When indexing diffs, we can avoid indexing the diff parts under
XNQ and instead combine the parts in the read-only search
interface. This results in better indexing performance and
10-15% smaller Xapian indices.
mid => 'Q', # Message-ID (full/exact), this is mostly uniQue
);
mid => 'Q', # Message-ID (full/exact), this is mostly uniQue
);
+my $non_quoted_body = 'XNQ XDFN XDFA XDFB XDFHH XDFCTX XDFPRE XDFPOST';
my %prob_prefix = (
# for mairix compatibility
s => 'S',
my %prob_prefix = (
# for mairix compatibility
s => 'S',
c => 'XCC',
tcf => 'XTO XCC A',
a => 'XTO XCC A',
c => 'XCC',
tcf => 'XTO XCC A',
a => 'XTO XCC A',
- b => 'XNQ XQUOT',
- bs => 'XNQ XQUOT S',
+ b => $non_quoted_body . ' XQUOT',
+ bs => $non_quoted_body . ' XQUOT S',
n => 'XFN',
q => 'XQUOT',
n => 'XFN',
q => 'XQUOT',
+ nq => $non_quoted_body,
dfn => 'XDFN',
dfa => 'XDFA',
dfb => 'XDFB',
dfn => 'XDFN',
dfa => 'XDFA',
dfb => 'XDFB',
dfblob => 'XDFPRE XDFPOST',
# default:
dfblob => 'XDFPRE XDFPOST',
# default:
- '' => 'XM S A XNQ XQUOT XFN',
+ '' => 'XM S A XQUOT XFN ' . $non_quoted_body,
);
# not documenting m: and mid: for now, the using the URLs works w/o Xapian
);
# not documenting m: and mid: for now, the using the URLs works w/o Xapian
-sub index_text_inc ($$$) {
- my ($tg, $text, $pfx) = @_;
+sub index_diff_inc ($$$$) {
+ my ($tg, $text, $pfx, $xnq) = @_;
+ if (@$xnq) {
+ $tg->index_text(join("\n", @$xnq), 1, 'XNQ');
+ $tg->increase_termpos;
+ @$xnq = ();
+ }
$tg->index_text($text, 1, $pfx);
$tg->increase_termpos;
}
sub index_old_diff_fn {
$tg->index_text($text, 1, $pfx);
$tg->increase_termpos;
}
sub index_old_diff_fn {
- my ($tg, $seen, $fa, $fb) = @_;
+ my ($tg, $seen, $fa, $fb, $xnq) = @_;
# no renames or space support for traditional diffs,
# find the number of leading common paths to strip:
# no renames or space support for traditional diffs,
# find the number of leading common paths to strip:
$fa = join('/', @fa);
$fb = join('/', @fb);
if ($fa eq $fb) {
$fa = join('/', @fa);
$fb = join('/', @fb);
if ($fa eq $fb) {
- index_text_inc($tg, $fa,'XDFN') unless $seen->{$fa}++;
+ unless ($seen->{$fa}++) {
+ index_diff_inc($tg, $fa, 'XDFN', $xnq);
+ }
my ($tg, $lines, $doc) = @_;
my %seen;
my $in_diff;
my ($tg, $lines, $doc) = @_;
my %seen;
my $in_diff;
+ my @xnq;
+ my $xnq = \@xnq;
foreach (@$lines) {
if ($in_diff && s/^ //) { # diff context
foreach (@$lines) {
if ($in_diff && s/^ //) { # diff context
- index_text_inc($tg, $_, 'XDFCTX');
+ index_diff_inc($tg, $_, 'XDFCTX', $xnq);
} elsif (/^-- $/) { # email signature begins
$in_diff = undef;
} elsif (m!^diff --git ("?a/.+) ("?b/.+)\z!) {
my ($fa, $fb) = ($1, $2);
my $fn = (split('/', git_unquote($fa), 2))[1];
} elsif (/^-- $/) { # email signature begins
$in_diff = undef;
} elsif (m!^diff --git ("?a/.+) ("?b/.+)\z!) {
my ($fa, $fb) = ($1, $2);
my $fn = (split('/', git_unquote($fa), 2))[1];
- index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
+ $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
$fn = (split('/', git_unquote($fb), 2))[1];
$fn = (split('/', git_unquote($fb), 2))[1];
- index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
+ $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
$in_diff = 1;
# traditional diff:
} elsif (m/^diff -(.+) (\S+) (\S+)$/) {
my ($opt, $fa, $fb) = ($1, $2, $3);
$in_diff = 1;
# traditional diff:
} elsif (m/^diff -(.+) (\S+) (\S+)$/) {
my ($opt, $fa, $fb) = ($1, $2, $3);
# only support unified:
next unless $opt =~ /[uU]/;
