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

$innf = 0;
$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 (/^\./)
    {
    # Some of the PCRE man pages used to contain instances of .br. However,
    # they should have all been removed because they cause trouble in some
    # (other) automated systems that translate man pages to HTML. Complain if
    # we find .br or .in (another macro that is deprecated).

    if (/^\.br/ || /^\.in/)
      {
      print STDERR "\n*** Deprecated macro encountered - rewrite needed\n";
      print STDERR "*** $_\n";
      die "*** Processing abandoned\n";
      }

    # Instead of .br, relevent "literal" sections are enclosed in .nf/.fi.

    elsif (/^\.nf/)
      {
      $innf = 1;
      }

    elsif (/^\.fi/)
      {
      $innf = 0;
      }

    # 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
      }
      
    # .EX/.EE are used in the pcredemo page to bracket the entire program,
    # which is unmodified except for turning backslash into "\e".
    
    elsif (/^\.EX\s*$/)
      {
      print TEMP "<PRE>\n";
      while (<STDIN>)
        {
        last if /^\.EE\s*$/; 
        s/\\e/\\/g;
        s/&/&amp;/g;   
        s/</&lt;/g;
        s/>/&gt;/g;
        print TEMP; 
        }   
      }     

    # 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;
    }

  # Add <br> to the end of a non-literal line if we are within .nf/.fi

  $_ .= "<br>\n" if (!$inpre && $innf);

