#!/home/jc/bin/perl5 -dw
#
# txt2html [file]..
#
# This is a prototype program to read text and convert it to html.
# This program acts like a conventional Unix filter program: It reads
# the file(s) named on the command line, or stdin if there are no file
# names on the command line; it writes to stdout.
$| = 1; # Don't buffer output.
$mintxtlen = 20; # Shorter lines treated as non-text.
print "\n"; # Boilerplate.
@txt = <>; # Slurp up all the text.
for $line (@txt) {
$line =~ s"\s+$"$/"; # Trim away white stuff.
if ($line eq $/) { # Blank line?
$newpar = 1; # Treat as end of paragraph.
if ($mailhdrs) { # Are we processing mail headers?
$line = "
\n";
$mailhdrs = 0;
} else {
$line = "\n";
}
next;
}
if ($line =~ /^\s*[\-= ][\-= ]+\s*$/) {
$line = "
\n";
next;
}
# Check for things that look like nmbered or bulleted lists:
if ($line =~ /^\s*\w\s*[\-.:]\s*/) {
$line = "
" . $line;
# $line =~ s/(.*)\n$/
$1
\n/;
next;
}
if ($newpar) { # Check for formatted lines.
if (length($line) < $mintxtlen) { # Too short.
$line =~ s"$/$"
$/";
next;
}
# Check for all capitalized words to try to spot things that
# look like addresses or titles:
if (!($line =~ /\s[a-z]/)) {
$line =~ s"$/$"
$/";
next;
}
}
if (($lbl,$val) = ($line =~ /^(From\s|From:|Path:|Subject:)\s*(.*)/)) { # Mail hdr?
if ($line =~ "^(From|From:|Path:)\s*(.*)!(.+)!(.+)!(.+)") {
$line = "$1 $3!$4!$5\n"; # Trim to host2!host1!user.
}
if ($lbl eq 'Subject:') {
splice @txt,0,0, "$val\n";
}
$line = "\n" . $line
if ! defined($mailhdrs);
++$mailhdrs;
}
$newpar = 0;
}
for $line (@txt) {
print $line;
}
print "\n"; # Boilerplate.