Working towards better nroff scraping...

4 files changed, 97 insertions, 31 deletions

diff --git a/ChangeLog b/ChangeLog
index cf8568e..9ed5822 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,9 +1,15 @@
+2007-10-26  Donal K. Fellows  <donal.k.fellows@man.ac.uk>
+
+	* tools/man2help2.tcl, tools/man2tcl.c: Made some of the tooling code
+	to do man->other formats work better with current manpage set. Long
+	way still to go.
+
 2007-10-25  Zoran Vasiljevic <vasiljevic@users.sourceforge.net>
 
 	* generic/tclThread.c: Added TclpMasterLock/Unlock arround calls to
-	ForgetSyncObject in Tcl_MutexFinalize and Tcl_ConditionFinalize
-	to prevent from garbling the internal lists that track sync objects.
-	(Tcl Bug #1726873)
+	ForgetSyncObject in Tcl_MutexFinalize and Tcl_ConditionFinalize to
+	prevent from garbling the internal lists that track sync objects. [Bug
+	1726873]
 
 2007-10-24  Donal K. Fellows  <donal.k.fellows@man.ac.uk>
 
@@ -22,8 +28,8 @@
 
 2007-10-18  David Gravereaux <davygrvy@pobox.com>
 
-	* tools/mkdepend.tcl: sort the dep list for a more humanly
-	readable output.
+	* tools/mkdepend.tcl: sort the dep list for a more humanly readable
+	output.
 
 2007-10-18  Don Porter	<dgp@users.sourceforge.net>
 
@@ -31,14 +37,14 @@
 	values get pulled out of the dictionary, even if they are integer
 	valued.
 
-	* generic/tclCompCmds.c (TclCompileReturnCmd):	Added code to
-	more optimally compile [return -level 0 $x] to "push $x". [RFE 1794073]
+	* generic/tclCompCmds.c (TclCompileReturnCmd):	Added code to more
+	optimally compile [return -level 0 $x] to "push $x". [RFE 1794073]
 
 	* compat/tmpnam.c (removed):	The routine tmpnam() is no longer
-	* unix/Makefile.in:	called by Tcl source code.  Remove autogoo
-	* unix/configure.in:	the supplied a replacement version on
-	* win/tcl.dsp:		systems where the routine was not available.
-	[RFE 1811848].
+	* unix/Makefile.in:	called by Tcl source code. Remove autogoo the
+	* unix/configure.in:	supplied a replacement version on systems
+	* win/tcl.dsp:		where the routine was not available. [RFE
+	1811848]
 
 	* unix/configure:	autoconf-2.59
 
@@ -46,8 +52,8 @@
 
 2007-10-17  David Gravereaux <davygrvy@pobox.com>
 
-	* tools/mkdepend.tcl:	Improved defense from malformed object
-	list infile.
+	* tools/mkdepend.tcl:	Improved defense from malformed object list
+	infile.
 
 2007-10-17  Donal K. Fellows  <donal.k.fellows@manchester.ac.uk>
 
diff --git a/tools/man2help.tcl b/tools/man2help.tcl
index 91fcb89..c61085b 100644
--- a/tools/man2help.tcl
+++ b/tools/man2help.tcl
@@ -6,14 +6,15 @@
 #
 # Copyright (c) 1996 by Sun Microsystems, Inc.
 #
-# RCS: @(#) $Id: man2help.tcl,v 1.14 2003/06/04 23:40:18 mistachkin Exp $
+# RCS: @(#) $Id: man2help.tcl,v 1.15 2007/10/26 14:32:16 dkf Exp $
 # 
 
 #
 # PASS 1
 #
 
-set man2tclprog [file join [file dirname [info script]] man2tcl.exe]
+set man2tclprog [file join [file dirname [info script]] \
+	man2tcl[file extension [info nameofexecutable]]]
 
 proc generateContents {basename version files} {
     global curID topics
diff --git a/tools/man2help2.tcl b/tools/man2help2.tcl
index 62b5b2e..7cd020c 100644
--- a/tools/man2help2.tcl
+++ b/tools/man2help2.tcl
@@ -9,7 +9,7 @@
 # See the file "license.terms" for information on usage and redistribution
 # of this file, and for a DISCLAIMER OF ALL WARRANTIES.
 # 
-# RCS: @(#) $Id: man2help2.tcl,v 1.15 2004/07/07 12:08:43 dkf Exp $
+# RCS: @(#) $Id: man2help2.tcl,v 1.16 2007/10/26 14:32:16 dkf Exp $
 # 
 