# only support unified:
next unless $opt =~ /[uU]/;
- $in_diff = index_old_diff_fn($tg, \%seen, $fa, $fb);
+ $in_diff = index_old_diff_fn($tg, \%seen, $fa, $fb,
+ $xnq);
} elsif (m!^--- ("?a/.+)!) {
my $fn = (split('/', git_unquote($1), 2))[1];
} elsif (m!^--- ("?a/.+)!) {
my $fn = (split('/', git_unquote($1), 2))[1];
- index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
+ $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
$in_diff = 1;
} elsif (m!^\+\+\+ ("?b/.+)!) {
my $fn = (split('/', git_unquote($1), 2))[1];
$in_diff = 1;
} elsif (m!^\+\+\+ ("?b/.+)!) {
my $fn = (split('/', git_unquote($1), 2))[1];
- index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
+ $seen{$fn}++ or index_diff_inc($tg, $fn, 'XDFN', $xnq);
$in_diff = 1;
} elsif (/^--- (\S+)/) {
$in_diff = $1;
$in_diff = 1;
} elsif (/^--- (\S+)/) {
$in_diff = $1;
} elsif (defined $in_diff && /^\+\+\+ (\S+)/) {
} elsif (defined $in_diff && /^\+\+\+ (\S+)/) {
- $in_diff = index_old_diff_fn($tg, \%seen, $in_diff, $1);
+ $in_diff = index_old_diff_fn($tg, \%seen, $in_diff, $1,
+ $xnq);
} elsif ($in_diff && s/^\+//) { # diff added
} elsif ($in_diff && s/^\+//) { # diff added
- index_text_inc($tg, $_, 'XDFB');
+ index_diff_inc($tg, $_, 'XDFB', $xnq);
} elsif ($in_diff && s/^-//) { # diff removed
} elsif ($in_diff && s/^-//) { # diff removed
- index_text_inc($tg, $_, 'XDFA');
+ index_diff_inc($tg, $_, 'XDFA', $xnq);
} elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) {
my ($ba, $bb) = ($1, $2);
index_git_blob_id($doc, 'XDFPRE', $ba);
} elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) {
my ($ba, $bb) = ($1, $2);
index_git_blob_id($doc, 'XDFPRE', $ba);
# traditional diff w/o -p
} elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) {
# hunk header context
# traditional diff w/o -p
} elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) {
# hunk header context
- index_text_inc($tg, $1, 'XDFHH');
+ index_diff_inc($tg, $1, 'XDFHH', $xnq);
# ignore the following lines:
# ignore the following lines:
- } elsif (/^(?:dis)similarity index/) {
- } elsif (/^(?:old|new) mode/) {
- } elsif (/^(?:deleted|new) file mode/) {
- } elsif (/^(?:copy|rename) (?:from|to) /) {
- } elsif (/^(?:dis)?similarity index /) {
- } elsif (/^\\ No newline at end of file/) {
- } elsif (/^Binary files .* differ/) {
+ } elsif (/^(?:dis)similarity index/ ||
+ /^(?:old|new) mode/ ||
+ /^(?:deleted|new) file mode/ ||
+ /^(?:copy|rename) (?:from|to) / ||
+ /^(?:dis)?similarity index / ||
+ /^\\ No newline at end of file/ ||
+ /^Binary files .* differ/) {
+ push @xnq, $_;
} elsif ($_ eq '') {
$in_diff = undef;
} else {
} elsif ($_ eq '') {
$in_diff = undef;
} else {
warn "non-diff line: $_\n" if DEBUG && $_ ne '';
$in_diff = undef;
}
}
warn "non-diff line: $_\n" if DEBUG && $_ ne '';
$in_diff = undef;
}
}
+
+ $tg->index_text(join("\n", @xnq), 1, 'XNQ');
+ $tg->increase_termpos;
}
sub index_body ($$$) {
my ($tg, $lines, $doc) = @_;
my $txt = join("\n", @$lines);
}
sub index_body ($$$) {
my ($tg, $lines, $doc) = @_;
my $txt = join("\n", @$lines);
- $tg->index_text($txt, !!$doc, $doc ? 'XNQ' : 'XQUOT');
- $tg->increase_termpos;
- # does it look like a diff?
- if ($doc && $txt =~ /^(?:diff|---|\+\+\+) /ms) {
- $txt = undef;
- index_diff($tg, $lines, $doc);
+ if ($doc) {
+ # does it look like a diff?
+ if ($txt =~ /^(?:diff|---|\+\+\+) /ms) {
+ $txt = undef;
+ index_diff($tg, $lines, $doc);
+ } else {
+ $tg->index_text($txt, 1, 'XNQ');
+ }
+ } else {
+ $tg->index_text($txt, 0, 'XQUOT');