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 base 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);
24 $self->{references} // '',
34 sub load_from_data ($$) {
35 my ($self) = $_[0]; # data = $_[1]
42 # To: and Cc: are stored to optimize HDR/XHDR in NNTP since
43 # some NNTP clients will use that for message displays.
44 # NNTP only, and only stored in Over(view), not Xapian
54 ) = split(/\n/, $_[1]);
60 # drop NNTP-only fields which aren't relevant to PSGI results:
61 # saves ~80K on a 200 item search result:
62 # TODO: we may need to keep some of these for JMAP...
63 my ($f) = delete @$self{qw(from tid to cc bytes lines)};
64 # ghosts don't have ->{from}
65 $self->{from_name} = join(', ', PublicInbox::Address::names($f // ''));
69 sub parse_references ($$$) {
70 my ($smsg, $hdr, $mids) = @_;
71 my $refs = references($hdr);
72 push(@$refs, @$mids) if scalar(@$mids) > 1;
73 return $refs if scalar(@$refs) == 0;
75 # prevent circular references here:
76 my %seen = ( ($smsg->{mid} // '') => 1 );
78 foreach my $ref (@$refs) {
79 if (length($ref) > PublicInbox::MID::MAX_MID_SIZE) {
80 warn "References: <$ref> too long, ignoring\n";
83 $seen{$ref} //= push(@keep, $ref);
85 $smsg->{references} = '<'.join('> <', @keep).'>' if @keep;
89 # used for v2, Import and v1 non-SQLite WWW code paths
91 my ($self, $hdr, $sync) = @_;
92 for my $f (qw(From To Cc Subject)) {
93 my @all = $hdr->header($f);
94 my $val = join(', ', @all);
96 # MIME decoding can create NULs, replace them with spaces
97 # to protect git and NNTP clients
100 # rare: in case headers have wide chars (not RFC2047-encoded)
103 # lower-case fields for read-only stuff
104 $self->{lc($f)} = $val;
106 # Capitalized From/Subject for git-fast-import
107 next if $f eq 'To' || $f eq 'Cc';
108 if (scalar(@all) > 1) {
111 $val =~ tr/\0\t\n/ /;
113 $self->{$f} = $val if $val ne '';
116 $self->{-ds} = [ my @ds = msg_datestamp($hdr, $sync->{autime}) ];
117 $self->{-ts} = [ my @ts = msg_timestamp($hdr, $sync->{cotime}) ];
118 $self->{ds} //= $ds[0]; # no zone
119 $self->{ts} //= $ts[0];
120 $self->{mid} //= mids($hdr)->[0];
123 # no strftime, that is locale-dependent and not for RFC822
124 my @DoW = qw(Sun Mon Tue Wed Thu Fri Sat);
125 my @MoY = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
127 sub date ($) { # for NNTP
129 my $ds = $self->{ds};
130 return unless defined $ds;
131 my ($sec, $min, $hour, $mday, $mon, $year, $wday) = gmtime($ds);
132 "$DoW[$wday], " . sprintf("%02d $MoY[$mon] %04d %02d:%02d:%02d +0000",
133 $mday, $year+1900, $hour, $min, $sec);
136 sub internaldate { # for IMAP
138 my ($sec, $min, $hour, $mday, $mon, $year) = gmtime($self->{ts} // 0);
139 sprintf("%02d-$MoY[$mon]-%04d %02d:%02d:%02d +0000",
140 $mday, $year+1900, $hour, $min, $sec);
143 our $REPLY_RE = qr/^re:\s+/i;
145 # TODO: see RFC 5256 sec 2.1 "Base Subject" and evaluate compatibility
146 # w/ existing indices...
147 sub subject_normalized ($) {
149 $subj =~ s/\A\s+//s; # no leading space
150 $subj =~ s/\s+\z//s; # no trailing space
151 $subj =~ s/\s+/ /gs; # no redundant spaces
152 $subj =~ s/\.+\z//; # no trailing '.'
153 $subj =~ s/$REPLY_RE//igo; # remove reply prefix
157 # returns the number of bytes to add if given a non-CRLF arg
158 sub crlf_adjust ($) {
159 if (index($_[0], "\r\n") < 0) {
160 # common case is LF-only, every \n needs an \r;
161 # so favor a cheap tr// over an expensive m//g
163 } else { # count number of '\n' w/o '\r', expensive:
164 scalar(my @n = ($_[0] =~ m/(?<!\r)\n/g));
168 sub set_bytes { $_[0]->{bytes} = $_[2] + crlf_adjust($_[1]) }