  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	ph10	97	#! /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	ph10	111	}
28	ph10	97	$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	ph10	111	$inpara = 1;
38	ph10	97	}
39
40
41			# Main program
42
43	ph10	99	$innf = 0;
44	ph10	97	$inpara = 0;
45			$inpre = 0;
46			$wrotetext = 0;
47			$toc = 0;
48			$ref = 1;
49
50			while ($#ARGV >= 0 && $ARGV[0] =~ /^-/)
51			{
52	ph10	111	$toc = 1 if $ARGV[0] eq "-toc";
53	ph10	97	shift;
54	ph10	111	}
55	ph10	97
56			# Initial output to STDOUT
57
58			print <<End ;
59			<html>
60			<head>
61			<title>$ARGV[0] specification</title>
62			</head>
63			<body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
64			<h1>$ARGV[0] man page</h1>
65			<p>
66			Return to the <a href="index.html">PCRE index page</a>.
67			</p>
68	ph10	111	<p>
69	ph10	97	This page is part of the PCRE HTML documentation. It was generated automatically
70			from the original man page. If there is any nonsense in it, please consult the
71			man page, in case the conversion went wrong.
72	ph10	111	<br>
73	ph10	97	End
74
75			print "<ul>\n" if ($toc);
76
77			open(TEMP, ">/tmp/$$") \|\| die "Can't open /tmp/$$ for output\n";
78
79			while (<STDIN>)
80			{
81	ph10	111	# Handle lines beginning with a dot
82
83	ph10	97	if (/^\./)
84			{
85	ph10	111	# Some of the PCRE man pages used to contain instances of .br. However,
86	ph10	99	# they should have all been removed because they cause trouble in some
87			# (other) automated systems that translate man pages to HTML. Complain if
88	ph10	111	# we find .br or .in (another macro that is deprecated).
89
90	ph10	99	if (/^\.br/ \|\| /^\.in/)
91	ph10	97	{
92	ph10	99	print STDERR "\n*** Deprecated macro encountered - rewrite needed\n";
93	ph10	111	print STDERR "*** $_\n";
94	ph10	99	die "*** Processing abandoned\n";
95	ph10	97	}
96	ph10	111
97	ph10	99	# Instead of .br, relevent "literal" sections are enclosed in .nf/.fi.
98	ph10	111
99	ph10	99	elsif (/^\.nf/)
100			{
101	ph10	111	$innf = 1;
102			}
103
104	ph10	99	elsif (/^\.fi/)
105			{
106	ph10	111	$innf = 0;
107			}
108
109	ph10	97	# Handling .sp is subtle. If it is inside a literal section, do nothing if
110			# the next line is a non literal text line; similarly, if not inside a
111			# literal section, do nothing if a literal follows. The point being that
112			# the <pre> and </pre> that delimit literal sections will do the spacing.
113			# Always skip if no previous output.
114	ph10	111
115	ph10	97	elsif (/^\.sp/)
116			{
117			if ($wrotetext)
118			{
119			$_ = <STDIN>;
120			if ($inpre)
121			{
122	ph10	111	print TEMP "\n" if (/^[\s.]/);
123	ph10	97	}
124			else
125	ph10	111	{
126			print TEMP "<br>\n<br>\n" if (!/^[\s.]/);
127	ph10	97	}
128			redo; # Now process the lookahead line we just read
129	ph10	111	}
130			}
131	ph10	97	elsif (/^\.TP/ \|\| /^\.PP/ \|\| /^\.P/)
132			{
133	ph10	111	&new_para();
134			}
135	ph10	97	elsif (/^\.SH\s("?)(.)\1/)
136			{
137			# Ignore the NAME section
138			if ($2 =~ /^NAME\b/)
139			{
140			<STDIN>;
141			next;
142	ph10	111	}
143
144	ph10	97	&end_para();
145	ph10	111	my($title) = &do_line($2);
146	ph10	97	if ($toc)
147			{
148	ph10	111	printf("<li><a name=\"TOC%d\" href=\"#SEC%d\">$title</a>\n",
149	ph10	97	$ref, $ref);
150	ph10	111	printf TEMP ("<br><a name=\"SEC%d\" href=\"#TOC1\">$title</a><br>\n",
151	ph10	97	$ref, $ref);
152			$ref++;
153			}
154			else
155	ph10	111	{
156			print TEMP "<br><b>\n$title\n</b><br>\n";
157			}
158			}
159	ph10	97	elsif (/^\.SS\s("?)(.)\1/)
160			{
161			&end_para();
162	ph10	111	my($title) = &do_line($2);
163			print TEMP "<br><b>\n$title\n</b><br>\n";
164			}
165	ph10	97	elsif (/^\.B\s(.)/)
166			{
167			&new_para() if (!$inpara);
168	ph10	111	$_ = &do_line($1);
169	ph10	97	s/"(.*?)"/$1/g;
170			print TEMP "<b>$_</b>\n";
171	ph10	111	$wrotetext = 1;
172	ph10	97	}
173			elsif (/^\.I\s(.)/)
174			{
175			&new_para() if (!$inpara);
176	ph10	111	$_ = &do_line($1);
177	ph10	97	s/"(.*?)"/$1/g;
178			print TEMP "<i>$_</i>\n";
179	ph10	111	$wrotetext = 1;
180	ph10	97	}
181	ph10	111
182	ph10	97	# A comment that starts "HREF" takes the next line as a name that
183			# is turned into a hyperlink, using the text given, which might be
184			# in a special font. If it ends in () or (digits) or punctuation, they
185			# aren't part of the link.
186	ph10	111
187	ph10	97	elsif (/^\.\\"\s*HREF/)
188			{
189			$_=<STDIN>;
190	ph10	111	chomp;
191	ph10	97	$_ = &do_line($_);
192	ph10	111	$_ =~ s/\s+$//;
193	ph10	97	$_ =~ /^(?:<.>)?([^<(]+)(?:\(\))?(?:<\/.>)?(?:\(\d+\))?[.,;:]?$/;
194	ph10	111	print TEMP "<a href=\"$1.html\">$_</a>\n";
195			}
196
197	ph10	97	# A comment that starts "HTML" inserts literal HTML
198	ph10	111
199	ph10	97	elsif (/^\.\\"\sHTML\s(.*)/)
200			{
201	ph10	111	print TEMP $1;
202			}
203
204	ph10	97	# A comment that starts < inserts that HTML at the end of the
205			# next input line - so as not to get a newline between them.
206	ph10	111
207	ph10	97	elsif (/^\.\\"\s(<.>)/)
208			{
209	ph10	111	my($markup) = $1;
210	ph10	97	$_=<STDIN>;
211			chomp;
212			$_ = &do_line($_);
213	ph10	111	$_ =~ s/\s+$//;
214	ph10	97	print TEMP "$_$markup\n";
215	ph10	111	}
216
217	ph10	97	# A comment that starts JOIN joins the next two lines together, with one
218	ph10	111	# space between them. Then that line is processed. This is used in some
219	ph10	97	# displays where two lines are needed for the "man" version. JOINSH works
220			# the same, except that it assumes this is a shell command, so removes
221	ph10	111	# continuation backslashes.
222
223	ph10	97	elsif (/^\.\\"\s*JOIN(SH)?/)
224			{
225			my($one,$two);
226			$one = <STDIN>;
227			$two = <STDIN>;
228			$one =~ s/\s\\e\s$// if (defined($1));
229			chomp($one);
230			$two =~ s/^\s+//;
231			$_ = "$one $two";
232			redo; # Process the joined lines
233	ph10	111	}
234	ph10	429
235			# .EX/.EE are used in the pcredemo page to bracket the entire program,
236			# which is unmodified except for turning backslash into "\e".
237
238			elsif (/^\.EX\s*$/)
239			{
240			print TEMP "<PRE>\n";
241			while (<STDIN>)
242			{
243			last if /^\.EE\s*$/;
244			s/\\e/\\/g;
245			s/&/&/g;
246			s/</</g;
247			s/>/>/g;
248			print TEMP;
249			}
250			}
251	ph10	97
252			# Ignore anything not recognized
253	ph10	111
254	ph10	97	next;
255	ph10	111	}
256	ph10	97
257			# Line does not begin with a dot. Replace blank lines with new paragraphs
258	ph10	111
259	ph10	97	if (/^\s*$/)
260			{
261			&end_para() if ($wrotetext);
262	ph10	111	next;
263			}
264
265	ph10	97	# Convert fonts changes and output an ordinary line. Ensure that indented
266	ph10	111	# lines are marked as literal.
267	ph10	97
268	ph10	111	$_ = &do_line($_);
269	ph10	97	&new_para() if (!$inpara);
270	ph10	111
271	ph10	97	if (/^\s/)
272			{
273			if (!$inpre)
274			{
275			print TEMP "<pre>\n";
276			$inpre = 1;
277	ph10	111	}
278	ph10	97	}
279			elsif ($inpre)
280			{
281	ph10	111	print TEMP "</pre>\n";
282			$inpre = 0;
283			}
284
285	ph10	99	# Add <br> to the end of a non-literal line if we are within .nf/.fi
286	ph10	111
287			$_ .= "<br>\n" if (!$inpre && $innf);
288
289			print TEMP;
290			$wrotetext = 1;
291			}
292
293	ph10	97	# The TOC, if present, will have been written - terminate it
294
295			print "</ul>\n" if ($toc);
296
297			# Copy the remainder to the standard output
298
299			close(TEMP);
300			open(TEMP, "/tmp/$$") \|\| die "Can't open /tmp/$$ for input\n";
301	ph10	111
302	ph10	97	print while (<TEMP>);
303
304			print <<End ;
305			<p>
306			Return to the <a href="index.html">PCRE index page</a>.
307			</p>
308			End
309
310			close(TEMP);
311			unlink("/tmp/$$");
312
313			# End