Contents of /code/trunk/132html

#! /usr/bin/perl -w

# Script to turn PCRE man pages into HTML


# Subroutine to handle font changes and other escapes

sub do_line {
my($s) = $_[0];

$s =~ s/</&#60;/g;                   # Deal with < and >
$s =~ s/>/&#62;/g;
$s =~ s"\\fI(.*?)\\f[RP]"<i>$1</i>"g;
$s =~ s"\\fB(.*?)\\f[RP]"<b>$1</b>"g;
$s =~ s"\\e"\\"g;
$s =~ s/(?<=Copyright )\(c\)/&copy;/g;
$s;
}

# Subroutine to ensure not in a paragraph

sub end_para {
if ($inpara)
  {
  print TEMP "</PRE>\n" if ($inpre);
  print TEMP "</P>\n";
  } 
$inpara = $inpre = 0;
$wrotetext = 0;
}

# Subroutine to start a new paragraph

sub new_para {
&end_para();
print TEMP "<P>\n";
$inpara = 1;  
}


# Main program

$inpara = 0;
$inpre = 0;
$wrotetext = 0;
$toc = 0;
$ref = 1;

while ($#ARGV >= 0 && $ARGV[0] =~ /^-/)
  {
  $toc = 1 if $ARGV[0] eq "-toc"; 
  shift;
  } 

# Initial output to STDOUT

print <<End ;
<html>
<head>
<title>$ARGV[0] specification</title>
</head>
<body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
<h1>$ARGV[0] man page</h1>
<p>
Return to the <a href="index.html">PCRE index page</a>.
</p>
<p> 
This page is part of the PCRE HTML documentation. It was generated automatically
from the original man page. If there is any nonsense in it, please consult the
man page, in case the conversion went wrong.
<br> 
End

print "<ul>\n" if ($toc);

open(TEMP, ">/tmp/$$") || die "Can't open /tmp/$$ for output\n";

while (<STDIN>)
  {
  # Handle lines beginning with a dot 
   
  if (/^\./)
    {
    if (/^\.br/)
      {
      print TEMP "<br>\n" if $wrotetext; 
      }
      
    # Handling .sp is subtle. If it is inside a literal section, do nothing if
    # the next line is a non literal text line; similarly, if not inside a
    # literal section, do nothing if a literal follows. The point being that
    # the <pre> and </pre> that delimit literal sections will do the spacing.
    # Always skip if no previous output.
       
    elsif (/^\.sp/)
      {
      if ($wrotetext)
        {
        $_ = <STDIN>;
        if ($inpre)
          {
          print TEMP "\n" if (/^[\s.]/); 
          }
        else
          {      
          print TEMP "<br>\n<br>\n" if (!/^[\s.]/); 
          }
        redo;    # Now process the lookahead line we just read
        }  
      }     
    elsif (/^\.TP/ || /^\.PP/ || /^\.P/)
      {
      &new_para(); 
      }    
    elsif (/^\.SH\s*("?)(.*)\1/)
      {
      # Ignore the NAME section
      if ($2 =~ /^NAME\b/)
        {
        <STDIN>;
        next;
        }     
 
      &end_para();
      my($title) = &do_line($2); 
      if ($toc)
        {
        printf("<li><a name=\"TOC%d\" href=\"#SEC%d\">$title</a>\n", 
          $ref, $ref);
        printf TEMP ("<br><a name=\"SEC%d\" href=\"#TOC1\">$title</a><br>\n", 
          $ref, $ref);
        $ref++;
        }
      else
        {       
        print TEMP "<br><b>\n$title\n</b><br>\n"; 
        } 
      }     
    elsif (/^\.SS\s*("?)(.*)\1/)
      {
      &end_para();
      my($title) = &do_line($2); 
      print TEMP "<br><b>\n$title\n</b><br>\n"; 
      }     
    elsif (/^\.B\s*(.*)/)
      {
      &new_para() if (!$inpara);
      $_ = &do_line($1); 
      s/"(.*?)"/$1/g;
      print TEMP "<b>$_</b>\n";
      $wrotetext = 1; 
      }
    elsif (/^\.I\s*(.*)/)
      {
      &new_para() if (!$inpara);
      $_ = &do_line($1); 
      s/"(.*?)"/$1/g;
      print TEMP "<i>$_</i>\n";
      $wrotetext = 1; 
      }
      
    # A comment that starts "HREF" takes the next line as a name that
    # is turned into a hyperlink, using the text given, which might be
    # in a special font. If it ends in () or (digits) or punctuation, they
    # aren't part of the link.
    
    elsif (/^\.\\"\s*HREF/)
      {
      $_=<STDIN>;
      chomp; 
      $_ = &do_line($_);
      $_ =~ s/\s+$//; 
      $_ =~ /^(?:<.>)?([^<(]+)(?:\(\))?(?:<\/.>)?(?:\(\d+\))?[.,;:]?$/;
      print TEMP "<a href=\"$1.html\">$_</a>\n"; 
      }    
      