 # Global variables used by these scripts:
@@ -425,6 +425,21 @@ proc macro {name args} {
 	}
 	VE {}
 	VS {}
+	QW {
+	    formattedText "``[lindex $args 0]''[lindex $args 1] "
+	}
+	MT {
+	    text "``'' "
+	}
+	PQ {
+	    formattedText \
+		"(``[lindex $args 0]''[lindex $args 1])[lindex $args 2] "
+	}
+	QR {
+	    formattedText "``[lindex $args 0]"
+	    dash
+	    formattedText "[lindex $args 1]''[lindex $args 2] "
+	}
 	default {
 	    puts stderr "Unknown macro: .$name [join $args " "]"
 	}
@@ -516,13 +531,12 @@ proc formattedText {text} {
 		dash
 		set text [string range $text [expr {$index+2}] end]
 	    }
-	    | {
+	    & - | {
 		set text [string range $text [expr {$index+2}] end]
 	    }
-	    o {
-		text "\\'"
-		regexp {'([^']*)'(.*)} $text all ch text
-		text $chars($ch)
+	    ( {
+		char [string range $text $index [expr {$index+3}]]
+		set text [string range $text [expr {$index+4}] end]
 	    }
 	    default {
 		puts stderr "Unknown sequence: \\$c"
@@ -665,32 +679,51 @@ proc char {name} {
     global file state
 
     switch -exact $name {
-        \\o {
+        {\o} {
 	    set state(intl) 1
 	}
-	\\\  {
+	{\ } {
 	    textSetup
 	    puts -nonewline $file " "
 	}
-	\\0 {
+	{\0} {
 	    textSetup
 	    puts -nonewline $file " \\emspace "
 	}
-	\\\\ {
+	{\\} - {\e} {
 	    textSetup
 	    puts -nonewline $file "\\\\"
 	}
-	\\(+- {
+	{\(+-} {
 	    textSetup
 	    puts -nonewline $file "\\'b1 "
 	}
-	\\% -
-	\\| {
+	{\%} - {\|} {
 	}
-	\\(bu {
+	{\(->} {
+	    textSetup
+	    puts -nonewline $file "->"
+	}
+	{\(bu} {
 	    textSetup
 	    puts -nonewline $file "\\bullet "
 	}
+	{\(co} {
+	    textSetup
+	    puts -nonewline $file "\\'a9 "
+	}
+	{\(mu} {
+	    textSetup
+	    puts -nonewline $file "\\'d7 "
+	}
+	{\(em} {
+	    textSetup
+	    puts -nonewline $file "-"
+	}
+	{\(fm} {
+	    textSetup
+	    puts -nonewline $file "\\'27 "
+	}
 	default {
 	    puts stderr "Unknown character: $name"
 	}
diff --git a/tools/man2tcl.c b/tools/man2tcl.c
index a90b729..8ddeb20 100644
--- a/tools/man2tcl.c
+++ b/tools/man2tcl.c
@@ -15,7 +15,7 @@
  * See the file "license.terms" for information on usage and redistribution of
  * this file, and for a DISCLAIMER OF ALL WARRANTIES.
  *
- * RCS: @(#) $Id: man2tcl.c,v 1.11 2007/01/19 08:17:35 mistachkin Exp $
+ * RCS: @(#) $Id: man2tcl.c,v 1.12 2007/10/26 14:32:16 dkf Exp $
  */
 
 static char sccsid[] = "@(#) man2tcl.c 1.3 95/08/12 17:34:08";
@@ -329,6 +329,9 @@ DoText(
 		p++;
 	    } else if (*p == '&') {
 		p++;
+	    } else if (*p == '0') {
+		PRINT(("text { }\n"));
+		p++;
 	    } else if (*p == '(') {
 		if ((p[1] == 0) || (p[2] == 0)) {
 		    fprintf(stderr, "Bad \\( sequence on line %d.\n",
@@ -338,6 +341,13 @@ DoText(
 		    PRINT(("char {\\(%c%c}\n", p[1], p[2]));
 		    p += 3;
 		}
+	    } else if (*p == 'N' && *(p+1) == '\'') {
+		int ch;
+
+		p += 2;
+		sscanf(p,"%d",&ch);
+		PRINT(("text \\u%04x", ch));
+		while(*p&&*p!='\'') p++;
 	    } else if (*p != 0) {
 		PRINT(("char {\\%c}\n", *p));
 		p++;
@@ -377,7 +387,23 @@ QuoteText(
     }
     for ( ; count > 0; string++, count--) {
 	switch (*string) {
-	case '$': case '[': case '{': case ' ': case ';': case '\\':
+	case '\\':
+	    if (*(string+1) == 'N' && *(string+2) == '\'') {
+		int ch;
+
+		string += 3;
+		count -= 3;
+		sscanf(string,"%d",&ch);
+		PRINT(("\\u%04x", ch));
+		while(count>0&&*string!='\'') {string++;count--;}
+		continue;
+	    } else if (*(string+1) == '0') {
+		PRINT(("\\ "));
+		string++;
+		count--;
+		continue;
+	    }
+	case '$': case '[': case '{': case ' ': case ';':
 	case '"': case '\t':
 	    PRINTC('\\');
 	default: