3 # Man page to help file converter
4 # Copyright (C) 1994, 1995, 1998, 2000, 2001, 2002, 2003, 2004, 2005,
5 # 2007, 2010 Free Software Foundation, Inc.
7 # Originally written by:
8 # 2002 Andrew V. Samoilov
12 # Completely rewritten in Perl by:
13 # 2010 Alexandr Prenko
15 # This program is free software; you can redistribute it and/or modify
16 # it under the terms of the GNU General Public License as published by
17 # the Free Software Foundation; either version 2 of the License, or
18 # (at your option) any later version.
20 # This program is distributed in the hope that it will be useful,
21 # but WITHOUT ANY WARRANTY; without even the implied warranty of
22 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
23 # GNU General Public License for more details.
25 # You should have received a copy of the GNU General Public License
26 # along with this program; if not, write to the Free Software
27 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
30 # \brief Source: man page to help file converter
35 # Perl have no static variables, so this hash emulates them
37 "string_len anchor_flag" => 0,
38 "string_len lc_link_flag" => 0,
39 "handle_link old" => undef
43 my $CHAR_LINK_START = chr(01); # Ctrl-A
44 my $CHAR_LINK_POINTER = chr(02); # Ctrl-B
45 my $CHAR_LINK_END = chr(03); # Ctrl-C
46 my $CHAR_NODE_END = chr(04); # Ctrl-D
47 my $CHAR_ALTERNATE = chr(05); # Ctrl-E
48 my $CHAR_NORMAL = chr(06); # Ctrl-F
49 my $CHAR_VERSION = chr(07); # Ctrl-G
50 my $CHAR_FONT_BOLD = chr(010); # Ctrl-H
51 my $CHAR_FONT_NORMAL = chr(013); # Ctrl-K
52 my $CHAR_FONT_ITALIC = chr(024); # Ctrl-T
55 my $col = 0; # Current output column
56 my $out_row = 1; # Current output row
57 my $in_row = 0; # Current input row
58 my $no_split_flag = 0; # Flag: Don't split section on next ".SH"
59 my $skip_flag = 0; # Flag: Skip this section.
62 # 2 = title skipped, skipping text
63 my $link_flag = 0; # Flag: Next line is a link
64 my $verbatim_flag = 0; # Flag: Copy input to output verbatim
65 my $node = 0; # Flag: This line is an original ".SH"
67 my $c_out; # Output filename
68 my $f_out; # Output file
70 my $c_in; # Current input filename
72 my $indentation; # Indentation level, n spaces
73 my $tp_flag; # Flag: .TP paragraph
74 # 1 = this line is .TP label,
75 # 2 = first line of label description.
82 my ($str, $chars) = @_;
84 if (! defined $chars || $chars eq "")
91 $str = $strtok unless defined $str;
92 return undef unless defined $str;
95 $str =~ s/^[$chars]+//;
96 ($result, $strtok) = split /[$chars]+/, $str, 2;
97 ($result, $strtok) = split /[$chars]+/, $strtok, 2 if defined $result && $result eq "";
98 $strtok = undef if ! defined $strtok || $strtok eq "";
104 "node" => undef, # Section name
105 "lname" => undef, # Translated .SH, undef if not translated
107 "heading_level" => undef
111 my $nodes = struct_node();
112 my $cnode; # Current node
114 # Report error in input
118 warn sprintf "man2hlp: %s in file \"%s\" on line %d\n", $message, $c_in, $in_row;
121 # Do open, exit if it fails
124 my ($mode, $filename) = @_;
127 unless (open $f, $mode, $filename)
129 warn sprintf("man2hlp: Cannot open file \"%s\" ($!)\n", $filename);
135 # Do close, exit if it fails
141 warn "man2hlp: Cannot close file ($!)\n";
154 # Calculate the length of string
158 my $anchor_flag = \$static{"string_len anchor_flag"}; # Flag: Inside hypertext anchor name ho4u_v_Ariom
159 my $lc_link_flag = \$static{"string_len lc_link_flag"}; # Flag: Inside hypertext link target name
160 my $backslash_flag = 0; # Flag: Backslash quoting
161 my $len = 0; # Result: the length of the string
164 foreach my $c (split //, $buffer)
166 if ($c eq $CHAR_LINK_POINTER)
168 $$lc_link_flag = 1; # Link target name starts
170 elsif ($c eq $CHAR_LINK_END)
172 $$lc_link_flag = 0; # Link target name ends
174 elsif ($c eq $CHAR_NODE_END)
176 # Node anchor name starts
