Skip to content

Commit

Permalink
20070803
Browse files Browse the repository at this point in the history 
    - Set LANG=C when generating the manpage
    - Detect html attachments even when mimetype is application/octet-stream
    - Fix: stop optimizing blanks, might be signed (spotted by jmc at dolorespark.org)
  • Loading branch information
@raforg
raforg committed Aug 3, 2007
1 parent 22b14bd commit c69ac07
Show file tree
Hide file tree
Showing 2 changed files with 91 additions and 83 deletions.
6 changes: 6 additions & 0 deletions CHANGELOG
View file
Original file line number Diff line number Diff line change
@@ -1,3 +1,9 @@
20070803

- Set LANG=C when generating the manpage
- Detect html attachments even when mimetype is application/octet-stream
- Fix: stop optimizing blanks, might be signed (spotted by jmc at dolorespark.org)

20060525

- Fixed signal handler (wasn't calling exit after cleaning up!)
Expand Down
168 changes: 85 additions & 83 deletions textmail
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@ use strict;

# textmail - mail filter to replace MS Word/HTML attachments with plain text
#
# Copyright (C) 2003-2005 raf <raf@raf.org>
# Copyright (C) 2003-2007 raf <raf@raf.org>
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
Expand All @@ -20,7 +20,7 @@ use strict;
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
# or visit http://www.gnu.org/copyleft/gpl.html
#
# 20051129 raf <raf@raf.org>
# 20070803 raf <raf@raf.org>

=head1 NAME
Expand Down Expand Up @@ -348,82 +348,17 @@ C<http://raf.org/minimail/>
=head1 AUTHOR
20051129 raf <raf@raf.org>
20070803 raf <raf@raf.org>
=head1 URL
C<http://raf.org/textmail/>
=cut

# Doco functions: usage and manpage (via $PAGER or as nroff or html)

sub help
{
print
"usage: textmail [options]\n",
"options:\n",
" -h - Print the help message then exit\n",
" -m - Print the manpage then exit\n",
" -w - Print the manpage in html format then exit\n",
" -r - Print the manpage in nroff format then exit\n",
" -M - Output in mailbox format\n",
" -T - Output in raw mail format (for smtp)\n",
" -W - Don't replace MS Word attachments with text\n",
" -E - Don't replace MS Excel attachments with csv\n",
" -H - Don't replace HTML attachments with text\n",
" -R - Don't replace RTF attachments with text\n",
" -P - Don't replace PDF attachments with text\n",
" -U - Don't translate winmail.dat attachments\n",
" -L - Don't reduce appledouble attachments\n",
" -I - Don't delete image attachments\n",
" -A - Don't delete audio attachments\n",
" -V - Don't delete video attachments\n",
" -X - Don't delete MS Windows executable attachments\n",
" -B - Don't recode text that was base64-encoded\n",
" -S - Don't replace spaces in filenames with underscores\n",
" -Z - Do translate signed content (discards signatures)\n",
" -O - Delete all application/octet-stream attachments\n",
" -! - Delete all application/* attachments\n",
" -D hdrs - Delete headers (list of header prefixes and filenames)\n",
" -K types - Keep attachments (list of mimetypes and filenames)\n",
" -f - On translation error, keep translation, not original\n",
" -? - Print paths of helper applications then exit\n",
"\n",
"Filters a mail message or mbox, replacing MS Word, MS Excel, HTML, RTF and PDF\n",
"attachments with the plain text contained therein. By default, the following\n",
"attachments are also deleted: image, audio, video and MS Windows executables.\n",
"MS winmail.dat attachments are replaced by any attachments contained therein\n",
"which are then replaced by text or deleted in the same fashion. Any of these\n",
"actions can be suppressed with the command line options. Mail headers can also\n",
"be selectively deleted.\n";
exit;
}

sub man
{
my $noquotes = (`pod2man -h 2>&1` =~ /--quotes=/) ? '--quotes=none' : '';
system "pod2man $noquotes $0 | nroff -man | " . ($ENV{PAGER} || 'more');
exit;
}

sub nroff
{
my $noquotes = (`pod2man -h 2>&1` =~ /--quotes=/) ? '--quotes=none' : '';
system "pod2man $noquotes $0";
exit;
}
# Functions from minimail: see http://raf.org/minimail/

