3 # Man page to help file converter
4 # Copyright (C) 1994-2024
5 # The Free Software Foundation, Inc.
7 # Originally written by:
8 # Andrew V. Samoilov, 2002
10 # Andrew Borodin <aborodin@vmail.ru>, 2010
12 # Completely rewritten in Perl by:
13 # Alexandr Prenko, 2010
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 3 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, see <http://www.gnu.org/licenses/>.
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;
206 my $font_change_flag = 0;
210 return if $skip_flag;
214 # Attempt to handle backslash quoting
215 foreach (split //, $buffer)
217 if ($_ eq '\\' && !$backslash_flag)
229 $buffer = strtok($buffer, " \t\n");
230 # Repeat for each word
231 while (defined $buffer)
236 $len = string_len($buffer);
237 # Words are separated by spaces
245 print $f_out ' ' while $col++ < $indentation;
247 # Attempt to handle backslash quoting
248 foreach (split //, $buffer)
250 # handle quotes: \(lq, \(rq, \(dq
251 if ($quotes_flag != 0)
253 if (($_ eq 'l' || $_ eq 'r' || $_ eq 'd') && $quotes_flag == 1)
255 # continue quotes handling
259 elsif ($_ eq 'q' && $quotes_flag == 2)
261 # finish quotes handling
268 print $f_out '(' . $_;
269 print_error "Syntax error: unsupported \\(" . $_ . " command";
272 # handle \fR, \fB, \fI and \fP commands
273 if ($font_change_flag)
277 print $f_out $CHAR_FONT_BOLD;
281 print $f_out $CHAR_FONT_ITALIC;
283 elsif ($_ eq 'R' || $_ eq 'P')
285 print $f_out $CHAR_FONT_NORMAL;
289 print $f_out 'f' . $_;
290 print_error "Syntax error: unsupported \\f" . $_ . " command";
293 $font_change_flag = 0;
296 if ($_ eq '(' && $backslash_flag)
302 if ($_ eq 'f' && $backslash_flag)
304 $font_change_flag = 1;
308 if ($_ eq '\\' && !$backslash_flag)
314 $font_change_flag = 0;
322 $buffer = strtok(undef, " \t\n");
327 # Like print_string but with printf-like syntax
330 print_string sprintf shift, @_;
333 # Handle NODE and .SH commands. is_sh is 1 for .SH, 0 for NODE
334 # FIXME: Consider to remove first parameter
337 my ($buffer, $is_sh) = @_;
338 my ($len, $heading_level);
340 # If we already skipped a section, don't skip another
341 $skip_flag = 0 if $skip_flag == 2;
343 # Get the command parameters
344 $buffer = strtok(undef, "");
345 if (! defined $buffer)
347 print_error "Syntax error: .SH: no title";
353 $buffer =~ s/^"// and $buffer =~ s/"$//;
354 # Calculate heading level
356 $heading_level++ while substr($buffer, $heading_level, 1) eq ' ';
357 # Heading level must be even
358 if ($heading_level % 2)
360 print_error "Syntax error: .SH: odd heading level";
364 # Don't start a new section
366 print_string $buffer;
373 # Skipping title and marking text for skipping
378 $buffer = substr($buffer, $heading_level);
379 if (! $is_sh || ! $node)
381 # Start a new section, but omit empty section names
384 printf $f_out "%s[%s]", $CHAR_NODE_END, $buffer;
388 # Add section to the linked list
389 if (! defined $cnode)
395 $cnode->{'next'} = struct_node();
396 $cnode = $cnode->{'next'};
398 $cnode->{'node'} = $buffer;
399 $cnode->{'lname'} = undef;
400 $cnode->{'next'} = undef;
401 $cnode->{'heading_level'} = $heading_level;
405 $cnode->{'lname'} = $buffer;
406 print_string $buffer;
410 } # Start new section
415 # Convert character from the macro name to the font marker
420 'R' => $CHAR_FONT_NORMAL,
