* [PATCH 1/4] v2writable: allow tracking parallel versions
2018-04-05 21:45 [PATCH 0/4] a few more v2 updates Eric Wong (Contractor, The Linux Foundation)
@ 2018-04-05 21:45 ` Eric Wong (Contractor, The Linux Foundation)
2018-04-05 21:45 ` [PATCH 2/4] v2writable: refer to git each repository as "epoch" Eric Wong (Contractor, The Linux Foundation)
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Eric Wong (Contractor, The Linux Foundation) @ 2018-04-05 21:45 UTC (permalink / raw)
To: meta
For upgrades, this will let users keep an old version
running while performing "public-inbox-index" on the
newest version.
---
lib/PublicInbox/Msgmap.pm | 9 ++++++---
lib/PublicInbox/V2Writable.pm | 17 +++++++++++------
2 files changed, 17 insertions(+), 9 deletions(-)
diff --git a/lib/PublicInbox/Msgmap.pm b/lib/PublicInbox/Msgmap.pm
index 5c37e16..f5f8843 100644
--- a/lib/PublicInbox/Msgmap.pm
+++ b/lib/PublicInbox/Msgmap.pm
@@ -92,9 +92,12 @@ sub last_commit {
$self->meta_accessor('last_commit', $commit);
}
-sub last_commit_n {
- my ($self, $i, $commit) = @_;
- $self->meta_accessor('last_commit'.$i, $commit);
+# v2 uses this to keep track of how up-to-date Xapian is
+# old versions may be automatically GC'ed away in the future,
+# but it's a trivial amount of storage.
+sub last_commit_xap {
+ my ($self, $version, $i, $commit) = @_;
+ $self->meta_accessor("last_xap$version-$i", $commit);
}
sub created_at {
diff --git a/lib/PublicInbox/V2Writable.pm b/lib/PublicInbox/V2Writable.pm
index 8323846..ea116f1 100644
--- a/lib/PublicInbox/V2Writable.pm
+++ b/lib/PublicInbox/V2Writable.pm
@@ -71,7 +71,7 @@ sub new {
lock_path => "$dir/inbox.lock",
# limit each repo to 1GB or so
rotate_bytes => int((1024 * 1024 * 1024) / $PACKING_FACTOR),
- last_commit => [],
+ last_commit => [], # git repo -> commit
};
$self->{partitions} = count_partitions($self) || nproc();
bless $self, $class;
@@ -339,15 +339,20 @@ sub purge {
$purges;
}
+sub last_commit_part ($$;$) {
+ my ($self, $i, $cmt) = @_;
+ my $v = PublicInbox::Search::SCHEMA_VERSION();
+ $self->{mm}->last_commit_xap($v, $i, $cmt);
+}
+
sub set_last_commits ($) {
my ($self) = @_;
defined(my $max_git = $self->{max_git}) or return;
- my $mm = $self->{mm};
my $last_commit = $self->{last_commit};
foreach my $i (0..$max_git) {
defined(my $cmt = $last_commit->[$i]) or next;
$last_commit->[$i] = undef;
- $mm->last_commit_n($i, $cmt);
+ last_commit_part($self, $i, $cmt);
}
}
@@ -673,13 +678,13 @@ sub reindex_oid {
# only update last_commit for $i on reindex iff newer than current
sub update_last_commit {
my ($self, $git, $i, $cmt) = @_;
- my $last = $self->{mm}->last_commit_n($i);
+ my $last = last_commit_part($self, $i);
if (defined $last && is_ancestor($git, $last, $cmt)) {
my @cmd = (qw(rev-list --count), "$last..$cmt");
chomp(my $n = $git->qx(@cmd));
return if $n ne '' && $n == 0;
}
- $self->{mm}->last_commit_n($i, $cmt);
