Add a complete suite of tests covering ASCII characters in UTF-8 spelling.

Most of this goes to nogroff.in rather than ascii.in for now
because groff-1.23.0 is buggy as hell in this respect
and i'm too lazy to debug the horrific function make_glyph_node()
in the file groff/src/roff/troff/node.cpp right now.

diff --git a/regress/usr.bin/mandoc/char/unicode/ascii.in b/regress/usr.bin/mandoc/char/unicode/ascii.in
index b8e561a..cbefe4e 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/ascii.in
+++ b/regress/usr.bin/mandoc/char/unicode/ascii.in
@@ -1,10 +1,11 @@
-.\" $OpenBSD: ascii.in,v 1.4 2017/07/04 14:53:23 schwarze Exp $
-.TH CHAR-UNICODE-ASCII 1 "October 27, 2014"
+.\" $OpenBSD: ascii.in,v 1.5 2024/05/16 18:49:00 schwarze Exp $
+.TH CHAR-UNICODE-ASCII 1 "May 16, 2024"
 .SH NAME
 char-unicode-ascii \- Unicode characters in the ASCII range
 .SH DESCRIPTION
 .nf
 BEGINTEST
+\[u0020]\N'32'         SPACE
 \[u0022]\N'34'\(dq     QUOTATION MARK
 \[u0023]\N'35'\(sh     NUMBER SIGN
 \[u0024]\N'36'\(Do     DOLLAR SIGN

diff --git a/regress/usr.bin/mandoc/char/unicode/ascii.out_ascii b/regress/usr.bin/mandoc/char/unicode/ascii.out_ascii
index 8a63135..2540bca 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/ascii.out_ascii
+++ b/regress/usr.bin/mandoc/char/unicode/ascii.out_ascii
@@ -5,6 +5,7 @@ N\bNA\bAM\bME\bE
 
 D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      BEGINTEST
+          SPACE
      """  QUOTATION MARK
      ###  NUMBER SIGN
      $$$  DOLLAR SIGN
@@ -27,4 +28,4 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      ~~~~ TILDE
      ENDTEST
 
-OpenBSD                        October 27, 2014          CHAR-UNICODE-ASCII(1)
+OpenBSD                          May 16, 2024            CHAR-UNICODE-ASCII(1)

diff --git a/regress/usr.bin/mandoc/char/unicode/ascii.out_html b/regress/usr.bin/mandoc/char/unicode/ascii.out_html
index 264fa35..441362a 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/ascii.out_html
+++ b/regress/usr.bin/mandoc/char/unicode/ascii.out_html
@@ -1,3 +1,4 @@
+       SPACE
 """     QUOTATION MARK
 ###    NUMBER SIGN
 $$$    DOLLAR SIGN

diff --git a/regress/usr.bin/mandoc/char/unicode/ascii.out_utf8 b/regress/usr.bin/mandoc/char/unicode/ascii.out_utf8
index 8a63135..2540bca 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/ascii.out_utf8
+++ b/regress/usr.bin/mandoc/char/unicode/ascii.out_utf8
@@ -5,6 +5,7 @@ N\bNA\bAM\bME\bE
 
 D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      BEGINTEST
+          SPACE
      """  QUOTATION MARK
      ###  NUMBER SIGN
      $$$  DOLLAR SIGN
@@ -27,4 +28,4 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      ~~~~ TILDE
      ENDTEST
 
-OpenBSD                        October 27, 2014          CHAR-UNICODE-ASCII(1)
+OpenBSD                          May 16, 2024            CHAR-UNICODE-ASCII(1)

diff --git a/regress/usr.bin/mandoc/char/unicode/nogroff.in b/regress/usr.bin/mandoc/char/unicode/nogroff.in
index 70853b5..cfeba6f 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/nogroff.in
+++ b/regress/usr.bin/mandoc/char/unicode/nogroff.in
@@ -1,5 +1,5 @@
-.\" $OpenBSD: nogroff.in,v 1.7 2024/05/15 19:37:24 schwarze Exp $
-.TH CHAR-UNICODE-NOGROFF 1 "June 2, 2021"
+.\" $OpenBSD: nogroff.in,v 1.8 2024/05/16 18:49:00 schwarze Exp $
+.TH CHAR-UNICODE-NOGROFF 1 "May 16, 2024"
 .SH NAME
 char-unicode-nogroff \- characters handled differently by groff
 .SH DESCRIPTION
@@ -37,6 +37,27 @@ BEGINTEST
 \[u001D]\N'29' <control> INFORMATION SEPARATOR THREE
 \[u001E]\N'30' <control> INFORMATION SEPARATOR TWO
 \[u001F]\N'31' <control> INFORMATION SEPARATOR INE
+\[u0021]\N'33' EXCLAMATION MARK
+\[u0025]\N'37' PERCENT SIGN
+\[u0026]\N'38' AMPERSAND
+\[u0028]\N'40' LEFT PARENTHESIS
+\[u0029]\N'41' RIGHT PARENTHESIS
+\[u002A]\N'42' ASTERISK
+\[u002C]\N'44' COMMA
+\[u002D]\N'45' HYPHEN-MINUS
+\[u002E]\N'46' FULL STOP
+\[u0030]\N'48' DIGIT ZERO
+\[u0031]\N'49' DIGIT ONE
+\[u0039]\N'57' DIGIT NINE
+\[u003A]\N'58' COLON
+\[u003B]\N'59' SEMICOLON
+\[u003C]\N'60' LESS-THAN SIGN
+\[u003E]\N'62' GREATER-THAN SIGN
+\[u003F]\N'63' QUESTION MARK
+\[u0041]\N'65' LATIN CAPITAL LETTER A
+\[u005A]\N'90' LATIN CAPITAL LETTER Z
+\[u0061]\N'97' LATIN SMALL LETTER A
+\[u007A]\N'122'        LATIN SMALL LETTER Z
 \[u007F]\N'127'        <control> DELETE
 \[u0080]\N'128'        <control> 0x80
 \[u0081]\N'129'        <control> 0x81