178 # Ugly hack to prevent loss of one space
181 # Don't add control characters to the length
182 next if ord($c) >= 0 && ord($c) < 32;
183 # Attempt to handle backslash quoting
184 if ($c eq '\\' && !$backslash_flag)
190 # Increase length if not inside anchor name or link target name
191 $len++ if !$$anchor_flag && !$$lc_link_flag;
192 if ($$anchor_flag && $c eq ']')
194 # Node anchor name ends
205 my $len; # The length of current word
206 my $backslash_flag = 0;
209 return if $skip_flag;
213 # Attempt to handle backslash quoting
214 foreach (split //, $buffer)
216 if ($_ eq '\\' && !$backslash_flag)
228 $buffer = strtok($buffer, " \t\n");
229 # Repeat for each word
230 while (defined $buffer)
235 $len = string_len($buffer);
236 # Words are separated by spaces
244 print $f_out ' ' while $col++ < $indentation;
246 # Attempt to handle backslash quoting
247 foreach (split //, $buffer)
249 if ($_ eq '\\' && !$backslash_flag)
261 $buffer = strtok(undef, " \t\n");
266 # Like print_string but with printf-like syntax
269 print_string sprintf shift, @_;
272 # Handle NODE and .SH commands. is_sh is 1 for .SH, 0 for NODE
273 # FIXME: Consider to remove first parameter
276 my ($buffer, $is_sh) = @_;
277 my ($len, $heading_level);
279 # If we already skipped a section, don't skip another
280 $skip_flag = 0 if $skip_flag == 2;
282 # Get the command parameters
283 $buffer = strtok(undef, "");
284 if (! defined $buffer)
286 print_error "Syntax error: .SH: no title";
292 $buffer =~ s/^"// and $buffer =~ s/"$//;
293 # Calculate heading level
295 $heading_level++ while substr($buffer, $heading_level, 1) eq ' ';
296 # Heading level must be even
297 if ($heading_level % 2)
299 print_error "Syntax error: .SH: odd heading level";
303 # Don't start a new section
305 print_string $buffer;
312 # Skipping title and marking text for skipping
317 $buffer = substr($buffer, $heading_level);
318 if (! $is_sh || ! $node)
320 # Start a new section, but omit empty section names
323 printf $f_out "%s[%s]", $CHAR_NODE_END, $buffer;
327 # Add section to the linked list
328 if (! defined $cnode)
334 $cnode->{'next'} = struct_node();
335 $cnode = $cnode->{'next'};
337 $cnode->{'node'} = $buffer;
338 $cnode->{'lname'} = undef;
339 $cnode->{'next'} = undef;
340 $cnode->{'heading_level'} = $heading_level;
344 $cnode->{'lname'} = $buffer;
345 print_string $buffer;
349 } # Start new section
354 # Convert character from the macro name to the font marker
359 'R' => $CHAR_FONT_NORMAL,
360 'B' => $CHAR_FONT_BOLD,
361 'I' => $CHAR_FONT_ITALIC
363 return exists $font{$c} ? $font{$c} : chr(0);
367 # Handle alternate font commands (.BR, .IR, .RB, .RI, .BI, .IB)
368 # Return 0 if the command wasn't recognized, 1 otherwise
370 sub handle_alt_font($)
376 return 0 if length($buffer) != 3;
377 return 0 if substr($buffer, 0, 1) ne '.';
380 char_to_font substr($buffer, 1, 1),
381 char_to_font substr($buffer, 2, 1)
384 # Exclude names with unknown characters, .BB, .II and .RR
385 if ($font[0] eq chr(0) || $font[1] eq chr(0) || $font[0] eq $font[1])
390 my $p = strtok(undef, "");
391 return 1 unless defined $p;
395 my @p = split //, $p;
401 $in_quotes = !$in_quotes;
406 if ($p[0] eq ' ' && !$in_quotes)
409 # Don't change font if we are at the end
412 $alt_state = $alt_state ? 0 : 1;
413 $buffer .= $font[$alt_state];
417 shift @p while @p && $p[0] eq ' ';
425 # Turn off attributes if necessary
426 if ($font[$alt_state] ne $CHAR_FONT_NORMAL)
428 $buffer .= $CHAR_FONT_NORMAL;
431 print_string $buffer;
436 # Handle .IP and .TP commands. is_tp is 1 for .TP, 0 for .IP
453 # Handle all the roff dot commands. See man groff_man for details
454 sub handle_command($)
459 # Get the command name
460 $buffer = strtok($buffer, " \t");
462 if ($buffer eq ".SH")
465 handle_node $buffer, 1;
467 elsif ($buffer eq ".\\\"NODE")
469 handle_node $buffer, 0;
471 elsif ($buffer eq ".\\\"DONT_SPLIT\"")