+ last_commit_part($self, $i, $cmt);
}
sub git_dir_n ($$) { "$_[0]->{-inbox}->{mainrepo}/git/$_[1].git" }
@@ -688,7 +693,7 @@ sub last_commits {
my ($self, $max_git) = @_;
my $heads = [];
for (my $i = $max_git; $i >= 0; $i--) {
- $heads->[$i] = $self->{mm}->last_commit_n($i);
+ $heads->[$i] = last_commit_part($self, $i);
}
$heads;
}
--
EW
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH 2/4] v2writable: refer to git each repository as "epoch"
2018-04-05 21:45 [PATCH 0/4] a few more v2 updates Eric Wong (Contractor, The Linux Foundation)
2018-04-05 21:45 ` [PATCH 1/4] v2writable: allow tracking parallel versions Eric Wong (Contractor, The Linux Foundation)
@ 2018-04-05 21:45 ` Eric Wong (Contractor, The Linux Foundation)
2018-04-05 21:45 ` [PATCH 3/4] over: use only supported and safe SQLite APIs Eric Wong (Contractor, The Linux Foundation)
2018-04-05 21:45 ` [PATCH 4/4] search: index and allow searching by date-time Eric Wong (Contractor, The Linux Foundation)
3 siblings, 0 replies; 5+ messages in thread
From: Eric Wong (Contractor, The Linux Foundation) @ 2018-04-05 21:45 UTC (permalink / raw)
To: meta
This hopefully helps for people who try to understand
this design.
---
lib/PublicInbox/V2Writable.pm | 54 +++++++++++++++++++++----------------------
1 file changed, 27 insertions(+), 27 deletions(-)
diff --git a/lib/PublicInbox/V2Writable.pm b/lib/PublicInbox/V2Writable.pm
index ea116f1..b858dcc 100644
--- a/lib/PublicInbox/V2Writable.pm
+++ b/lib/PublicInbox/V2Writable.pm
@@ -69,7 +69,7 @@ sub new {
xpfx => $xpfx,
over => PublicInbox::OverIdxFork->new("$xpfx/over.sqlite3"),
lock_path => "$dir/inbox.lock",
- # limit each repo to 1GB or so
+ # limit each git repo (epoch) to 1GB or so
rotate_bytes => int((1024 * 1024 * 1024) / $PACKING_FACTOR),
last_commit => [], # git repo -> commit
};
@@ -81,9 +81,9 @@ sub init_inbox {
my ($self, $parallel) = @_;
$self->{parallel} = $parallel;
$self->idx_init;
- my $max_git = -1;
- git_dir_latest($self, \$max_git);
- $self->git_init($max_git >= 0 ? $max_git : 0);
+ my $epoch_max = -1;
+ git_dir_latest($self, \$epoch_max);
+ $self->git_init($epoch_max >= 0 ? $epoch_max : 0);
$self->done;
}
@@ -115,7 +115,7 @@ sub add {
my $nparts = $self->{partitions};
my $part = $num % $nparts;
- $self->{last_commit}->[$self->{max_git}] = $cmt;
+ $self->{last_commit}->[$self->{epoch_max}] = $cmt;
my $idx = $self->idx_part($part);
$idx->index_raw($len, $msgref, $num, $oid, $mid0, $mime);
my $n = $self->{transact_bytes} += $len;
@@ -242,7 +242,7 @@ sub purge_oids {
$self->done;
my $pfx = "$self->{-inbox}->{mainrepo}/git";
my $purges = [];
- foreach my $i (0..$self->{max_git}) {
+ foreach my $i (0..$self->{epoch_max}) {
my $git = PublicInbox::Git->new("$pfx/$i.git");
my $im = $self->import_init($git, 0, 1);
$purges->[$i] = $im->purge_oids($purge);
@@ -315,7 +315,7 @@ sub remove_internal {
