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
11 # Completely rewriten on perl by:
12 # 2010 Alexandr Prenko
14 # This program is free software; you can redistribute it and/or modify
15 # it under the terms of the GNU General Public License as published by
16 # the Free Software Foundation; either version 2 of the License, or
17 # (at your option) any later version.
19 # This program is distributed in the hope that it will be useful,
20 # but WITHOUT ANY WARRANTY; without even the implied warranty of
21 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
22 # GNU General Public License for more details.
24 # You should have received a copy of the GNU General Public License
25 # along with this program; if not, write to the Free Software
26 # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
29 # \brief Source: man page to help file converter
34 # Perl have no static variables, so this hash emulates them
36 "string_len anchor_flag" => 0,
37 "string_len lc_link_flag" => 0,
38 "handle_link old" => undef
42 my $CHAR_LINK_START = chr(01); # Ctrl-A
43 my $CHAR_LINK_POINTER = chr(02); # Ctrl-B
44 my $CHAR_LINK_END = chr(03); # Ctrl-C
45 my $CHAR_NODE_END = chr(04); # Ctrl-D
46 my $CHAR_ALTERNATE = chr(05); # Ctrl-E
47 my $CHAR_NORMAL = chr(06); # Ctrl-F
48 my $CHAR_VERSION = chr(07); # Ctrl-G
49 my $CHAR_FONT_BOLD = chr(010); # Ctrl-H
50 my $CHAR_FONT_NORMAL = chr(013); # Ctrl-K
51 my $CHAR_FONT_ITALIC = chr(024); # Ctrl-T
54 my $col = 0; # Current output column
55 my $out_row = 1; # Current output row
56 my $in_row = 0; # Current input row
57 my $no_split_flag = 0; # Flag: Don't split section on next ".SH"
58 my $skip_flag = 0; # Flag: Skip this section.
61 # 2 = title skipped, skipping text
62 my $link_flag = 0; # Flag: Next line is a link
63 my $verbatim_flag = 0; # Flag: Copy input to output verbatim
64 my $node = 0; # Flag: This line is an original ".SH"
66 my $c_out; # Output filename
67 my $f_out; # Output file
69 my $c_in; # Current input filename
71 my $indentation; # Indentation level, n spaces
72 my $tp_flag; # Flag: .TP paragraph
73 # 1 = this line is .TP label,
74 # 2 = first line of label description.
81 my ($str, $chars) = @_;
83 if (! defined $chars || $chars eq "")
90 $str = $strtok unless defined $str;
91 return undef unless defined $str;
94 $str =~ s/^[$chars]+//;
95 ($result, $strtok) = split /[$chars]+/, $str, 2;
96 ($result, $strtok) = split /[$chars]+/, $strtok, 2 if defined $result && $result eq "";
97 $strtok = undef if ! defined $strtok || $strtok eq "";
103 "node" => undef, # Section name
104 "lname" => undef, # Translated .SH, undef if not translated
106 "heading_level" => undef
110 my $nodes = struct_node();
111 my $cnode; # Current node
113 # Report error in input
117 warn sprintf "man2hlp: %s in file \"%s\" on line %d\n", $message, $c_in, $in_row;
120 # Do open, exit if it fails
123 my ($mode, $filename) = @_;
126 unless (open $f, $mode, $filename)
128 warn sprintf("man2hlp: Cannot open file \"%s\" ($!)\n", $filename);
134 # Do close, exit if it fails
140 warn "man2hlp: Cannot close file ($!)\n";
153 # Calculate the length of string
157 my $anchor_flag = \$static{"string_len anchor_flag"}; # Flag: Inside hypertext anchor name ho4u_v_Ariom
158 my $lc_link_flag = \$static{"string_len lc_link_flag"}; # Flag: Inside hypertext link target name
159 my $backslash_flag = 0; # Flag: Backslash quoting
160 my $len = 0; # Result: the length of the string
163 foreach my $c (split //, $buffer)
165 if ($c eq $CHAR_LINK_POINTER)
167 $$lc_link_flag = 1; # Link target name starts
169 elsif ($c eq $CHAR_LINK_END)
171 $$lc_link_flag = 0; # Link target name ends
173 elsif ($c eq $CHAR_NODE_END)
175 # Node anchor name starts
177 # Ugly hack to prevent loss of one space
180 # Don't add control characters to the length
181 next if ord($c) >= 0 && ord($c) < 32;
182 # Attempt to handle backslash quoting
183 if ($c eq '\\' && !$backslash_flag)
189 # Increase length if not inside anchor name or link target name
190 $len++ if !$$anchor_flag && !$$lc_link_flag;
191 if ($$anchor_flag && $c eq ']')
