1 # Copyright (C) 2015-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # A small/skeleton/slim representation of a message.
6 # This used to be "SearchMsg", but we split out overview
7 # indexing into over.sqlite3 so it's not just "search". There
8 # may be many of these objects loaded in memory at once for
9 # large threads in our WWW UI and the NNTP range responses.
10 package PublicInbox::Smsg;
13 use parent qw(Exporter);
14 our @EXPORT_OK = qw(subject_normalized);
15 use PublicInbox::MID qw(mids references);
16 use PublicInbox::Address;
17 use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
19 sub oidbin { pack('H*', $_[0]->{blob}) }
26 $self->{references} // '',
36 sub load_from_data ($$) {
37 my ($self) = $_[0]; # data = $_[1]
44 # To: and Cc: are stored to optimize HDR/XHDR in NNTP since
45 # some NNTP clients will use that for message displays.
46 # NNTP only, and only stored in Over(view), not Xapian
56 ) = split(/\n/, $_[1]);
62 # drop NNTP-only fields which aren't relevant to PSGI results:
63 # saves ~80K on a 200 item search result:
64 # TODO: we may need to keep some of these for JMAP...
65 my ($f) = delete @$self{qw(from tid to cc bytes lines)};
66 # ghosts don't have ->{from}
67 $self->{from_name} = join(', ', PublicInbox::Address::names($f // ''));
71 sub parse_references ($$$) {
72 my ($smsg, $hdr, $mids) = @_;
73 my $refs = references($hdr);
74 push(@$refs, @$mids) if scalar(@$mids) > 1;
75 return $refs if scalar(@$refs) == 0;
77 # prevent circular references here:
78 my %seen = ( ($smsg->{mid} // '') => 1 );
80 foreach my $ref (@$refs) {
81 if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) {
82 warn "References: <$ref> too long, ignoring\n";
85 $seen{$ref} //= push(@keep, $ref);
87 $smsg->{references} = '<'.join('> <', @keep).'>' if @keep;
91 # used for v2, Import and v1 non-SQLite WWW code paths
93 my ($self, $hdr, $sync) = @_;
94 for my $f (qw(From To Cc Subject)) {
95 my @all = $hdr->header($f);
96 my $val = join(', ', @all);
98 # MIME decoding can create NULs, replace them with spaces
99 # to protect git and NNTP clients
100 $val =~ tr/\0\t\n/ /;
102 # rare: in case headers have wide chars (not RFC2047-encoded)
105 # lower-case fields for read-only stuff
106 $self->{lc($f)} = $val;
108 # Capitalized From/Subject for git-fast-import
109 next if $f eq 'To' || $f eq 'Cc';
110 if (scalar(@all) > 1) {
113 $val =~ tr/\0\t\n/ /;
115 $self->{$f} = $val if $val ne '';
118 $self->{-ds} = [ my @ds = msg_datestamp($hdr, $sync->{autime}) ];
119 $self->{-ts} = [ my @ts = msg_timestamp($hdr, $sync->{cotime}) ];
120 $self->{ds} //= $ds[0]; # no zone
121 $self->{ts} //= $ts[0];
122 $self->{mid} //= mids($hdr)->[0];
125 # no strftime, that is locale-dependent and not for RFC822
126 my @DoW = qw(Sun Mon Tue Wed Thu Fri Sat);
127 my @MoY = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
129 sub date ($) { # for NNTP
131 my $ds = $self->{ds};
132 return unless defined $ds;
133 my ($sec, $min, $hour, $mday, $mon, $year, $wday) = gmtime($ds);
134 "$DoW[$wday], " . sprintf("%02d $MoY[$mon] %04d %02d:%02d:%02d +0000",
135 $mday, $year+1900, $hour, $min, $sec);
138 sub internaldate { # for IMAP
140 my ($sec, $min, $hour, $mday, $mon, $year) = gmtime($self->{ts} // 0);
141 sprintf("%02d-$MoY[$mon]-%04d %02d:%02d:%02d +0000",
142 $mday, $year+1900, $hour, $min, $sec);
145 our $REPLY_RE = qr/^re:\s+/i;
147 # TODO: see RFC 5256 sec 2.1 "Base Subject" and evaluate compatibility
148 # w/ existing indices...
149 sub subject_normalized ($) {
151 $subj =~ s/\A\s+//s; # no leading space
152 $subj =~ s/\s+\z//s; # no trailing space
153 $subj =~ s/\s+/ /gs; # no redundant spaces
154 $subj =~ s/\.+\z//; # no trailing '.'
155 $subj =~ s/$REPLY_RE//igo; # remove reply prefix
159 # returns the number of bytes to add if given a non-CRLF arg
160 sub crlf_adjust ($) {
161 if (index($_[0], "\r\n") < 0) {
162 # common case is LF-only, every \n needs an \r;
163 # so favor a cheap tr// over an expensive m//g
165 } else { # count number of '\n' w/o '\r', expensive:
166 scalar(my @n = ($_[0] =~ m/(?<!\r)\n/g));
170 sub set_bytes { $_[0]->{bytes} = $_[2] + crlf_adjust($_[1]) }