if (defined $mark) {
my $cmt = $im->get_mark($mark);
- $self->{last_commit}->[$self->{max_git}] = $cmt;
+ $self->{last_commit}->[$self->{epoch_max}] = $cmt;
}
if ($purge && scalar keys %$purge) {
return purge_oids($self, $purge);
@@ -347,9 +347,9 @@ sub last_commit_part ($$;$) {
sub set_last_commits ($) {
my ($self) = @_;
- defined(my $max_git = $self->{max_git}) or return;
+ defined(my $epoch_max = $self->{epoch_max}) or return;
my $last_commit = $self->{last_commit};
- foreach my $i (0..$max_git) {
+ foreach my $i (0..$epoch_max) {
defined(my $cmt = $last_commit->[$i]) or next;
$last_commit->[$i] = undef;
last_commit_part($self, $i, $cmt);
@@ -430,9 +430,9 @@ sub barrier {
}
sub git_init {
- my ($self, $new) = @_;
+ my ($self, $epoch) = @_;
my $pfx = "$self->{-inbox}->{mainrepo}/git";
- my $git_dir = "$pfx/$new.git";
+ my $git_dir = "$pfx/$epoch.git";
my @cmd = (qw(git init --bare -q), $git_dir);
PublicInbox::Import::run_die(\@cmd);
@@ -450,7 +450,7 @@ sub git_init {
PublicInbox::Import::run_die(\@cmd);
my $alt = "$all/objects/info/alternates";
- my $new_obj_dir = "../../git/$new.git/objects";
+ my $new_obj_dir = "../../git/$epoch.git/objects";
my %alts;
if (-e $alt) {
open(my $fh, '<', $alt) or die "open < $alt: $!\n";
@@ -491,26 +491,26 @@ sub importer {
$im->done;
$self->barrier(1);
$im = undef;
- my $git_dir = $self->git_init(++$self->{max_git});
+ my $git_dir = $self->git_init(++$self->{epoch_max});
my $git = PublicInbox::Git->new($git_dir);
return $self->import_init($git, 0);
}
}
- my $new = 0;
+ my $epoch = 0;
my $max;
my $latest = git_dir_latest($self, \$max);
if (defined $latest) {
my $git = PublicInbox::Git->new($latest);
my $packed_bytes = $git->packed_bytes;
if ($packed_bytes >= $self->{rotate_bytes}) {
- $new = $max + 1;
+ $epoch = $max + 1;
} else {
- $self->{max_git} = $max;
+ $self->{epoch_max} = $max;
return $self->import_init($git, $packed_bytes);
}
}
- $self->{max_git} = $new;
- $latest = $self->git_init($new);
+ $self->{epoch_max} = $epoch;
+ $latest = $self->git_init($epoch);
$self->import_init(PublicInbox::Git->new($latest), 0);
}
@@ -690,9 +690,9 @@ sub update_last_commit {
sub git_dir_n ($$) { "$_[0]->{-inbox}->{mainrepo}/git/$_[1].git" }
sub last_commits {
- my ($self, $max_git) = @_;
+ my ($self, $epoch_max) = @_;
my $heads = [];
- for (my $i = $max_git; $i >= 0; $i--) {
+ for (my $i = $epoch_max; $i >= 0; $i--) {
$heads->[$i] = last_commit_part($self, $i);
}
$heads;
@@ -710,10 +710,10 @@ sub is_ancestor ($$$) {
}
sub index_prepare {
- my ($self, $opts, $max_git, $ranges) = @_;
+ my ($self, $opts, $epoch_max, $ranges) = @_;
my $regen_max = 0;
my $head = $self->{-inbox}->{ref_head} || 'refs/heads/master';
- for (my $i = $max_git; $i >= 0; $i--) {
+ for (my $i = $epoch_max; $i >= 0; $i--) {
die "already indexing!\n" if $self->{index_pipe};
my $git_dir = git_dir_n($self, $i);
-d $git_dir or next; # missing parts are fine
