# leading "//" denotes protocol-relative (http:// or https://)
'//marc.info/?i=%s',
'//www.mail-archive.com/search?l=mid&q=%s',
- 'http://mid.gmane.org/%s',
+ 'nntp://news.gmane.org/%s',
'https://lists.debian.org/msgid-search/%s',
'//docs.FreeBSD.org/cgi/mid.cgi?db=mid&id=%s',
'https://www.w3.org/mid/%s',
'doc-url=/lurker&format=en.html&query=id:%s'
);
+sub PARTIAL_MAX () { 100 }
+
+sub search_partial ($$) {
+ my ($srch, $mid) = @_;
+ my $opt = { limit => PARTIAL_MAX, mset => 2 };
+ my @try = ("m:$mid*");
+ my $chop = $mid;
+ if ($chop =~ s/(\W+)(\w*)\z//) {
+ my ($delim, $word) = ($1, $2);
+ if (length($word)) {
+ push @try, "m:$chop$delim";
+ push @try, "m:$chop$delim*";
+ }
+ push @try, "m:$chop";
+ push @try, "m:$chop*";
+ }
+
+ # break out long words individually to search for, because
+ # too many messages begin with "Pine.LNX." (or "alpine" or "nycvar")
+ if ($mid =~ /\w{9,}/) {
+ my @long = ($mid =~ m!(\w{3,})!g);
+ push(@try, join(' ', map { "m:$_" } @long));
+
+ # is the last element long enough to not trigger excessive
+ # wildcard matches?
+ if (length($long[-1]) > 8) {
+ $long[-1] .= '*';
+ push(@try, join(' ', map { "m:$_" } @long));
+ }
+ }
+
+ foreach my $m (@try) {
+ my $mset = eval { $srch->query($m, $opt) };
+ if (ref($@) eq 'Search::Xapian::QueryParserError') {
+ # If Xapian can't handle the wildcard since it
+ # has too many results.
+ next;
+ }
+ my @mids = map {
+ my $doc = $_->get_document;
+ PublicInbox::SearchMsg->load_doc($doc)->mid;
+ } $mset->items;
+ return \@mids if scalar(@mids);
+ }
+}
+
sub ext_msg {
my ($ctx) = @_;
my $cur = $ctx->{-inbox};
my $mid = $ctx->{mid};
- eval { require PublicInbox::Search };
- my $have_xap = $@ ? 0 : 1;
- my (@nox, @ibx, @found);
+ eval { require PublicInbox::Msgmap };
+ my $have_mm = $@ ? 0 : 1;
+ my (@ibx, @found);
$ctx->{www}->{pi_config}->each_inbox(sub {
my ($other) = @_;
return if $other->{name} eq $cur->{name} || !$other->base_url;
- my $s = $other->search;
- if (!$s) {
- push @nox, $other;
- return;
- }
-
- # try to find the URL with Xapian to avoid forking
- my $doc_id = eval { $s->find_unique_doc_id('XMID' . $mid) };
- if ($@) {
- # xapian not configured properly for this repo
- push @nox, $other;
- return;
- }
+ my $mm = $other->mm or return;
- # maybe we found it!
- if (defined $doc_id) {
+ # try to find the URL with Msgmap to avoid forking
+ my $num = $mm->num_for($mid);
+ if (defined $num) {
push @found, $other;
} else {
# no point in trying the fork fallback if we
return exact($ctx, \@found, $mid) if @found;
- # Xapian not installed or configured for some repos,
- # do a full MID check (this is expensive...):
- if (@nox) {
- my $path = mid2path($mid);
- foreach my $other (@nox) {
- my (undef, $type, undef) = $other->path_check($path);
-
- if ($type && $type eq 'blob') {
- push @found, $other;
- }
- }
- }
- return exact($ctx, \@found, $mid) if @found;
-
# fall back to partial MID matching
- my $n_partial = 0;
my @partial;
+ my $n_partial = 0;
+ my $srch = $cur->search;
+ my $mids = search_partial($srch, $mid) if $srch;
+ if ($mids) {
+ $n_partial = scalar(@$mids);
+ push @partial, [ $cur, $mids ];
+ }
- eval { require PublicInbox::Msgmap };
- my $have_mm = $@ ? 0 : 1;
- if ($have_mm) {
- my $tmp_mid = $mid;
-again:
- unshift @ibx, $cur;
+ # can't find a partial match in current inbox, try the others:
+ if (!$n_partial && length($mid) >= 16) {
foreach my $ibx (@ibx) {
- my $mm = $ibx->mm or next;
- if (my $res = $mm->mid_prefixes($tmp_mid)) {
- $n_partial += scalar(@$res);
- push @partial, [ $ibx, $res ];
- }
- }
- # fixup common errors:
- if (!$n_partial && $tmp_mid =~ s,/[tTf],,) {
- goto again;
+ $srch = $ibx->search or next;
+ $mids = search_partial($srch, $mid) or next;
+ $n_partial += scalar(@$mids);
+ push @partial, [ $ibx, $mids];
+ last if $n_partial >= PARTIAL_MAX;
}
}
if ($n_partial) {
$code = 300;
my $es = $n_partial == 1 ? '' : 'es';
+ $n_partial .= '+' if ($n_partial == PARTIAL_MAX);
$s .= "\n$n_partial partial match$es found:\n\n";
my $cur_name = $cur->{name};
foreach my $pair (@partial) {