421 'B' => $CHAR_FONT_BOLD,
422 'I' => $CHAR_FONT_ITALIC
424 return exists $font{$c} ? $font{$c} : chr(0);
428 # Handle alternate font commands (.BR, .IR, .RB, .RI, .BI, .IB)
429 # Return 0 if the command wasn't recognized, 1 otherwise
431 sub handle_alt_font($)
437 return 0 if length($buffer) != 3;
438 return 0 if substr($buffer, 0, 1) ne '.';
441 char_to_font substr($buffer, 1, 1),
442 char_to_font substr($buffer, 2, 1)
445 # Exclude names with unknown characters, .BB, .II and .RR
446 if ($font[0] eq chr(0) || $font[1] eq chr(0) || $font[0] eq $font[1])
451 my $p = strtok(undef, "");
452 return 1 unless defined $p;
456 my @p = split //, $p;
462 $in_quotes = !$in_quotes;
467 if ($p[0] eq ' ' && !$in_quotes)
470 # Don't change font if we are at the end
473 $alt_state = $alt_state ? 0 : 1;
474 $buffer .= $font[$alt_state];
478 shift @p while @p && $p[0] eq ' ';
486 # Turn off attributes if necessary
487 if ($font[$alt_state] ne $CHAR_FONT_NORMAL)
489 $buffer .= $CHAR_FONT_NORMAL;
492 print_string $buffer;
497 # Handle .IP and .TP commands. is_tp is 1 for .TP, 0 for .IP
514 # Handle all the roff dot commands. See man groff_man for details
515 sub handle_command($)
520 # Get the command name
521 $buffer = strtok($buffer, " \t");
523 if ($buffer eq ".SH")
526 handle_node $buffer, 1;
528 elsif ($buffer eq ".\\\"NODE")
530 handle_node $buffer, 0;
532 elsif ($buffer eq ".\\\"DONT_SPLIT\"")
536 elsif ($buffer eq ".\\\"SKIP_SECTION\"")
540 elsif ($buffer eq ".\\\"LINK2\"")
542 # Next two input lines form a link
545 elsif ($buffer eq ".PP" || $buffer eq ".P" || $buffer eq ".LP")
552 elsif ($buffer eq ".nf")
554 # Following input lines are to be handled verbatim
558 elsif ($buffer eq ".I" || $buffer eq ".B" || $buffer eq ".SB")
560 # Bold text or italics text
561 my $backslash_flag = 0;
564 # Causes the text on the same line or the text on the
565 # next line to appear in boldface font, one point
566 # size smaller than the default font.
569 # FIXME: text is optional, so there is no error
571 my $p = strtok(undef, "");
574 print_error "Syntax error: .I | .B | .SB : no text";
578 $buffer = substr($buffer, 1, 1) eq 'I' ? $CHAR_FONT_ITALIC : $CHAR_FONT_BOLD;
580 # Attempt to handle backslash quoting
581 foreach (split //, $p)
583 if ($_ eq '\\' && !$backslash_flag)
591 print_string $buffer . $CHAR_FONT_NORMAL;
593 elsif ($buffer eq ".TP")
597 elsif ($buffer eq ".IP")
601 elsif ($buffer eq ".\\\"TOPICS")
605 print_error "Syntax error: .\\\"TOPICS must be first command";
608 $buffer = strtok(undef, "");
609 if (! defined $buffer)
611 print_error "Syntax error: .\\\"TOPICS: no text";
615 $buffer =~ s/^"// and $buffer =~ s/"$//;
618 elsif ($buffer eq ".br")
622 elsif ($buffer =~ /^\.\\"/)
624 # Comment { Hello from K.O. ;-) }
626 elsif ($buffer eq ".TH")
630 elsif ($buffer eq ".SM")
632 # Causes the text on the same line or the text on the
633 # next line to appear in a font that is one point
634 # size smaller than the default font.
635 $buffer = strtok(undef, "");
636 print_string $buffer if defined $buffer;
638 elsif (handle_alt_font($buffer) == 1)
642 elsif ($buffer eq ".RE")
648 # Other commands are ignored
649 print_error sprintf "Warning: unsupported command %s", $buffer;
657 'linkname' => undef, # Section name
658 'line' => undef, # Input line in ...