@@ -822,15 +822,15 @@ sub index_sync {
my ($self, $opts) = @_;
$opts ||= {};
my $ibx = $self->{-inbox};
- my $max_git;
- my $latest = git_dir_latest($self, \$max_git);
+ my $epoch_max;
+ my $latest = git_dir_latest($self, \$epoch_max);
return unless defined $latest;
$self->idx_init; # acquire lock
my $mm_tmp = $self->{mm}->tmp_clone;
- my $ranges = $opts->{reindex} ? [] : $self->last_commits($max_git);
+ my $ranges = $opts->{reindex} ? [] : $self->last_commits($epoch_max);
my ($min, $max) = $mm_tmp->minmax;
- my $regen = $self->index_prepare($opts, $max_git, $ranges);
+ my $regen = $self->index_prepare($opts, $epoch_max, $ranges);
$$regen += $max if $max;
my $D = {};
my @cmd = qw(log --raw -r --pretty=tformat:%h
@@ -838,7 +838,7 @@ sub index_sync {
# work backwards through history
my $last_commit = [];
- for (my $i = $max_git; $i >= 0; $i--) {
+ for (my $i = $epoch_max; $i >= 0; $i--) {
my $git_dir = git_dir_n($self, $i);
die "already reindexing!\n" if delete $self->{reindex_pipe};
-d $git_dir or next; # missing parts are fine
--
EW
^ permalink raw reply related [flat|nested] 5+ messages in thread* [PATCH 4/4] search: index and allow searching by date-time
2018-04-05 21:45 [PATCH 0/4] a few more v2 updates Eric Wong (Contractor, The Linux Foundation)
` (2 preceding siblings ...)
2018-04-05 21:45 ` [PATCH 3/4] over: use only supported and safe SQLite APIs Eric Wong (Contractor, The Linux Foundation)
@ 2018-04-05 21:45 ` Eric Wong (Contractor, The Linux Foundation)
3 siblings, 0 replies; 5+ messages in thread
From: Eric Wong (Contractor, The Linux Foundation) @ 2018-04-05 21:45 UTC (permalink / raw)
To: meta
Dscho found this useful for finding matching git commits based
on AuthorDate in git. Add it to the overview DB format, too;
so in the future we can support v2 repos without Xapian.
https://public-inbox.org/git/nycvar.QRO.7.76.6.1804041821420.55@ZVAVAG-6OXH6DA.rhebcr.pbec.zvpebfbsg.pbz
https://public-inbox.org/git/alpine.DEB.2.20.1702041206130.3496@virtualbox/
---
lib/PublicInbox/OverIdx.pm | 10 ++++++----
lib/PublicInbox/Search.pm | 8 +++++++-
lib/PublicInbox/SearchIdx.pm | 11 ++++++++---
lib/PublicInbox/SearchMsg.pm | 8 +++++---
t/over.t | 12 ++++++------
t/search.t | 7 +++++++
6 files changed, 39 insertions(+), 17 deletions(-)
diff --git a/lib/PublicInbox/OverIdx.pm b/lib/PublicInbox/OverIdx.pm
index 5c20f1f..28e4aa9 100644
--- a/lib/PublicInbox/OverIdx.pm
+++ b/lib/PublicInbox/OverIdx.pm
@@ -202,7 +202,7 @@ sub link_refs {
sub add_over {
my ($self, $values) = @_;
- my ($ts, $num, $mids, $refs, $xpath, $ddd) = @$values;
+ my ($ts, $ds, $num, $mids, $refs, $xpath, $ddd) = @$values;
my $old_tid;
my $vivified = 0;
@@ -232,11 +232,11 @@ sub add_over {
my $sid = $self->sid($xpath);
my $dbh = $self->{dbh};
my $sth = $dbh->prepare_cached(<<'');
-INSERT INTO over (num, tid, sid, ts, ddd)
-VALUES (?,?,?,?,?)
