[Bug 2903743]: Try to do the right thing when presented with old-style nroff.

4 files changed, 35 insertions, 43 deletions

diff --git a/ChangeLog b/ChangeLog
index 85119ca..5dd2fb8 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+2011-09-22  Donal K. Fellows  <dkf@users.sf.net>
+
+	* doc/re_syntax.n: [Bug 2903743]: Add more magic so that we can do at
+	least something sane on Solaris.
+	* tools/tcltk-man2html-utils.tcl (process-text): Teach the HTML
+	generator how to handle this magic.
+
 2011-09-21  Don Porter  <dgp@users.sourceforge.net>
 
 	* generic/tclThreadTest.c: Revise the thread exit handling of the
diff --git a/doc/re_syntax.n b/doc/re_syntax.n
index a53f58b..dacc41f 100644
--- a/doc/re_syntax.n
+++ b/doc/re_syntax.n
@@ -6,6 +6,8 @@
 '\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
 '\"
 .so man.macros
+.ie '\w'o''\w'\C'^o''' .ds qo \C'^o'
+.el .ds qo u
 .TH re_syntax n "8.1" Tcl "Tcl Built-In Commands"
 .BS
 .SH NAME
@@ -290,12 +292,12 @@ treatment is as if the enclosing delimiters were
 .QW \fB[.\fR \&
 and
 .QW \fB.]\fR .)
-For example, if \fBo\fR and \fB\N'244'\fR are the members of an
+For example, if \fBo\fR and \fB\*(qo\fR are the members of an
 equivalence class, then
 .QW \fB[[=o=]]\fR ,
-.QW \fB[[=\N'244'=]]\fR ,
+.QW \fB[[=\*(qo=]]\fR ,
 and
-.QW \fB[o\N'244']\fR \&
+.QW \fB[o\*(qo]\fR \&
 are all synonymous. An equivalence class may not be an endpoint of a range.
 .RS
 .PP
diff --git a/tools/tcltk-man2html-utils.tcl b/tools/tcltk-man2html-utils.tcl
index 938a1af..ef1f62a 100644
--- a/tools/tcltk-man2html-utils.tcl
+++ b/tools/tcltk-man2html-utils.tcl
@@ -110,6 +110,7 @@ proc htmlize-text {text {charmap {}}} {
     # contains some extras for use in nroff->html processing
     # build on the list passed in, if any
     lappend charmap \
+	"&ndash;" "&ndash;" \
 	{&}	{&amp;} \
 	{\\}	"&#92;" \
 	{\e}	"&#92;" \
@@ -143,8 +144,8 @@ proc process-text {text} {
 	    {\fP}	{\fR} \
 	    {\.}	. \
 	    {\(bu}	"&#8226;" \
+	    {\*(qo}	"&ocirc;" \
 	    ]
-    lappend charmap {\o'o^'} {&ocirc;} ; # o-circumflex in re_syntax.n
     lappend charmap {\-\|\-} --        ; # two hyphens
     lappend charmap {\-} -             ; # a hyphen
 