sub html
{
system "pod2html --noindex $0";
unlink glob 'pod2htm*';
exit;
}

# Minimail functions: see http://raf.org/minimail/

sub formail # rfc2822 + mboxrd format (see www.qmail.org/man/man5/mbox.html)
sub formail # rfc2822 + mboxrd format (see http://www.qmail.org/man/man5/mbox.html)
{
sub mime # rfc2045, rfc2046
{
Expand All @@ -449,12 +384,10 @@ sub formail # rfc2822 + mboxrd format (see www.qmail.org/man/man5/mbox.html)
{
if (/^--\Q$mail->{mime_boundary}\E(--)?/)
{
$text = substr($text, 0, -1) if substr($text, -1) eq "\n";

if ($state eq 'preamble')
{
$mail->{mime_preamble} = $text if length $text;
$state = 'part';
$mail->{mime_preamble} = $text if length $text;
}
elsif ($state eq 'part')
{
Expand Down Expand Up @@ -532,11 +465,11 @@ sub mail2str
$head .= join '', @{$mail->{headers}} if exists $mail->{headers};
my $body = '';
$body .= $mail->{body} if exists $mail->{body};
$body .= "$mail->{mime_preamble}\n" if exists $mail->{mime_preamble};
$body .= "$mail->{mime_preamble}" if exists $mail->{mime_preamble};
$body .= "--$mail->{mime_boundary}\n" if exists $mail->{mime_boundary} && !exists $mail->{mime_parts};
$body .= join "\n", map { "--$mail->{mime_boundary}\n" . mail2str($_) } @{$mail->{mime_parts}} if exists $mail->{mime_parts};
$body .= "\n--$mail->{mime_boundary}--" if exists $mail->{mime_boundary};
$body .= "\n$mail->{mime_epilogue}" if exists $mail->{mime_epilogue};
$body .= join("", map { "--$mail->{mime_boundary}\n" . mail2str($_) } @{$mail->{mime_parts}}) if exists $mail->{mime_parts};
$body .= "--$mail->{mime_boundary}--\n" if exists $mail->{mime_boundary};
$body .= "$mail->{mime_epilogue}" if exists $mail->{mime_epilogue};
$body .= mail2str($mail->{mime_message}) if exists $mail->{mime_message};
$body =~ s/^(>*From )/>$1/mg, $body =~ s/([^\n])\n?\z/$1\n\n/ if exists $mail->{mbox};
return $head . "\n" . $body;
Expand Down Expand Up @@ -753,13 +686,13 @@ sub newmail # rfc2822, rfc2045, rfc2046, rfc2183 (also rfc3282, rfc3066, rfc2424
sub decode
{
my ($d, $e) = @_;
return $e =~ /^base64$/i ? decode_base64($d) : $e =~ /^quoted-printable$/i ? decode_quoted_printable($d) : $d;
return $e =~ /^base64$/i ? decode_base64($d) : $e =~ /^quoted-printable$/i ? decode_quoted_printable($d) : substr($d, 0, -1);
}

sub encode
{
my ($d, $e) = @_;
return $e =~ /^base64$/i ? encode_base64($d) : $e =~ /^quoted-printable$/i ? encode_quoted_printable($d) : $d;
return $e =~ /^base64$/i ? encode_base64($d) : $e =~ /^quoted-printable$/i ? encode_quoted_printable($d) : $d . "\n";
}