+INSERT INTO over (num, tid, sid, ts, ds, ddd)
+VALUES (?,?,?,?,?,?)
my $n = 0;
- my @v = ($num, $tid, $sid, $ts);
+ my @v = ($num, $tid, $sid, $ts, $ds);
foreach (@v) { $sth->bind_param(++$n, $_) }
$sth->bind_param(++$n, $ddd, SQL_BLOB);
$sth->execute;
@@ -274,6 +274,7 @@ CREATE TABLE IF NOT EXISTS over (
tid INTEGER NOT NULL,
sid INTEGER,
ts INTEGER,
+ ds INTEGER,
ddd VARBINARY, /* doc-data-deflated */
UNIQUE (num)
)
@@ -281,6 +282,7 @@ CREATE TABLE IF NOT EXISTS over (
$dbh->do('CREATE INDEX IF NOT EXISTS idx_tid ON over (tid)');
$dbh->do('CREATE INDEX IF NOT EXISTS idx_sid ON over (sid)');
$dbh->do('CREATE INDEX IF NOT EXISTS idx_ts ON over (ts)');
+ $dbh->do('CREATE INDEX IF NOT EXISTS idx_ds ON over (ds)');
$dbh->do(<<'');
CREATE TABLE IF NOT EXISTS counter (
diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm
index 9eb0728..34ebd1a 100644
--- a/lib/PublicInbox/Search.pm
+++ b/lib/PublicInbox/Search.pm
@@ -9,7 +9,8 @@ use warnings;
# values for searching
use constant TS => 0; # Received: header in Unix time
-use constant YYYYMMDD => 1; # for searching in the WWW UI
+use constant YYYYMMDD => 1; # Date: header for searching in the WWW UI
+use constant DT => 2; # Date: YYYYMMDDHHMMSS
use Search::Xapian qw/:standard/;
use PublicInbox::SearchMsg;
@@ -88,6 +89,9 @@ our @HELP = (
date range as YYYYMMDD e.g. d:19931002..20101002
Open-ended ranges such as d:19931002.. and d:..20101002
are also supported
+EOF
+ 'dt:' => <<EOF,
+date-time range as YYYYMMDDhhmmss (e.g. dt:19931002011000..19931002011200)
EOF
'b:' => 'match within message body, including text attachments',
'nq:' => 'match non-quoted text within message body',
@@ -258,6 +262,8 @@ sub qp {
$qp->set_stemming_strategy(STEM_SOME);
$qp->add_valuerangeprocessor(
Search::Xapian::NumberValueRangeProcessor->new(YYYYMMDD, 'd:'));
+ $qp->add_valuerangeprocessor(
+ Search::Xapian::NumberValueRangeProcessor->new(DT, 'dt:'));
while (my ($name, $prefix) = each %bool_pfx_external) {
$qp->add_boolean_prefix($name, $prefix);
diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index 06bce70..4256263 100644
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -282,9 +282,14 @@ sub add_message {
$smsg->{lines} = $mime->body_raw =~ tr!\n!\n!;
defined $bytes or $bytes = length($mime->as_string);
$smsg->{bytes} = $bytes;
+
add_val($doc, PublicInbox::Search::TS(), $smsg->ts);
- my $yyyymmdd = strftime('%Y%m%d', gmtime($smsg->ds));
- add_val($doc, PublicInbox::Search::YYYYMMDD, $yyyymmdd);
+ my @ds = gmtime($smsg->ds);
+ my $yyyymmdd = strftime('%Y%m%d', @ds);
+ add_val($doc, PublicInbox::Search::YYYYMMDD(), $yyyymmdd);
+ my $dt = strftime('%Y%m%d%H%M%S', @ds);
+ add_val($doc, PublicInbox::Search::DT(), $dt);
+ my @vals = ($smsg->{ts}, $smsg->{ds});
my $tg = $self->term_generator;
@@ -355,7 +360,7 @@ sub add_message {
utf8::encode($data);
$data = compress($data);
- my @vals = ($smsg->ts, $num, $mids, $refs, $xpath, $data);
+ push @vals, $num, $mids, $refs, $xpath, $data;
$self->{over}->add_over(\@vals);
$doc->add_boolean_term('Q' . $_) foreach @$mids;