475 elsif ($buffer eq ".\\\"SKIP_SECTION\"")
479 elsif ($buffer eq ".\\\"LINK2\"")
481 # Next two input lines form a link
484 elsif ($buffer eq ".PP" || $buffer eq ".P" || $buffer eq ".LP")
491 elsif ($buffer eq ".nf")
493 # Following input lines are to be handled verbatim
497 elsif ($buffer eq ".I" || $buffer eq ".B" || $buffer eq ".SB")
499 # Bold text or italics text
500 my $backslash_flag = 0;
503 # Causes the text on the same line or the text on the
504 # next line to appear in boldface font, one point
505 # size smaller than the default font.
508 # FIXME: text is optional, so there is no error
510 my $p = strtok(undef, "");
513 print_error "Syntax error: .I | .B | .SB : no text";
517 $buffer = substr($buffer, 1, 1) eq 'I' ? $CHAR_FONT_ITALIC : $CHAR_FONT_BOLD;
519 # Attempt to handle backslash quoting
520 foreach (split //, $p)
522 if ($_ eq '\\' && !$backslash_flag)
530 print_string $buffer . $CHAR_FONT_NORMAL;
532 elsif ($buffer eq ".TP")
536 elsif ($buffer eq ".IP")
540 elsif ($buffer eq ".\\\"TOPICS")
544 print_error "Syntax error: .\\\"TOPICS must be first command";
547 $buffer = strtok(undef, "");
548 if (! defined $buffer)
550 print_error "Syntax error: .\\\"TOPICS: no text";
554 $buffer =~ s/^"// and $buffer =~ s/"$//;
557 elsif ($buffer eq ".br")
561 elsif ($buffer =~ /^\.\\"/)
563 # Comment { Hello from K.O. ;-) }
565 elsif ($buffer eq ".TH")
569 elsif ($buffer eq ".SM")
571 # Causes the text on the same line or the text on the
572 # next line to appear in a font that is one point
573 # size smaller than the default font.
574 $buffer = strtok(undef, "");
575 print_string $buffer if defined $buffer;
577 elsif (handle_alt_font($buffer) == 1)
583 # Other commands are ignored
584 print_error sprintf "Warning: unsupported command %s", $buffer;
592 'linkname' => undef, # Section name
593 'line' => undef, # Input line in ...
599 my $links = struct_links();
606 my $old = \$static{"handle_link old"};
613 # Old format link, not supported
615 elsif ($link_flag == 2)
617 # First part of new format link
618 # Bold text or italics text
619 if (substr($buffer, 0, 2) eq '.I' || substr($buffer, 0, 2) eq '.B')
621 $buffer =~ s/^..[\s\t]*//;
627 elsif ($link_flag == 3)
629 # Second part of new format link
635 # "Layout\&)," -- "Layout" should be highlighted, but not "),"
636 ($$old, $amp_arg) = split /\\&/, $$old, 2;
637 $amp_arg = "" unless defined $amp_arg;
638 printf_string "%s%s%s%s%s%s\n", $CHAR_LINK_START, $$old,
639 $CHAR_LINK_POINTER, $buffer, $CHAR_LINK_END, $amp_arg;
641 # Add to the linked list
642 if (defined $current_link)
644 $current_link->{'next'} = struct_links();
645 $current_link = $current_link->{'next'};
646 $current_link->{'next'} = undef;
650 $current_link = $links;
652 $current_link->{'linkname'} = $buffer;
653 $current_link->{'filename'} = $c_in;
654 $current_link->{'line'} = $in_row;
660 my $len; # Length of input line
661 my $c_man; # Manual filename
662 my $c_tmpl; # Template filename
663 my $f_man; # Manual file
664 my $f_tmpl; # Template file
665 my $buffer; # Full input line
667 my $outfile_buffer; # Large buffer to keep the output file
668 my $cont_start; # Start of [Contents]
669 my $file_end; # Length of the output file
671 # Validity check for arguments
674 warn "Usage: man2hlp file.man template_file helpfile\n";
682 # First stage - process the manual, write to the output file
684 $f_man = fopen_check "<", $c_man;
685 $f_out = fopen_check ">", $c_out;
688 # Repeat for each input line
691 # Remove terminating newline
694 my $input_line; # Input line without initial "\&"
696 if (substr($buffer, 0, 2) eq '\\&')
698 $input_line = substr($buffer, 2);
702 $input_line = $buffer;
706 $len = length($input_line);
710 # Copy the line verbatim
711 if ($input_line eq ".fi")
717 print_string $input_line;
724 handle_link $input_line;
726 elsif (substr($buffer, 0, 1) eq '.')
