X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FEml.pm;h=8b999e1a88ef8d11b78e7c5bd0c022dbd7d0f7be;hb=HEAD;hp=4508bd84c6b1b3457fa7b9247331b94694eabc51;hpb=b2f2d28e20f7bf9e7234d3134ec91cd9549001b1;p=public-inbox.git diff --git a/lib/PublicInbox/Eml.pm b/lib/PublicInbox/Eml.pm index 4508bd84..8b999e1a 100644 --- a/lib/PublicInbox/Eml.pm +++ b/lib/PublicInbox/Eml.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2020 all contributors +# Copyright (C) all contributors # License: AGPL-3.0+ # # Lazy MIME parser, it still slurps the full message but keeps short @@ -28,7 +28,7 @@ package PublicInbox::Eml; use strict; use v5.10.1; use Carp qw(croak); -use Encode qw(find_encoding decode encode); # stdlib +use Encode qw(find_encoding); # stdlib use Text::Wrap qw(wrap); # stdlib, we need Perl 5.6+ for $huge use MIME::Base64 3.05; # Perl 5.10.0 / 5.9.2 use MIME::QuotedPrint 3.05; # ditto @@ -38,9 +38,12 @@ my $MIME_Header = find_encoding('MIME-Header'); use PublicInbox::EmlContentFoo qw(parse_content_type parse_content_disposition); $PublicInbox::EmlContentFoo::STRICT_PARAMS = 0; -our $MAXPARTS = 1000; # same as SpamAssassin -our $MAXDEPTH = 20; # seems enough, Perl sucks, here -our $MAXBOUNDLEN = 2048; # same as postfix +our $mime_parts_limit = 1000; # same as SpamAssassin (not in postfix AFAIK) + +# the rest of the limit names are taken from postfix: +our $mime_nesting_limit = 20; # seems enough, Perl sucks, here +our $mime_boundary_length_limit = 2048; # same as postfix +our $header_size_limit = 102400; # same as postfix my %MIME_ENC = (qp => \&enc_qp, base64 => \&encode_base64); my %MIME_DEC = (qp => \&dec_qp, base64 => \&decode_base64); @@ -48,7 +51,9 @@ $MIME_ENC{quotedprint} = $MIME_ENC{'quoted-printable'} = $MIME_ENC{qp}; $MIME_DEC{quotedprint} = $MIME_DEC{'quoted-printable'} = $MIME_DEC{qp}; $MIME_ENC{$_} = \&identity_codec for qw(7bit 8bit binary); -my %DECODE_ADDRESS = map { $_ => 1 } qw(From To Cc Sender Reply-To); +my %DECODE_ADDRESS = map { + ($_ => 1, "Resent-$_" => 1) +} qw(From To Cc Sender Reply-To Bcc); my %DECODE_FULL = ( Subject => 1, 'Content-Description' => 1, @@ -57,6 +62,14 @@ my %DECODE_FULL = ( our %STR_TYPE = (text => 1); our %STR_SUBTYPE = (plain => 1, html => 1); +# message/* subtypes we descend into +our %MESSAGE_DESCEND = ( + news => 1, # RFC 1849 (obsolete, but archives are forever) + rfc822 => 1, # RFC 2046 + rfc2822 => 1, # gmime handles this (but not rfc5322) + global => 1, # RFC 6532 +); + my %re_memo; sub re_memo ($) { my ($k) = @_; @@ -68,30 +81,59 @@ sub re_memo ($) { /ismx } +sub hdr_truncate ($) { + my $len = length($_[0]); + substr($_[0], $header_size_limit, $len) = ''; + my $end = rindex($_[0], "\n"); + if ($end >= 0) { + ++$end; + substr($_[0], $end, $len) = ''; + warn "header of $len bytes truncated to $end bytes\n"; + } else { + $_[0] = ''; + warn <= 0) { + # likely on *nix + my $hdr = substr($$ref, 0, $pos + 2, ''); # sv_chop on $$ref + chop($hdr); # lower SvCUR + hdr_truncate($hdr) if length($hdr) > $header_size_limit; + bless { hdr => \$hdr, crlf => "\n", bdy => $ref }, __PACKAGE__; + } elsif ($$ref =~ /\r?