$doc->add_boolean_term('XNUM' . $num) if defined $num;
diff --git a/lib/PublicInbox/SearchMsg.pm b/lib/PublicInbox/SearchMsg.pm
index d43853a..3278802 100644
--- a/lib/PublicInbox/SearchMsg.pm
+++ b/lib/PublicInbox/SearchMsg.pm
@@ -9,6 +9,7 @@ use warnings;
use PublicInbox::MID qw/mid_clean mid_mime/;
use PublicInbox::Address;
use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
+use Time::Local qw(timegm);
sub new {
my ($class, $mime) = @_;
@@ -44,7 +45,6 @@ sub to_doc_data {
$self->cc,
$oid,
$mid0,
- $self->ds,
$self->{bytes},
$self->{lines}
);
@@ -65,7 +65,6 @@ sub load_from_data ($$) {
$self->{blob},
$self->{mid},
- $self->{ds},
$self->{bytes},
$self->{lines}
) = split(/\n/, $_[1]);
@@ -75,7 +74,10 @@ sub load_expand {
my ($self) = @_;
my $doc = $self->{doc};
my $data = $doc->get_data or return;
- $self->{ts} = get_val($doc, &PublicInbox::Search::TS);
+ $self->{ts} = get_val($doc, PublicInbox::Search::TS());
+ my $dt = get_val($doc, PublicInbox::Search::DT());
+ my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $dt);
+ $self->{ds} = timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy);
utf8::decode($data);
load_from_data($self, $data);
$self;
diff --git a/t/over.t b/t/over.t
index 2a7e8d1..c0d9d5e 100644
--- a/t/over.t
+++ b/t/over.t
@@ -38,21 +38,21 @@ is($y, $x + 1, 'integer tid for ghost increases');
my $ddd = compress('');
foreach my $s ('', undef) {
- $over->add_over([0, 98, [ 'a' ], [], $s, $ddd]);
- $over->add_over([0, 99, [ 'b' ], [], $s, $ddd]);
+ $over->add_over([0, 0, 98, [ 'a' ], [], $s, $ddd]);
+ $over->add_over([0, 0, 99, [ 'b' ], [], $s, $ddd]);
my $msgs = [ map { $_->{num} } @{$over->get_thread('a')} ];
is_deeply([98], $msgs,
'messages not linked by empty subject');
}
-$over->add_over([0, 98, [ 'a' ], [], 's', $ddd]);
-$over->add_over([0, 99, [ 'b' ], [], 's', $ddd]);
+$over->add_over([0, 0, 98, [ 'a' ], [], 's', $ddd]);
+$over->add_over([0, 0, 99, [ 'b' ], [], 's', $ddd]);
foreach my $mid (qw(a b)) {
my $msgs = [ map { $_->{num} } @{$over->get_thread('a')} ];
is_deeply([98, 99], $msgs, 'linked messages by subject');
}
-$over->add_over([0, 98, [ 'a' ], [], 's', $ddd]);
-$over->add_over([0, 99, [ 'b' ], ['a'], 'diff', $ddd]);
+$over->add_over([0, 0, 98, [ 'a' ], [], 's', $ddd]);
+$over->add_over([0, 0, 99, [ 'b' ], ['a'], 'diff', $ddd]);
foreach my $mid (qw(a b)) {
my $msgs = [ map { $_->{num} } @{$over->get_thread($mid)} ];
is_deeply([98, 99], $msgs, "linked messages by Message-ID: <$mid>");
diff --git a/t/search.t b/t/search.t
index c9bef71..2f7b795 100644
--- a/t/search.t
+++ b/t/search.t
@@ -170,6 +170,13 @@ sub filter_mids {
# body
$res = $ro->query('goodbye');
is($res->[0]->mid, 'last@s', 'got goodbye message body');
+
+ # datestamp
+ $res = $ro->query('dt:20101002000001..20101002000001');
+ @res = filter_mids($res);
+ is_deeply(\@res, ['ghost-message@s'], 'exact Date: match works');
+ $res = $ro->query('dt:20101002000002..20101002000002');
+ is_deeply($res, [], 'exact Date: match down to the second');
}
# long message-id
--
EW
^ permalink raw reply related [flat|nested] 5+ messages in thread