@@ -1063,25 +1064,17 @@ proc output-directive {line} {
 	    output-IP-list .IP .IP $rest
 	    return
 	}
-	.PP {
+	.PP - .sp {
 	    man-puts <P>
 	}
 	.RS {
 	    output-RS-list
 	    return
 	}
-	.RE {
-	    manerror "unexpected .RE"
-	    return
-	}
 	.br {
 	    man-puts <BR>
 	    return
 	}
-	.DE {
-	    manerror "unexpected .DE"
-	    return
-	}
 	.DS {
 	    if {[next-op-is .ta rest]} {
 		# skip the leading .ta directive if it is there
@@ -1109,16 +1102,6 @@ proc output-directive {line} {
 	    }
 	    return
 	}
-	.CE {
-	    manerror "unexpected .CE"
-	    return
-	}
-	.sp {
-	    man-puts <P>
-	}
-	.ta {
-	    manerror "ignoring $line"
-	}
 	.nf {
 	    if {[match-text @more .fi]} {
 		foreach more [split $more \n] {
@@ -1174,13 +1157,11 @@ proc output-directive {line} {
 		manerror "ignoring $line"
 	    }
 	}
-	.fi {
-	    manerror "ignoring $line"
+	.RE - .DE - .CE {
+	    manerror "unexpected $code"
+	    return
 	}
-	.na -
-	.ad -
-	.UL -
-	.ne {
+	.ta - .fi - .na - .ad - .UL - .ie - .el - .ne {
 	    manerror "ignoring $line"
 	}
 	default {
diff --git a/tools/tcltk-man2html.tcl b/tools/tcltk-man2html.tcl
index b347abf..e4845a6 100755
--- a/tools/tcltk-man2html.tcl
+++ b/tools/tcltk-man2html.tcl
@@ -360,7 +360,7 @@ proc make-man-pages {html args} {
 		    continue
 		}
 		switch -exact -- $code {
-		    .if - .nr - .ti - .in -
+		    .if - .nr - .ti - .in - .ie - .el -
 		    .ad - .na - .so - .ne - .AS - .VE - .VS - . {
 			# ignore
 			continue
@@ -379,21 +379,22 @@ proc make-man-pages {html args} {
 			lappend manual(text) "$code [unquote $rest]"
 		    }
 		    .QW {
-			set rest [regexp -all -inline {\"(?:[^""]+)\"|\S+} $rest]
-			addbuffer $LQ [unquote [lindex $rest 0]] $RQ \
-			    [unquote [lindex $rest 1]]
+			lassign [regexp -all -inline {\"(?:[^""]+)\"|\S+} $rest] \
+			    inQuote afterwards
+			addbuffer $LQ [unquote $inQuote] $RQ [unquote $afterwards]
 		    }
 		    .PQ {
-			set rest [regexp -all -inline {\"(?:[^""]+)\"|\S+} $rest]
-			addbuffer ( $LQ [unquote [lindex $rest 0]] $RQ \
-			    [unquote [lindex $rest 1]] ) \
-			    [unquote [lindex $rest 2]]
+			lassign [regexp -all -inline {\"(?:[^""]+)\"|\S+} $rest] \
+			    inQuote punctuation afterwards
+			addbuffer ( $LQ [unquote $inQuote] $RQ \
+			    [unquote $punctuation] ) \
+			    [unquote $afterwards]
 		    }
 		    .QR {
-			set rest [regexp -all -inline {\"(?:[^""]+)\"|\S+} $rest]
-			addbuffer $LQ [unquote [lindex $rest 0]] - \
-			    [unquote [lindex $rest 1]] $RQ \
-			    [unquote [lindex $rest 2]]
+			lassign [regexp -all -inline {\"(?:[^""]+)\"|\S+} $rest] \
+			    rangeFrom rangeTo afterwards
+			addbuffer $LQ [unquote $rangeFrom] "&ndash;" \
+			    [unquote $rangeTo] $RQ [unquote $afterwards]
 		    }
 		    .MT {
 			addbuffer $LQ$RQ
@@ -404,7 +405,7 @@ proc make-man-pages {html args} {
 		    }
 		    .BS - .BE - .br - .fi - .sp - .nf {
 			flushbuffer
-			if {"$rest" ne {}} {
+			if {$rest ne ""} {
 			    if {!$verbose} {
 				puts stderr ""
 			    }
@@ -435,8 +436,9 @@ proc make-man-pages {html args} {
 		    }
 		    .OP {
 			flushbuffer
+			lassign $rest cmdName dbName dbClass
 			lappend manual(text) [concat .OP [process-text \
-				"\\fB[lindex $rest 0]\\fR \\fB[lindex $rest 1]\\fR \\fB[lindex $rest 2]\\fR"]]
+				"\\fB$cmdName\\fR \\fB$dbName\\fR \\fB$dbClass\\fR"]]
 		    }
 		    .PP - .LP {
 			flushbuffer