\n(\r?\n)/s) { + my $hdr = substr($$ref, 0, $+[0], ''); # sv_chop on $$ref substr($hdr, -(length($1))) = ''; # lower SvCUR + hdr_truncate($hdr) if length($hdr) > $header_size_limit; bless { hdr => \$hdr, crlf => $1, bdy => $ref }, __PACKAGE__; } elsif ($$ref =~ /^[a-z0-9-]+[ \t]*:/ims && $$ref =~ /(\r?\n)\z/s) { # body is optional :P - bless { hdr => \($$ref), crlf => $1 }, __PACKAGE__; - } else { # nothing useful - my $hdr = $$ref = ''; - bless { hdr => \$hdr, crlf => "\n" }, __PACKAGE__; + my $hdr = substr($$ref, 0, $header_size_limit + 1); + hdr_truncate($hdr) if length($hdr) > $header_size_limit; + bless { hdr => \$hdr, crlf => $1 }, __PACKAGE__; + } else { # just a body w/o header? + my $hdr = ''; + my $eol = ($$ref =~ /(\r?\n)/) ? $1 : "\n"; + bless { hdr => \$hdr, crlf => $eol, bdy => $ref }, __PACKAGE__; } } sub new_sub { my (undef, $ref) = @_; # special case for messages like <85k5su9k59.fsf_-_@lola.goethe.zz> - $$ref =~ /\A(?:(\r?\n))/gs or goto &new; - my $hdr = substr($$ref, 0, pos($$ref), ''); # sv_chop on $$ref + $$ref =~ /\A(\r?\n)/s or return new(undef, $ref); + my $hdr = substr($$ref, 0, $+[0], ''); # sv_chop on $$ref bless { hdr => \$hdr, crlf => $1, bdy => $ref }, __PACKAGE__; } @@ -102,6 +144,7 @@ sub header_raw { my $re = re_memo($_[1]); my @v = (${ $_[0]->{hdr} } =~ /$re/g); for (@v) { + utf8::decode($_); # SMTPUTF8 # for compatibility w/ Email::Simple::Header, s/\s+\z//s; s/\A\s+//s; @@ -118,13 +161,25 @@ sub ct ($) { } # returns a queue of sub-parts iff it's worth descending into -# TODO: descend into message/rfc822 parts (Email::MIME didn't) sub mp_descend ($$) { my ($self, $nr) = @_; # or $once for top-level - my $bnd = ct($self)->{attributes}->{boundary} // return; # single-part - return if $bnd eq '' || length($bnd) >= $MAXBOUNDLEN; + my $ct = ct($self); + my $type = lc($ct->{type}); + if ($type eq 'message' && $MESSAGE_DESCEND{lc($ct->{subtype})}) { + my $nxt = new(undef, body_raw($self)); + $self->{-call_cb} = $nxt->{is_submsg} = 1; + return [ $nxt ]; + } + return if $type ne 'multipart'; + my $bnd = $ct->{attributes}->{boundary} // return; # single-part + return if $bnd eq '' || length($bnd) >= $mime_boundary_length_limit; $bnd = quotemeta($bnd); + # this is a multipart message that didn't get descended into in + # public-inbox <= 1.5.0, so ensure we call the user callback for + # this part to not break PSGI downloads. + $self->{-call_cb} = $self->{is_submsg}; + # "multipart" messages can exist w/o a body my $bdy = ($nr ? delete($self->{bdy}) : \(body_raw($self))) or return; @@ -132,8 +187,8 @@ sub mp_descend ($$) { # *sigh* just the regexp match alone seems to bump RSS by # length($$bdy) on a ~30M string: my $epilogue_missing; - if ($$bdy =~ /((?:\r?\n)?^--$bnd--[ \t]*\r?$)/gsm) { - substr($$bdy, pos($$bdy) - length($1)) = ''; + if ($$bdy =~ /(?:\r?\n)?^--$bnd--[ \t]*\r?