193 # Node anchor name ends
204 my $len; # The length of current word
205 my $backslash_flag = 0;
208 return if $skip_flag;
212 # Attempt to handle backslash quoting
213 foreach (split //, $buffer)
215 if ($_ eq '\\' && !$backslash_flag)
227 $buffer = strtok($buffer, " \t\n");
228 # Repeat for each word
229 while (defined $buffer)
234 $len = string_len($buffer);
235 # Words are separated by spaces
243 print $f_out ' ' while $col++ < $indentation;
245 # Attempt to handle backslash quoting
246 foreach (split //, $buffer)
248 if ($_ eq '\\' && !$backslash_flag)
260 $buffer = strtok(undef, " \t\n");
265 # Like print_string but with printf-like syntax
268 print_string sprintf shift, @_;
271 # Handle NODE and .SH commands. is_sh is 1 for .SH, 0 for NODE
272 # FIXME: Consider to remove first parameter
275 my ($buffer, $is_sh) = @_;
276 my ($len, $heading_level);
278 # If we already skipped a section, don't skip another
279 $skip_flag = 0 if $skip_flag == 2;
281 # Get the command parameters
282 $buffer = strtok(undef, "");
283 if (! defined $buffer)
285 print_error "Syntax error: .SH: no title";
291 $buffer =~ s/^"// and $buffer =~ s/"$//;
292 # Calculate heading level
294 $heading_level++ while substr($buffer, $heading_level, 1) eq ' ';
295 # Heading level must be even
296 if ($heading_level % 2)
298 print_error "Syntax error: .SH: odd heading level";
302 # Don't start a new section
304 print_string $buffer;
311 # Skipping title and marking text for skipping
316 $buffer = substr($buffer, $heading_level);
317 if (! $is_sh || ! $node)
319 # Start a new section, but omit empty section names
322 printf $f_out "%s[%s]", $CHAR_NODE_END, $buffer;
326 # Add section to the linked list
327 if (! defined $cnode)
333 $cnode->{'next'} = struct_node();
334 $cnode = $cnode->{'next'};
336 $cnode->{'node'} = $buffer;
337 $cnode->{'lname'} = undef;
338 $cnode->{'next'} = undef;
339 $cnode->{'heading_level'} = $heading_level;
343 $cnode->{'lname'} = $buffer;
344 print_string $buffer;
348 } # Start new section
353 # Convert character from the macro name to the font marker
358 'R' => $CHAR_FONT_NORMAL,
359 'B' => $CHAR_FONT_BOLD,
360 'I' => $CHAR_FONT_ITALIC
362 return exists $font{$c} ? $font{$c} : chr(0);
366 # Handle alternate font commands (.BR, .IR, .RB, .RI, .BI, .IB)
367 # Return 0 if the command wasn't recognized, 1 otherwise
369 sub handle_alt_font($)
375 return 0 if length($buffer) != 3;
376 return 0 if substr($buffer, 0, 1) ne '.';
379 char_to_font substr($buffer, 1, 1),
380 char_to_font substr($buffer, 2, 1)
383 # Exclude names with unknown characters, .BB, .II and .RR
384 if ($font[0] eq chr(0) || $font[1] eq chr(0) || $font[0] eq $font[1])
389 my $p = strtok(undef, "");
390 return 1 unless defined $p;
394 my @p = split //, $p;
400 $in_quotes = !$in_quotes;
405 if ($p[0] eq ' ' && !$in_quotes)
408 # Don't change font if we are at the end
411 $alt_state = $alt_state ? 0 : 1;
412 $buffer .= $font[$alt_state];
416 shift @p while $p[0] eq ' ';
424 # Turn off attributes if necessary
425 if ($font[$alt_state] ne $CHAR_FONT_NORMAL)
427 $buffer .= $CHAR_FONT_NORMAL;
430 print_string $buffer;
435 # Handle .IP and .TP commands. is_tp is 1 for .TP, 0 for .IP
452 # Handle all the roff dot commands. See man groff_man for details
453 sub handle_command($)
458 # Get the command name
459 $buffer = strtok($buffer, " \t");
461 if ($buffer eq ".SH")
464 handle_node $buffer, 1;
466 elsif ($buffer eq ".\\\"NODE")
468 handle_node $buffer, 0;
470 elsif ($buffer eq ".\\\"DONT_SPLIT\"")
474 elsif ($buffer eq ".\\\"SKIP_SECTION\"")
478 elsif ($buffer eq ".\\\"LINK2\"")
480 # Next two input lines form a link
483 elsif ($buffer eq ".PP" || $buffer eq ".P" || $buffer eq ".LP")
490 elsif ($buffer eq ".nf")
492 # Following input lines are to be handled verbatim
496 elsif ($buffer eq ".I" || $buffer eq ".B" || $buffer eq ".SB")
498 # Bold text or italics text
499 my $backslash_flag = 0;
502 # Causes the text on the same line or the text on the
503 # next line to appear in boldface font, one point
504 # size smaller than the default font.