    # A comment that starts "HTML" inserts literal HTML
      
    elsif (/^\.\\"\s*HTML\s*(.*)/)
      {
      print TEMP $1; 
      }   
      
    # A comment that starts < inserts that HTML at the end of the
    # *next* input line - so as not to get a newline between them.
    
    elsif (/^\.\\"\s*(<.*>)/)
      {
      my($markup) = $1; 
      $_=<STDIN>;
      chomp;
      $_ = &do_line($_);
      $_ =~ s/\s+$//; 
      print TEMP "$_$markup\n";
      }     
      
    # A comment that starts JOIN joins the next two lines together, with one
    # space between them. Then that line is processed. This is used in some 
    # displays where two lines are needed for the "man" version. JOINSH works
    # the same, except that it assumes this is a shell command, so removes
    # continuation backslashes.  
    
    elsif (/^\.\\"\s*JOIN(SH)?/)
      {
      my($one,$two);
      $one = <STDIN>;
      $two = <STDIN>;
      $one =~ s/\s*\\e\s*$// if (defined($1));
      chomp($one);
      $two =~ s/^\s+//;
      $_ = "$one $two";
      redo;            # Process the joined lines
      }    

    # Ignore anything not recognized
    
    next;
    }  

  # Line does not begin with a dot. Replace blank lines with new paragraphs
  
  if (/^\s*$/)
    {
    &end_para() if ($wrotetext);
    next; 
    }  
 
  # Convert fonts changes and output an ordinary line. Ensure that indented
  # lines are marked as literal. 

  $_ = &do_line($_); 
  &new_para() if (!$inpara);
  
  if (/^\s/)
    {
    if (!$inpre)
      {
      print TEMP "<pre>\n";
      $inpre = 1;
      }    
    }
  elsif ($inpre)
    {
    print TEMP "</pre>\n"; 
    $inpre = 0; 
    }       
 
  print TEMP; 
  $wrotetext = 1; 
  } 
  
# The TOC, if present, will have been written - terminate it

print "</ul>\n" if ($toc);

# Copy the remainder to the standard output

close(TEMP);
open(TEMP, "/tmp/$$") || die "Can't open /tmp/$$ for input\n";
  
print while (<TEMP>);

print <<End ;
<p>
Return to the <a href="index.html">PCRE index page</a>.
</p>
End

close(TEMP);
unlink("/tmp/$$");