$/sm) { + substr($$bdy, $-[0]) = ''; } else { $epilogue_missing = 1; } @@ -150,7 +205,7 @@ sub mp_descend ($$) { # + 3 since we don't want the last part # processed to include any other excluded # parts ($nr starts at 1, and I suck at math) - $MAXPARTS + 3 - $nr); + $mime_parts_limit + 3 - $nr); if (@parts) { # the usual path if we got this far: undef $bdy; # release memory ASAP if $nr > 0 @@ -158,14 +213,15 @@ sub mp_descend ($$) { # compatibility with Email::MIME $parts[-1] =~ s/\n\r?\n\z/\n/s if $epilogue_missing; - @parts = grep /[^ \t\r\n]/s, @parts; # ignore empty parts + # ignore empty parts + @parts = map { new_sub(undef, \$_) } grep /[^ \t\r\n]/s, @parts; # Keep "From: someone..." from preamble in old, # buggy versions of git-send-email, otherwise drop it # There's also a case where quoted text showed up in the # preamble # <20060515162817.65F0F1BBAE@citi.umich.edu> - unshift(@parts, $pre) if $pre =~ /:/s; + unshift(@parts, new_sub(undef, \$pre)) if index($pre, ':') >= 0; return \@parts; } # "multipart", but no boundary found, treat as single part @@ -182,22 +238,30 @@ sub mp_descend ($$) { # $cb - user-supplied callback sub # $arg - user-supplied arg (think pthread_create) # $once - unref body scalar during iteration +# $all - used by IMAP server, only sub each_part { - my ($self, $cb, $arg, $once) = @_; + my ($self, $cb, $arg, $once, $all) = @_; my $p = mp_descend($self, $once // 0) or - return $cb->([$self, 0, 0], $arg); + return $cb->([$self, 0, 1], $arg); + + $cb->([$self, 0, 0], $arg) if ($all || $self->{-call_cb}); # rare + $p = [ $p, 0 ]; my @s; # our virtual stack my $nr = 0; - while ((scalar(@{$p->[0]}) || ($p = pop @s)) && ++$nr <= $MAXPARTS) { + while ((scalar(@{$p->[0]}) || ($p = pop @s)) && + ++$nr <= $mime_parts_limit) { ++$p->[-1]; # bump index my (undef, @idx) = @$p; @idx = (join('.', @idx)); my $depth = ($idx[0] =~ tr/././) + 1; - my $sub = new_sub(undef, \(shift @{$p->[0]})); - if ($depth < $MAXDEPTH && (my $nxt = mp_descend($sub, $nr))) { + my $sub = shift @{$p->[0]}; + if ($depth < $mime_nesting_limit && + (my $nxt = mp_descend($sub, $nr))) { push(@s, $p) if scalar @{$p->[0]}; $p = [ $nxt, @idx, 0 ]; + ($all || $sub->{-call_cb}) and + $cb->([$sub, $depth, @idx], $arg); } else { # a leaf node $cb->([$sub, $depth, @idx], $arg); } @@ -237,7 +301,7 @@ sub subparts { if ($$bdy =~ /^--\Q$bnd\E--[ \t]*\r?\n(.+)\z/sm) { $self->{epilogue} = $1; } - map { new_sub(undef, \$_) } @$parts; + @$parts; } sub parts_set { @@ -271,10 +335,18 @@ sub body_set { } sub body_str_set { - my ($self, $body_str) = @_; - my $charset = ct($self)->{attributes}->{charset} or - Carp::confess('body_str was given, but no charset is defined'); - body_set($self, \(encode($charset, $body_str, Encode::FB_CROAK))); + my ($self, $str) = @_; + my $cs = ct($self)->{attributes}->{charset} // + croak('body_str was given, but no charset is defined'); + my $enc = find_encoding($cs) // croak "unknown encoding `$cs'"; + my $tmp; + { + my @w; + local $SIG{__WARN__} = sub { push @w, @_ }; + $tmp = $enc->encode($str, Encode::FB_WARN); + croak(@w) if @w; + }; + body_set($self, \$tmp); } sub content_type { scalar header($_[0], 'Content-Type') } @@ -288,14 +360,15 @@ sub header_set { $pfx .