728 # The line is a roff command
729 handle_command $input_line;
733 #A normal line, just output it
734 print_string $input_line;
736 # .TP label processed as usual line
747 if ($col >= $indentation)
753 print $f_out " " while ++$col < $indentation;
761 # First stage ends here, closing the manual
763 # Second stage - process the template file
764 $f_tmpl = fopen_check "<", $c_tmpl;
767 # Repeat for each input line
772 if (defined $lc_node)
776 $cnode->{'lname'} = $buffer;
777 chomp $cnode->{'lname'};
783 my $char_node_end = index($buffer, $CHAR_NODE_END);
784 $lc_node = $char_node_end < 0 ? undef : substr($buffer, $char_node_end);
786 if (defined $lc_node && substr($lc_node, 1, 1) eq '[')
788 my $p = index($lc_node, ']');
790 if (substr($lc_node, 1, 6) eq '[main]')
796 if (! defined $cnode)
802 $cnode->{'next'} = struct_node();
803 $cnode = $cnode->{'next'};
805 $cnode->{'node'} = substr($lc_node, 2, $p-2);
806 $cnode->{'lname'} = undef;
807 $cnode->{'next'} = undef;
808 $cnode->{'heading_level'} = 0;
821 print $f_out $buffer;
824 $cont_start = tell $f_out;
825 if ($cont_start <= 0)
833 printf $f_out "\004[Contents]\n%s\n\n", $topics;
837 print $f_out "\004[Contents]\n";
840 for ($current_link = $links; defined $current_link && defined $current_link->{'linkname'};)
843 my $next = $current_link->{'next'};
845 if ($current_link->{'linkname'} eq "Contents")
851 for ($cnode = $nodes; defined $cnode && defined $cnode->{'node'}; $cnode = $cnode->{'next'})
853 if ($cnode->{'node'} eq $current_link->{'linkname'})
862 $buffer = sprintf "Stale link \"%s\"", $current_link->{'linkname'};
863 $c_in = $current_link->{'filename'};
864 $in_row = $current_link->{'line'};
868 $current_link = $next;
871 for ($cnode = $nodes; defined $cnode && defined $cnode->{'node'};)
873 my $next = $cnode->{'next'};
874 $lc_node = $cnode->{'node'};
876 if (defined $lc_node && $lc_node ne '') {
877 printf $f_out " %*s\001%s\002%s\003", $cnode->{'heading_level'},
878 "", $cnode->{'lname'} ? $cnode->{'lname'} : $lc_node, $lc_node;
884 $file_end = tell $f_out;
887 if (($file_end <= 0) || ($file_end - $cont_start <= 0))
894 fclose_check $f_tmpl;
895 # Second stage ends here, closing all files, note the end of output
898 # Third stage - swap two parts of the output file.
899 # First, open the output file for reading and load it into the memory.
901 $outfile_buffer = '';
902 $f_out = fopen_check '<', $c_out;
903 $outfile_buffer .= $_ while <$f_out>;
905 # Now the output file is in the memory
907 # Again open output file for writing
908 $f_out = fopen_check '>', $c_out;
910 # Write part after the "Contents" node
911 print $f_out substr($outfile_buffer, $cont_start, $file_end - $cont_start);
913 # Write part before the "Contents" node
914 print $f_out substr($outfile_buffer, 0, $cont_start-1);