664 my $links = struct_links();
671 my $old = \$static{"handle_link old"};
678 # Old format link, not supported
680 elsif ($link_flag == 2)
682 # First part of new format link
683 # Bold text or italics text
684 if (substr($buffer, 0, 2) eq '.I' || substr($buffer, 0, 2) eq '.B')
686 $buffer =~ s/^..[\s\t]*//;
692 elsif ($link_flag == 3)
694 # Second part of new format link
700 # "Layout\&)," -- "Layout" should be highlighted, but not "),"
701 ($$old, $amp_arg) = split /\\&/, $$old, 2;
702 $amp_arg = "" unless defined $amp_arg;
703 printf_string "%s%s%s%s%s%s\n", $CHAR_LINK_START, $$old,
704 $CHAR_LINK_POINTER, $buffer, $CHAR_LINK_END, $amp_arg;
706 # Add to the linked list
707 if (defined $current_link)
709 $current_link->{'next'} = struct_links();
710 $current_link = $current_link->{'next'};
711 $current_link->{'next'} = undef;
715 $current_link = $links;
717 $current_link->{'linkname'} = $buffer;
718 $current_link->{'filename'} = $c_in;
719 $current_link->{'line'} = $in_row;
725 my $len; # Length of input line
726 my $c_man; # Manual filename
727 my $c_tmpl; # Template filename
728 my $f_man; # Manual file
729 my $f_tmpl; # Template file
730 my $buffer; # Full input line
732 my $outfile_buffer; # Large buffer to keep the output file
733 my $cont_start; # Start of [Contents]
734 my $file_end; # Length of the output file
736 # Validity check for arguments
739 warn "Usage: man2hlp file.man template_file helpfile\n";
747 # First stage - process the manual, write to the output file
749 $f_man = fopen_check "<", $c_man;
750 $f_out = fopen_check ">", $c_out;
753 # Repeat for each input line
756 # Remove terminating newline
759 my $input_line; # Input line without initial "\&"
761 if (substr($buffer, 0, 2) eq '\\&')
763 $input_line = substr($buffer, 2);
767 $input_line = $buffer;
771 $len = length($input_line);
775 # Copy the line verbatim
776 if ($input_line eq ".fi")
782 print_string $input_line;
789 handle_link $input_line;
791 elsif (substr($buffer, 0, 1) eq '.')
793 # The line is a roff command
794 handle_command $input_line;
798 #A normal line, just output it
799 print_string $input_line;
801 # .TP label processed as usual line
812 if ($col >= $indentation)
818 print $f_out " " while ++$col < $indentation;
826 # First stage ends here, closing the manual
828 # Second stage - process the template file
829 $f_tmpl = fopen_check "<", $c_tmpl;
832 # Repeat for each input line
837 if (defined $lc_node)
841 $cnode->{'lname'} = $buffer;
842 chomp $cnode->{'lname'};
848 my $char_node_end = index($buffer, $CHAR_NODE_END);
849 $lc_node = $char_node_end < 0 ? undef : substr($buffer, $char_node_end);
851 if (defined $lc_node && substr($lc_node, 1, 1) eq '[')
853 my $p = index($lc_node, ']');
855 if (substr($lc_node, 1, 6) eq '[main]')
861 if (! defined $cnode)
867 $cnode->{'next'} = struct_node();
868 $cnode = $cnode->{'next'};
870 $cnode->{'node'} = substr($lc_node, 2, $p-2);
871 $cnode->{'lname'} = undef;
872 $cnode->{'next'} = undef;
873 $cnode->{'heading_level'} = 0;
886 print $f_out $buffer;
889 $cont_start = tell $f_out;
890 if ($cont_start <= 0)
898 printf $f_out "\004[Contents]\n%s\n\n", $topics;
902 print $f_out "\004[Contents]\n";
905 for ($current_link = $links; defined $current_link && defined $current_link->{'linkname'};)
908 my $next = $current_link->{'next'};
910 if ($current_link->{'linkname'} eq "Contents")
916 for ($cnode = $nodes; defined $cnode && defined $cnode->{'node'}; $cnode = $cnode->{'next'})
918 if ($cnode->{'node'} eq $current_link->{'linkname'})
927 $buffer = sprintf "Stale link \"%s\"", $current_link->{'linkname'};
928 $c_in = $current_link->{'filename'};
929 $in_row = $current_link->{'line'};
933 $current_link = $next;
936 for ($cnode = $nodes; defined $cnode && defined $cnode->{'node'};)
938 my $next = $cnode->{'next'};
939 $lc_node = $cnode->{'node'};
941 if (defined $lc_node && $lc_node ne '') {
942 printf $f_out " %*s\001%s\002%s\003", $cnode->{'heading_level'},
943 "", $cnode->{'lname'} ? $cnode->{'lname'} : $lc_node, $lc_node;
949 $file_end = tell $f_out;
952 if (($file_end <= 0) || ($file_end - $cont_start <= 0))
959 fclose_check $f_tmpl;
960 # Second stage ends here, closing all files, note the end of output
963 # Third stage - swap two parts of the output file.
964 # First, open the output file for reading and load it into the memory.
966 $outfile_buffer = '';
967 $f_out = fopen_check '<', $c_out;
968 $outfile_buffer .= $_ while <$f_out>;
970 # Now the output file is in the memory
972 # Again open output file for writing
973 $f_out = fopen_check '>', $c_out;
975 # Write part after the "Contents" node
976 print $f_out substr($outfile_buffer, $cont_start, $file_end - $cont_start);
978 # Write part before the "Contents" node
979 print $f_out substr($outfile_buffer, 0, $cont_start-1);