* [PATCH 7/9] v2: pass smsg in more places
2020-03-20 8:18 7% ` [PATCH 0/9] preserve time and date of initial commit Eric Wong
@ 2020-03-20 8:18 5% ` Eric Wong
0 siblings, 0 replies; 2+ results
From: Eric Wong @ 2020-03-20 8:18 UTC (permalink / raw)
To: meta
We can pass fewer order-dependent args to V2Writable::do_idx and
SearchIdxShard::index_raw by passing the smsg object, instead.
---
lib/PublicInbox/SearchIdxShard.pm | 27 ++++++----------
lib/PublicInbox/V2Writable.pm | 52 +++++++++++++++++++------------
2 files changed, 42 insertions(+), 37 deletions(-)
diff --git a/lib/PublicInbox/SearchIdxShard.pm b/lib/PublicInbox/SearchIdxShard.pm
index d29e6090..21e81b16 100644
--- a/lib/PublicInbox/SearchIdxShard.pm
+++ b/lib/PublicInbox/SearchIdxShard.pm
@@ -67,20 +67,19 @@ sub shard_worker_loop ($$$$$) {
$self->remove_by_oid($oid, $mid);
} else {
chomp $line;
- my ($len, $artnum, $oid, $mid0, $autime, $cotime) =
+ my ($bytes, $num, $blob, $mid, $autime, $cotime) =
split(/ /, $line);
$self->begin_txn_lazy;
- my $n = read($r, my $msg, $len) or die "read: $!\n";
- $n == $len or die "short read: $n != $len\n";
+ my $n = read($r, my $msg, $bytes) or die "read: $!\n";
+ $n == $bytes or die "short read: $n != $bytes\n";
my $mime = PublicInbox::MIME->new(\$msg);
- $artnum = int($artnum);
$self->{autime} = $autime;
$self->{cotime} = $cotime;
my $smsg = bless {
- bytes => $len,
- num => $artnum,
- blob => $oid,
- mid => $mid0,
+ bytes => $bytes,
+ num => $num + 0,
+ blob => $blob,
+ mid => $mid,
}, 'PublicInbox::Smsg';
$self->add_message($mime, $smsg);
}
@@ -90,23 +89,17 @@ sub shard_worker_loop ($$$$$) {
# called by V2Writable
sub index_raw {
- my ($self, $bytes, $msgref, $artnum, $oid, $mid0, $mime, $times) = @_;
+ my ($self, $msgref, $mime, $smsg, $times) = @_;
my $at = $times->{autime} // time;
my $ct = $times->{cotime} // time;
if (my $w = $self->{w}) {
- print $w "$bytes $artnum $oid $mid0 $at $ct\n", $$msgref or
- die "failed to write shard $!\n";
+ print $w join(' ', @$smsg{qw(bytes num blob mid)}, $at, $ct),
+ "\n", $$msgref or die "failed to write shard $!\n";
} else {
$$msgref = undef;
$self->begin_txn_lazy;
$self->{autime} = $at;
$self->{cotime} = $ct;
- my $smsg = bless {
- bytes => $bytes,
- num => $artnum,
- blob => $oid,
- mid => $mid0,
- }, 'PublicInbox::Smsg';
$self->add_message($mime, $smsg);
}
}
diff --git a/lib/PublicInbox/V2Writable.pm b/lib/PublicInbox/V2Writable.pm
index 34dd139b..b5332da4 100644
--- a/lib/PublicInbox/V2Writable.pm
+++ b/lib/PublicInbox/V2Writable.pm
@@ -148,18 +148,12 @@ sub add {
}
# indexes a message, returns true if checkpointing is needed
-sub do_idx ($$$$$$$) {
- my ($self, $msgref, $mime, $len, $num, $oid, $mid0) = @_;
- my $smsg = bless {
- bytes => $len,
- num => $num,
- blob => $oid,
- mid => $mid0,
- }, 'PublicInbox::Smsg';
+sub do_idx ($$$$) {
+ my ($self, $msgref, $mime, $smsg) = @_;
$self->{over}->add_overview($mime, $smsg, $self);
- my $idx = idx_shard($self, $num % $self->{shards});
- $idx->index_raw($len, $msgref, $num, $oid, $mid0, $mime, $self);
- my $n = $self->{transact_bytes} += $len;
+ my $idx = idx_shard($self, $smsg->{num} % $self->{shards});
+ $idx->index_raw($msgref, $mime, $smsg, $self);
+ my $n = $self->{transact_bytes} += $smsg->{bytes};
$n >= (PublicInbox::SearchIdx::BATCH_BYTES * $self->{shards});
}
@@ -186,8 +180,10 @@ sub _add {
$cmt = $im->get_mark($cmt);
$self->{last_commit}->[$self->{epoch_max}] = $cmt;
- my ($oid, $len, $msgref) = @{$im->{last_object}};
- if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) {
+ my $msgref;
+ my $smsg = bless { mid => $mid0, num => $num }, 'PublicInbox::Smsg';
+ ($smsg->{blob}, $smsg->{bytes}, $msgref) = @{$im->{last_object}};
+ if (do_idx($self, $msgref, $mime, $smsg)) {
$self->checkpoint;
}
@@ -557,17 +553,21 @@ W: $list
}
# make sure we really got the OID:
- my ($oid, $type, $len) = $self->{-inbox}->git->check($expect_oid);
- $oid eq $expect_oid or die "BUG: $expect_oid not found after replace";
+ my ($blob, $type, $bytes) = $self->{-inbox}->git->check($expect_oid);
+ $blob eq $expect_oid or die "BUG: $expect_oid not found after replace";
# don't leak FDs to Xapian:
$self->{-inbox}->git->cleanup;
# reindex modified messages:
for my $smsg (@$need_reindex) {
- my $num = $smsg->{num};