sub choose_encoding # rfc2822, rfc2045
Expand Down Expand Up @@ -928,6 +861,75 @@ sub winmail
return ($badtnef) ? $m : map { newmail(%$_) } @attachment;
}
# Doco functions: usage and manpage (via $PAGER or as nroff or html)
$ENV{LANG} = 'C';
sub help
{
print
"usage: textmail [options]\n",
"options:\n",
" -h - Print the help message then exit\n",
" -m - Print the manpage then exit\n",
" -w - Print the manpage in html format then exit\n",
" -r - Print the manpage in nroff format then exit\n",
" -M - Output in mailbox format\n",
" -T - Output in raw mail format (for smtp)\n",
" -W - Don't replace MS Word attachments with text\n",
" -E - Don't replace MS Excel attachments with csv\n",
" -H - Don't replace HTML attachments with text\n",
" -R - Don't replace RTF attachments with text\n",
" -P - Don't replace PDF attachments with text\n",
" -U - Don't translate winmail.dat attachments\n",
" -L - Don't reduce appledouble attachments\n",
" -I - Don't delete image attachments\n",
" -A - Don't delete audio attachments\n",
" -V - Don't delete video attachments\n",
" -X - Don't delete MS Windows executable attachments\n",
" -B - Don't recode text that was base64-encoded\n",
" -S - Don't replace spaces in filenames with underscores\n",
" -Z - Do translate signed content (discards signatures)\n",
" -O - Delete all application/octet-stream attachments\n",
" -! - Delete all application/* attachments\n",
" -D hdrs - Delete headers (list of header prefixes and filenames)\n",
" -K types - Keep attachments (list of mimetypes and filenames)\n",
" -f - On translation error, keep translation, not original\n",
" -? - Print paths of helper applications then exit\n",
"\n",
"Filters a mail message or mbox, replacing MS Word, MS Excel, HTML, RTF and PDF\n",
"attachments with the plain text contained therein. By default, the following\n",
"attachments are also deleted: image, audio, video and MS Windows executables.\n",
"MS winmail.dat attachments are replaced by any attachments contained therein\n",
"which are then replaced by text or deleted in the same fashion. Any of these\n",
"actions can be suppressed with the command line options. Mail headers can also\n",
"be selectively deleted.\n";
exit;
}
sub man
{
my $noquotes = (`pod2man -h 2>&1` =~ /--quotes=/) ? '--quotes=none' : '';
system "pod2man $noquotes $0 | nroff -man | " . ($ENV{PAGER} || 'more');
exit;
}
sub nroff
{
my $noquotes = (`pod2man -h 2>&1` =~ /--quotes=/) ? '--quotes=none' : '';
system "pod2man $noquotes $0";
exit;
}
sub html
{
system "pod2html --noindex $0";
unlink glob 'pod2htm*';
exit;
}
# Parse command line
my %opt;
use Getopt::Std;
help unless getopts 'hmrwMTWEHRPLUIAVXBSZO!D:K:f?', \%opt;
Expand Down Expand Up @@ -980,7 +982,7 @@ if (!$removing || (($? || !defined $tmp || ! -d $tmp) && !mkdir($tmp = "/tmp/tex
exit;
};
# Filter the mail message on stdin into text on stdout
# Filter mail message(s) on stdin into text on stdout
formail(sub { <> }, sub
{
Expand All @@ -994,7 +996,7 @@ rmdir $tmp or system "rm -rf $tmp";
BEGIN { $SIG{INT} = $SIG{QUIT} = $SIG{TERM} = sub { rmdir $tmp or system "rm -rf $tmp" if defined $tmp; exit } }
# Print paths to help applications then exit
# Print paths to helper applications then exit
sub paths
{
Expand Down Expand Up @@ -1030,7 +1032,7 @@ sub textmail
if ($remove_html && isa($entity, 'multipart/alternative') && @parts == 2)
{
if (isa($parts[0], 'text/plain') && isa($parts[1], 'text/html') || isa($parts[1], 'text/plain') && isa($parts[0], 'text/html'))
if (isa($parts[0], 'text/plain') && isa($parts[1], 'text/html', qr/\.html?$/i) || isa($parts[1], 'text/plain') && isa($parts[0], 'text/html', qr/\.html?$/i))
{
my $plain = $parts[isa($parts[0], 'text/plain') ? 0 : 1];
@{$plain->{headers}} = (grep(!/^content-/i, @{$entity->{headers}}), grep { /^content-/i } @{$plain->{headers}});
Expand Down Expand Up @@ -1090,7 +1092,7 @@ sub textmail
# Replace HTML attachments with plain text (via lynx -dump)
if ($remove_html && isa($parts[$i], 'text/html'))
if ($remove_html && isa($parts[$i], 'text/html', qr/\.html?$/i))
{
$parts[$i] = translate($parts[$i], 'html,htm', 'txt', (defined $lynx) ? "$lynx -dump -force_html" : undef);
next;
Expand Down

0 comments on commit c69ac07

Please sign in to comment.