diff --git a/regress/usr.bin/mandoc/char/unicode/nogroff.out_ascii b/regress/usr.bin/mandoc/char/unicode/nogroff.out_ascii
index 8bd4d37..3c06879 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/nogroff.out_ascii
+++ b/regress/usr.bin/mandoc/char/unicode/nogroff.out_ascii
@@ -37,6 +37,27 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      <GS><GS>  <control> INFORMATION SEPARATOR THREE
      <RS><RS>  <control> INFORMATION SEPARATOR TWO
      <US><US>  <control> INFORMATION SEPARATOR INE
+     !!   EXCLAMATION MARK
+     %%   PERCENT SIGN
+     &&   AMPERSAND
+     ((   LEFT PARENTHESIS
+     ))   RIGHT PARENTHESIS
+     **   ASTERISK
+     ,,   COMMA
+     --   HYPHEN-MINUS
+     ..   FULL STOP
+     00   DIGIT ZERO
+     11   DIGIT ONE
+     99   DIGIT NINE
+     ::   COLON
+     ;;   SEMICOLON
+     <<   LESS-THAN SIGN
+     >>   GREATER-THAN SIGN
+     ??   QUESTION MARK
+     AA   LATIN CAPITAL LETTER A
+     ZZ   LATIN CAPITAL LETTER Z
+     aa   LATIN SMALL LETTER A
+     zz   LATIN SMALL LETTER Z
      <DEL><DEL>     <control> DELETE
      <80><80>  <control> 0x80
      <81><81>  <control> 0x81
@@ -98,4 +119,4 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      <?>  <undefined>
      ENDTEST
 
-OpenBSD                          June 2, 2021          CHAR-UNICODE-NOGROFF(1)
+OpenBSD                          May 16, 2024          CHAR-UNICODE-NOGROFF(1)

diff --git a/regress/usr.bin/mandoc/char/unicode/nogroff.out_html b/regress/usr.bin/mandoc/char/unicode/nogroff.out_html
index 5c48e9d..8f57109 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/nogroff.out_html
+++ b/regress/usr.bin/mandoc/char/unicode/nogroff.out_html
@@ -30,6 +30,27 @@
 ��       <control> INFORMATION SEPARATOR THREE
 ��       <control> INFORMATION SEPARATOR TWO
 ��       <control> INFORMATION SEPARATOR INE
+!!     EXCLAMATION MARK
+%%     PERCENT SIGN
+&&     AMPERSAND
+((     LEFT PARENTHESIS
+))     RIGHT PARENTHESIS
+**     ASTERISK
+,,     COMMA
+--     HYPHEN-MINUS
+..     FULL STOP
+00     DIGIT ZERO
+11     DIGIT ONE
+99     DIGIT NINE
+::     COLON
+;;     SEMICOLON
+<<       LESS-THAN SIGN
+>>       GREATER-THAN SIGN
+??     QUESTION MARK
+AA     LATIN CAPITAL LETTER A
+ZZ     LATIN CAPITAL LETTER Z
+aa     LATIN SMALL LETTER A
+zz     LATIN SMALL LETTER Z
 ��       <control> DELETE
 ��       <control> 0x80
 ��       <control> 0x81

diff --git a/regress/usr.bin/mandoc/char/unicode/nogroff.out_utf8 b/regress/usr.bin/mandoc/char/unicode/nogroff.out_utf8
index bad1c23..429c53f 100644 (file)
--- a/regress/usr.bin/mandoc/char/unicode/nogroff.out_utf8
+++ b/regress/usr.bin/mandoc/char/unicode/nogroff.out_utf8
@@ -37,6 +37,27 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      ��   <control> INFORMATION SEPARATOR THREE
      ��   <control> INFORMATION SEPARATOR TWO
      ��   <control> INFORMATION SEPARATOR INE
+     !!   EXCLAMATION MARK
+     %%   PERCENT SIGN
+     &&   AMPERSAND
+     ((   LEFT PARENTHESIS
+     ))   RIGHT PARENTHESIS
+     **   ASTERISK
+     ,,   COMMA
+     --   HYPHEN-MINUS
+     ..   FULL STOP
+     00   DIGIT ZERO
+     11   DIGIT ONE
+     99   DIGIT NINE
+     ::   COLON
+     ;;   SEMICOLON
+     <<   LESS-THAN SIGN
+     >>   GREATER-THAN SIGN
+     ??   QUESTION MARK
+     AA   LATIN CAPITAL LETTER A
+     ZZ   LATIN CAPITAL LETTER Z
+     aa   LATIN SMALL LETTER A
+     zz   LATIN SMALL LETTER Z
      ��   <control> DELETE
      ��   <control> 0x80
      ��   <control> 0x81
@@ -98,4 +119,4 @@ D\bDE\bES\bSC\bCR\bRI\bIP\bPT\bTI\bIO\bON\bN
      􏿿     <undefined>
      ENDTEST
 
-OpenBSD                          June 2, 2021          CHAR-UNICODE-NOGROFF(1)
+OpenBSD                          May 16, 2024          CHAR-UNICODE-NOGROFF(1)