# End
1	#! /usr/bin/perl -w
2
3	# Script to turn PCRE man pages into HTML
4
5
6	# Subroutine to handle font changes and other escapes
7
8	sub do_line {
9	my($s) = $_[0];
10
11	$s =~ s/</</g; # Deal with < and >
12	$s =~ s/>/>/g;
13	$s =~ s"\\fI(.*?)\\f[RP]"<i>$1</i>"g;
14	$s =~ s"\\fB(.*?)\\f[RP]"<b>$1</b>"g;
15	$s =~ s"\\e"\\"g;
16	$s =~ s/(?<=Copyright )\(c\)/©/g;
17	$s;
18	}
19
20	# Subroutine to ensure not in a paragraph
21
22	sub end_para {
23	if ($inpara)
24	{
25	print TEMP "</PRE>\n" if ($inpre);
26	print TEMP "</P>\n";
27	}
28	$inpara = $inpre = 0;
29	$wrotetext = 0;
30	}
31
32	# Subroutine to start a new paragraph
33
34	sub new_para {
35	&end_para();
36	print TEMP "<P>\n";
37	$inpara = 1;
38	}
39
40
41	# Main program
42
43	$inpara = 0;
44	$inpre = 0;
45	$wrotetext = 0;
46	$toc = 0;
47	$ref = 1;
48
49	while ($#ARGV >= 0 && $ARGV[0] =~ /^-/)
50	{
51	$toc = 1 if $ARGV[0] eq "-toc";
52	shift;
53	}
54
55	# Initial output to STDOUT
56
57	print <<End ;
58	<html>
59	<head>
60	<title>$ARGV[0] specification</title>
61	</head>
62	<body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
63	<h1>$ARGV[0] man page</h1>
64	<p>
65	Return to the <a href="index.html">PCRE index page</a>.
66	</p>
67	<p>
68	This page is part of the PCRE HTML documentation. It was generated automatically
69	from the original man page. If there is any nonsense in it, please consult the
70	man page, in case the conversion went wrong.
71	<br>
72	End
73
74	print "<ul>\n" if ($toc);
75
76	open(TEMP, ">/tmp/$$") \|\| die "Can't open /tmp/$$ for output\n";
77
78	while (<STDIN>)
79	{
80	# Handle lines beginning with a dot
81
82	if (/^\./)
83	{
84	if (/^\.br/)
85	{
86	print TEMP "<br>\n" if $wrotetext;
87	}
88
89	# Handling .sp is subtle. If it is inside a literal section, do nothing if
90	# the next line is a non literal text line; similarly, if not inside a
91	# literal section, do nothing if a literal follows. The point being that
92	# the <pre> and </pre> that delimit literal sections will do the spacing.
93	# Always skip if no previous output.
94
95	elsif (/^\.sp/)
96	{
97	if ($wrotetext)
98	{
99	$_ = <STDIN>;
100	if ($inpre)
101	{
102	print TEMP "\n" if (/^[\s.]/);
103	}
104	else
105	{
106	print TEMP "<br>\n<br>\n" if (!/^[\s.]/);
107	}
108	redo; # Now process the lookahead line we just read
109	}
110	}
111	elsif (/^\.TP/ \|\| /^\.PP/ \|\| /^\.P/)
112	{
113	&new_para();
114	}
115	elsif (/^\.SH\s("?)(.)\1/)
116	{
117	# Ignore the NAME section
118	if ($2 =~ /^NAME\b/)
119	{
120	<STDIN>;
121	next;
122	}
123
124	&end_para();
125	my($title) = &do_line($2);
126	if ($toc)
127	{
128	printf("<li><a name=\"TOC%d\" href=\"#SEC%d\">$title</a>\n",
129	$ref, $ref);
130	printf TEMP ("<br><a name=\"SEC%d\" href=\"#TOC1\">$title</a><br>\n",
131	$ref, $ref);
132	$ref++;
133	}
134	else
135	{
136	print TEMP "<br><b>\n$title\n</b><br>\n";
137	}
138	}
139	elsif (/^\.SS\s("?)(.)\1/)
140	{
141	&end_para();
142	my($title) = &do_line($2);
143	print TEMP "<br><b>\n$title\n</b><br>\n";
144	}
145	elsif (/^\.B\s(.)/)
146	{
147	&new_para() if (!$inpara);
148	$_ = &do_line($1);
149	s/"(.*?)"/$1/g;
150	print TEMP "<b>$_</b>\n";
151	$wrotetext = 1;
152	}
153	elsif (/^\.I\s(.)/)
154	{
155	&new_para() if (!$inpara);
156	$_ = &do_line($1);
157	s/"(.*?)"/$1/g;
158	print TEMP "<i>$_</i>\n";
159	$wrotetext = 1;
160	}
161
162	# A comment that starts "HREF" takes the next line as a name that
163	# is turned into a hyperlink, using the text given, which might be
164	# in a special font. If it ends in () or (digits) or punctuation, they
165	# aren't part of the link.
166
167	elsif (/^\.\\"\s*HREF/)
168	{
169	$_=<STDIN>;
170	chomp;
171	$_ = &do_line($_);
172	$_ =~ s/\s+$//;
173	$_ =~ /^(?:<.>)?([^<(]+)(?:\(\))?(?:<\/.>)?(?:\(\d+\))?[.,;:]?$/;
174	print TEMP "<a href=\"$1.html\">$_</a>\n";
175	}
176
177	# A comment that starts "HTML" inserts literal HTML
178
179	elsif (/^\.\\"\sHTML\s(.*)/)
180	{
181	print TEMP $1;
182	}
183
184	# A comment that starts < inserts that HTML at the end of the
185	# next input line - so as not to get a newline between them.
186
187	elsif (/^\.\\"\s(<.>)/)
188	{
189	my($markup) = $1;
190	$_=<STDIN>;
191	chomp;
192	$_ = &do_line($_);
193	$_ =~ s/\s+$//;
194	print TEMP "$_$markup\n";
195	}
196
197	# A comment that starts JOIN joins the next two lines together, with one
198	# space between them. Then that line is processed. This is used in some
199	# displays where two lines are needed for the "man" version. JOINSH works
200	# the same, except that it assumes this is a shell command, so removes
201	# continuation backslashes.
202
203	elsif (/^\.\\"\s*JOIN(SH)?/)
204	{
205	my($one,$two);
206	$one = <STDIN>;
207	$two = <STDIN>;
208	$one =~ s/\s\\e\s$// if (defined($1));
209	chomp($one);
210	$two =~ s/^\s+//;
211	$_ = "$one $two";
212	redo; # Process the joined lines
213	}
214
215	# Ignore anything not recognized
216
217	next;
218	}
219
220	# Line does not begin with a dot. Replace blank lines with new paragraphs
221
222	if (/^\s*$/)
223	{
224	&end_para() if ($wrotetext);
225	next;
226	}
227
228	# Convert fonts changes and output an ordinary line. Ensure that indented
229	# lines are marked as literal.
230
231	$_ = &do_line($_);
232	&new_para() if (!$inpara);
233
234	if (/^\s/)
235	{
236	if (!$inpre)
237	{
238	print TEMP "<pre>\n";
239	$inpre = 1;
240	}
241	}
242	elsif ($inpre)
243	{
244	print TEMP "</pre>\n";
245	$inpre = 0;
246	}
247
248	print TEMP;
249	$wrotetext = 1;
250	}
251
252	# The TOC, if present, will have been written - terminate it
253
254	print "</ul>\n" if ($toc);
255
256	# Copy the remainder to the standard output
257
258	close(TEMP);
259	open(TEMP, "/tmp/$$") \|\| die "Can't open /tmp/$$ for input\n";
260
261	print while (<TEMP>);
262
263	print <<End ;
264	<p>
265	Return to the <a href="index.html">PCRE index page</a>.
266	</p>
267	End
268
269	close(TEMP);
270	unlink("/tmp/$$");
271
272	# End