascii -u -p -r1.4 -r1.5
--- regress/eqn/delim/basic.out_ascii
+++ regress/eqn/delim/basic.out_ascii
@@ -5,6 +5,6 @@ N\bNA\bAM\bME\bE
 
 D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      initial text <alpha> inline <beta> <gamma> inline [delta] <epsilon>
-     inline <zeta> final text
+     inline <zeta> inline $eta$ inline <theta> final text
 
 OpenBSD                         January 8, 2020                        OpenBSD
Index: basic.in
===================================================================
RCS file: /home/cvs/mandoc/mandoc/regress/eqn/delim/basic.in,v
retrieving revision 1.3
retrieving revision 1.4
diff -Lregress/eqn/delim/basic.in -Lregress/eqn/delim/basic.in -u -p -r1.3 -r1.4
--- regress/eqn/delim/basic.in
+++ regress/eqn/delim/basic.in
@@ -1,4 +1,4 @@
-.\" $OpenBSD: basic.in,v 1.3 2020/01/08 10:58:09 schwarze Exp $
+.\" $OpenBSD: basic.in,v 1.4 2020/01/08 12:09:14 schwarze Exp $
 .Dd $Mdocdate$
 .Dt DELIM-BASIC 1
 .Os
@@ -19,4 +19,13 @@ inline [delta]
 delim onepsilon
 .EN
 inline [zeta]
+.EQ
+delim $$
+delim off
+.EN
+inline $eta$
+.EQ
+delim on
+.EN
+inline $theta$
 final text
--
 To unsubscribe send an email to tech+unsubscribe@mandoc.bsd.lv

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: eqn(7) "delim" is undocumented (maybe)
  2020-01-07 18:32 ` Ingo Schwarze
@ 2020-01-10 10:53   ` Anthony J. Bentley
  0 siblings, 0 replies; 4+ messages in thread
From: Anthony J. Bentley @ 2020-01-10 10:53 UTC (permalink / raw)
  To: Ingo Schwarze; +Cc: tech

Hi Ingo,

Ingo Schwarze writes:
> Regarding your documentation patch, i'd like to do it slightly
> differently.

Your points make sense, and your patch is ok with me.

-- 
Anthony J. Bentley
--
 To unsubscribe send an email to tech+unsubscribe@mandoc.bsd.lv

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2020-01-10 10:53 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-12-08  8:42 eqn(7) "delim" is undocumented (maybe) Anthony J. Bentley
2020-01-07 18:32 ` Ingo Schwarze
2020-01-10 10:53   ` Anthony J. Bentley
2020-01-08 12:21 ` Ingo Schwarze

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).