= ': '; my $len = 78 - length($pfx); @vals = map {; + utf8::encode(my $v = $_); # to bytes, support SMTPUTF8 # folding differs from Email::Simple::Header, # we favor tabs for visibility (and space savings :P) if (length($_) >= $len && (/\n[^ \t]/s || !/\n/s)) { local $Text::Wrap::columns = $len; local $Text::Wrap::huge = 'overflow'; - $pfx . wrap('', "\t", $_) . $self->{crlf}; + $pfx . wrap('', "\t", $v) . $self->{crlf}; } else { - $pfx . $_ . $self->{crlf}; + $pfx . $v . $self->{crlf}; } } @vals; $$hdr =~ s!$re!shift(@vals) // ''!ge; # replace current headers, first @@ -309,14 +382,19 @@ sub header_str_set { my ($self, $name, @vals) = @_; for (@vals) { next unless /[^\x20-\x7e]/; - utf8::encode($_); # to octets # 39: int((75 - length("Subject: =?UTF-8?B?".'?=') ) / 4) * 3; - s/(.{1,39})/'=?UTF-8?B?'.encode_base64($1, '').'?='/ges; + s/(.{1,39})/ + my $x = $1; + utf8::encode($x); # to octets + '=?UTF-8?B?'.encode_base64($x, '').'?=' + /xges; } header_set($self, $name, @vals); } -sub mhdr_decode ($) { eval { $MIME_Header->decode($_[0]) } // $_[0] } +sub mhdr_decode ($) { + eval { $MIME_Header->decode($_[0], Encode::FB_DEFAULT) } // $_[0]; +} sub filename { my $dis = header_raw($_[0], 'Content-Disposition'); @@ -385,15 +463,20 @@ sub body { sub body_str { my ($self) = @_; my $ct = ct($self); - my $charset = $ct->{attributes}->{charset}; - if (!$charset) { - if ($STR_TYPE{$ct->{type}} && $STR_SUBTYPE{$ct->{subtype}}) { + my $cs = $ct->{attributes}->{charset} // do { + ($STR_TYPE{$ct->{type}} && $STR_SUBTYPE{$ct->{subtype}}) and return body($self); - } - Carp::confess("can't get body as a string for ", + croak("can't get body as a string for ", join("\n\t", header_raw($self, 'Content-Type'))); - } - decode($charset, body($self), Encode::FB_CROAK); + }; + my $enc = find_encoding($cs) or croak "unknown encoding `$cs'"; + my $tmp = body($self); + # workaround https://rt.cpan.org/Public/Bug/Display.html?id=139622 + my @w; + local $SIG{__WARN__} = sub { push @w, @_ }; + my $ret = $enc->decode($tmp, Encode::FB_WARN); + croak(@w) if @w; + $ret; } sub as_string { @@ -413,6 +496,33 @@ sub charset_set { sub crlf { $_[0]->{crlf} // "\n" } +sub raw_size { + my ($self) = @_; + my $len = length(${$self->{hdr}}); + defined($self->{bdy}) and + $len += length(${$self->{bdy}}) + length($self->{crlf}); + $len; +} + +# warnings to ignore when handling spam mailboxes and maybe other places +sub warn_ignore { + my $s = "@_"; + # Email::Address::XS warnings + $s =~ /^Argument contains empty / + || $s =~ /^Element at index [0-9]+.*? contains / + # PublicInbox::MsgTime + || $s =~ /^bogus TZ offset: .+?, ignoring and assuming \+0000/ + || $s =~ /^bad Date: .+? in / + # Encode::Unicode::UTF7 + || $s =~ /^Bad UTF7 data escape at / +} + +# this expects to be RHS in this assignment: "local $SIG{__WARN__} = ..." +sub warn_ignore_cb { + my $cb = $SIG{__WARN__} // \&CORE::warn; + sub { $cb->(@_) unless warn_ignore(@_) } +} + sub willneed { re_memo($_) for @_ } willneed(qw(From To Cc Date Subject Content-Type In-Reply-To References