507 # FIXME: text is optional, so there is no error
509 my $p = strtok(undef, "");
512 print_error "Syntax error: .I | .B | .SB : no text";
516 $buffer = substr($buffer, 1, 1) eq 'I' ? $CHAR_FONT_ITALIC : $CHAR_FONT_BOLD;
518 # Attempt to handle backslash quoting
519 foreach (split //, $p)
521 if ($_ eq '\\' && !$backslash_flag)
529 print_string $buffer . $CHAR_FONT_NORMAL;
531 elsif ($buffer eq ".TP")
535 elsif ($buffer eq ".IP")
539 elsif ($buffer eq ".\\\"TOPICS")
543 print_error "Syntax error: .\\\"TOPICS must be first command";
546 $buffer = strtok(undef, "");
547 if (! defined $buffer)
549 print_error "Syntax error: .\\\"TOPICS: no text";
553 $buffer =~ s/^"// and $buffer =~ s/"$//;
556 elsif ($buffer eq ".br")
560 elsif ($buffer =~ /^\.\\"/)
562 # Comment { Hello from K.O. ;-) }
564 elsif ($buffer eq ".TH")
568 elsif ($buffer eq ".SM")
570 # Causes the text on the same line or the text on the
571 # next line to appear in a font that is one point
572 # size smaller than the default font.
573 $buffer = strtok(undef, "");
574 print_string $buffer if defined $buffer;
576 elsif (handle_alt_font($buffer) == 1)
582 # Other commands are ignored
583 print_error sprintf "Warning: unsupported command %s", $buffer;
591 'linkname' => undef, # Section name
592 'line' => undef, # Input line in ...
598 my $links = struct_links();
605 my $old = \$static{"handle_link old"};
612 # Old format link, not supported
614 elsif ($link_flag == 2)
616 # First part of new format link
617 # Bold text or italics text
618 if (substr($buffer, 0, 2) eq '.I' || substr($buffer, 0, 2) eq '.B')
620 $buffer =~ s/^..[\s\t]*//;
626 elsif ($link_flag == 3)
628 # Second part of new format link
634 # "Layout\&)," -- "Layout" should be highlighted, but not "),"
635 ($$old, $amp_arg) = split /\\&/, $$old, 2;
636 $amp_arg = "" unless defined $amp_arg;
637 printf_string "%s%s%s%s%s%s\n", $CHAR_LINK_START, $$old,
638 $CHAR_LINK_POINTER, $buffer, $CHAR_LINK_END, $amp_arg;
640 # Add to the linked list
641 if (defined $current_link)
643 $current_link->{'next'} = struct_links();
644 $current_link = $current_link->{'next'};
645 $current_link->{'next'} = undef;
649 $current_link = $links;
651 $current_link->{'linkname'} = $buffer;
652 $current_link->{'filename'} = $c_in;
653 $current_link->{'line'} = $in_row;
659 my $len; # Length of input line
660 my $c_man; # Manual filename
661 my $c_tmpl; # Template filename
662 my $f_man; # Manual file
663 my $f_tmpl; # Template file
664 my $buffer; # Full input line
666 my $outfile_buffer; # Large buffer to keep the output file
667 my $cont_start; # Start of [Contents]
668 my $file_end; # Length of the output file
670 # Validity check for arguments
673 warn "Usage: man2hlp file.man template_file helpfile\n";
681 # First stage - process the manual, write to the output file
683 $f_man = fopen_check "<", $c_man;
684 $f_out = fopen_check ">", $c_out;
687 # Repeat for each input line
690 # Remove terminating newline
693 my $input_line; # Input line without initial "\&"
695 if (substr($buffer, 0, 2) eq '\\&')
697 $input_line = substr($buffer, 2);
701 $input_line = $buffer;
705 $len = length($input_line);
709 # Copy the line verbatim
710 if ($input_line eq ".fi")
716 print_string $input_line;
723 handle_link $input_line;
725 elsif (substr($buffer, 0, 1) eq '.')