- my $mid0 = $smsg->{mid};
- do_idx($self, \$raw, $new_mime, $len, $num, $oid, $mid0);
+ my $new_smsg = bless {
+ blob => $blob,
+ bytes => $bytes,
+ num => $smsg->{num},
+ mid => $smsg->{mid},
+ }, 'PublicInbox::Smsg';
+ do_idx($self, \$raw, $new_mime, $new_smsg);
}
$rewritten->{rewrites};
}
@@ -955,7 +955,13 @@ sub reindex_oid_m ($$$$;$) {
}
}
$sync->{nr}++;
- if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) {
+ my $smsg = bless {
+ bytes => $len,
+ num => $num,
+ blob => $oid,
+ mid => $mid0,
+ }, 'PublicInbox::Smsg';
+ if (do_idx($self, $msgref, $mime, $smsg)) {
reindex_checkpoint($self, $sync, $git);
}
}
@@ -1051,7 +1057,13 @@ sub reindex_oid ($$$$) {
$sync->{mm_tmp}->mid_delete($mid0) or
die "failed to delete <$mid0> for article #$num\n";
$sync->{nr}++;
- if (do_idx($self, $msgref, $mime, $len, $num, $oid, $mid0)) {
+ my $smsg = bless {
+ bytes => $len,
+ num => $num,
+ blob => $oid,
+ mid => $mid0,
+ }, 'PublicInbox::Smsg';
+ if (do_idx($self, $msgref, $mime, $smsg)) {
reindex_checkpoint($self, $sync, $git);
}
}
^ permalink raw reply related [relevance 5%]
* [PATCH 0/9] preserve time and date of initial commit
@ 2020-03-20 8:18 7% ` Eric Wong
2020-03-20 8:18 5% ` [PATCH 7/9] v2: pass smsg in more places Eric Wong
0 siblings, 1 reply; 2+ results
From: Eric Wong @ 2020-03-20 8:18 UTC (permalink / raw)
To: meta
For messages lacking Date and/or Received headers, search
queries for "d:YYYYMMDD..YYYYMMDD" ranges can be unreliable in
mirrors, as can the $INBOX_URL/?t=$TIMESTAMP query which
only hits SQLite.
Yes, this ended up being a lot of work to deal with corner case
messages (probably most of which are spam), but there's also a
lot of internal cleanups which made the end result easier to
follow, I think...
The main fix is actually in 1/9, but it's gross.
Patch 2/9 fixes a small window where a race can happen and
cause searches to be off by a minute.
Patches 3-8 cleanup the mess left in 1 and 2,
Finally, patch 9 fixes the corner-case-of-corner-cases for
dealing with multi-MID messages which require a one-off queue to
store the git commit/author times instead of overloading msgmap.
Eric Wong (9):
index: use git commit times on missing Date/Received
v2writable: preserve timestamps from import if generated
rename PublicInbox::SearchMsg => PublicInbox::Smsg
smsg: to_doc_data: use existing fields
overidx: parse_references: less error-prone args
*idx: pass $smsg in more places instead of many args
v2: pass smsg in more places
*idx: pass smsg in even more places
v2: SDBM-based multi Message-ID queue
Documentation/mknews.perl | 4 +-
Documentation/technical/data_structures.txt | 4 +-
MANIFEST | 4 +-
lib/PublicInbox/ExtMsg.pm | 3 +-
lib/PublicInbox/Feed.pm | 4 +-
lib/PublicInbox/Import.pm | 19 +++--
lib/PublicInbox/Inbox.pm | 2 +-
lib/PublicInbox/Mbox.pm | 4 +-
lib/PublicInbox/MsgTime.pm | 12 +--
lib/PublicInbox/MultiMidQueue.pm | 57 +++++++++++++
lib/PublicInbox/NNTP.pm | 14 ++--
lib/PublicInbox/Over.pm | 8 +-
lib/PublicInbox/OverIdx.pm | 30 +++----
lib/PublicInbox/Search.pm | 8 +-
lib/PublicInbox/SearchIdx.pm | 68 ++++++++++-----
lib/PublicInbox/SearchIdxShard.pm | 26 ++++--
lib/PublicInbox/SearchView.pm | 8 +-
lib/PublicInbox/{SearchMsg.pm => Smsg.pm} | 19 +++--
lib/PublicInbox/SolverGit.pm | 2 +-
lib/PublicInbox/V2Writable.pm | 84 ++++++++++++-------
lib/PublicInbox/View.pm | 2 +-
t/import.t | 14 ++--
t/index-git-times.t | 93 +++++++++++++++++++++
t/multi-mid.t | 27 +++++-
t/search-thr-index.t | 17 +++-
t/thread-cycle.t | 2 +-
26 files changed, 386 insertions(+), 149 deletions(-)
create mode 100644 lib/PublicInbox/MultiMidQueue.pm
rename lib/PublicInbox/{SearchMsg.pm => Smsg.pm} (92%)
create mode 100644 t/index-git-times.t
^ permalink raw reply [relevance 7%]
Results 1-2 of 2 | reverse | options above
-- pct% links below jump to the message on this page, permalinks otherwise --
2020-03-05 5:13 [PATCH] index: use git commit times on missing Date/Received Eric Wong
2020-03-20 8:18 7% ` [PATCH 0/9] preserve time and date of initial commit Eric Wong
2020-03-20 8:18 5% ` [PATCH 7/9] v2: pass smsg in more places Eric Wong
Code repositories for project(s) associated with this public inbox
https://80x24.org/public-inbox.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).