727 # The line is a roff command
728 handle_command $input_line;
732 #A normal line, just output it
733 print_string $input_line;
735 # .TP label processed as usual line
746 if ($col >= $indentation)
752 print $f_out " " while ++$col < $indentation;
760 # First stage ends here, closing the manual
762 # Second stage - process the template file
763 $f_tmpl = fopen_check "<", $c_tmpl;
766 # Repeat for each input line
771 if (defined $lc_node)
775 $cnode->{'lname'} = $buffer;
776 chomp $cnode->{'lname'};
782 my $char_node_end = index($buffer, $CHAR_NODE_END);
783 $lc_node = $char_node_end < 0 ? undef : substr($buffer, $char_node_end);
785 if (defined $lc_node && substr($lc_node, 1, 1) eq '[')
787 my $p = index($lc_node, ']');
789 if (substr($lc_node, 1, 6) eq '[main]')
795 if (! defined $cnode)
801 $cnode->{'next'} = struct_node();
802 $cnode = $cnode->{'next'};
804 $cnode->{'node'} = substr($lc_node, 2, $p-2);
805 $cnode->{'lname'} = undef;
806 $cnode->{'next'} = undef;
807 $cnode->{'heading_level'} = 0;
820 print $f_out $buffer;
823 $cont_start = tell $f_out;
824 if ($cont_start <= 0)
832 printf $f_out "\004[Contents]\n%s\n\n", $topics;
836 print $f_out "\004[Contents]\n";
839 for ($current_link = $links; defined $current_link && defined $current_link->{'linkname'};)
842 my $next = $current_link->{'next'};
844 if ($current_link->{'linkname'} eq "Contents")
850 for ($cnode = $nodes; defined $cnode && defined $cnode->{'node'}; $cnode = $cnode->{'next'})
852 if ($cnode->{'node'} eq $current_link->{'linkname'})
861 $buffer = sprintf "Stale link \"%s\"", $current_link->{'linkname'};
862 $c_in = $current_link->{'filename'};
863 $in_row = $current_link->{'line'};
867 $current_link = $next;
870 for ($cnode = $nodes; defined $cnode && defined $cnode->{'node'};)
872 my $next = $cnode->{'next'};
873 $lc_node = $cnode->{'node'};
875 if (defined $lc_node && $lc_node ne '') {
876 printf $f_out " %*s\001%s\002%s\003", $cnode->{'heading_level'},
877 "", $cnode->{'lname'} ? $cnode->{'lname'} : $lc_node, $lc_node;
883 $file_end = tell $f_out;
886 if (($file_end <= 0) || ($file_end - $cont_start <= 0))
893 fclose_check $f_tmpl;
894 # Second stage ends here, closing all files, note the end of output
897 # Third stage - swap two parts of the output file.
898 # First, open the output file for reading and load it into the memory.
900 $outfile_buffer = '';
901 $f_out = fopen_check '<', $c_out;
902 $outfile_buffer .= $_ while <$f_out>;
904 # Now the output file is in the memory
906 # Again open output file for writing
907 $f_out = fopen_check '>', $c_out;
909 # Write part after the "Contents" node
910 print $f_out substr($outfile_buffer, $cont_start, $file_end - $cont_start);
912 # Write part before the "Contents" node
913 print $f_out substr($outfile_buffer, 0, $cont_start-1);