From: Eric Wong <e@80x24.org>
To: meta@public-inbox.org
Subject: [PATCH 14/15] lei_mail_sync: mv_src: use transaction, check UNIQUE
Date: Thu, 21 Oct 2021 21:10:31 +0000 [thread overview]
Message-ID: <20211021211032.22666-15-e@80x24.org> (raw)
In-Reply-To: <20211021211032.22666-1-e@80x24.org>
We need a transaction across two SQL statements so readers
(which don't use flock) will see the result as atomic.
This may help against some occasional test failures I'm seeing
from t/lei-auto-watch.t and t/lei-watch.t, or make the problem
more apparent.
---
lib/PublicInbox/LeiMailSync.pm | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/lib/PublicInbox/LeiMailSync.pm b/lib/PublicInbox/LeiMailSync.pm
index e70cb5de2b6b..124eb96948c5 100644
--- a/lib/PublicInbox/LeiMailSync.pm
+++ b/lib/PublicInbox/LeiMailSync.pm
@@ -182,16 +182,20 @@ sub mv_src {
my ($self, $folder, $oidbin, $id, $newbn) = @_;
my $lk = $self->lock_for_scope;
my $fid = $self->{fmap}->{$folder} //= fid_for($self, $folder, 1);
+ $self->{dbh}->begin_work;
my $sth = $self->{dbh}->prepare_cached(<<'');
UPDATE blob2name SET name = ? WHERE fid = ? AND oidbin = ? AND name = ?
- my $nr = $sth->execute($newbn, $fid, $oidbin, $$id);
- if ($nr == 0) { # may race with a clear_src, ensure new value exists
+ # eval since unique constraint may fail due to race
+ my $nr = eval { $sth->execute($newbn, $fid, $oidbin, $$id) };
+ if (!defined($nr) || $nr == 0) { # $nr may be `0E0'
+ # may race with a clear_src, ensure new value exists
$sth = $self->{dbh}->prepare_cached(<<'');
INSERT OR IGNORE INTO blob2name (oidbin, fid, name) VALUES (?, ?, ?)
$sth->execute($oidbin, $fid, $newbn);
}
+ $self->{dbh}->commit;
}
# read-only, iterates every oidbin + UID or name for a given folder
next prev parent reply other threads:[~2021-10-21 21:10 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-10-21 21:10 [PATCH 00/15] use RENAME_NOREPLACE on Linux 3.15+ Eric Wong
2021-10-21 21:10 ` [PATCH 01/15] t/lei-{auto-watch,export-kw}: extra diagnostics on failure Eric Wong
2021-10-21 21:10 ` [PATCH 02/15] t/lei-import-maildir: rename fix (SR -> RS) Eric Wong
2021-10-21 21:10 ` [PATCH 03/15] t/lei-p2q: extra diagnostics Eric Wong
2023-09-21 10:23 ` Eric Wong
2021-10-21 21:10 ` [PATCH 04/15] lei/store: check for any unexpected process death Eric Wong
2021-10-21 21:10 ` [PATCH 05/15] lei note-event: drop unnecessary eval guard Eric Wong
2021-10-21 21:10 ` [PATCH 06/15] lei note-event: wq_io_do => wq_do Eric Wong
2021-10-21 21:10 ` [PATCH 07/15] lei_search: try harder to associate "lei index"-ed messages Eric Wong
2021-10-21 21:10 ` [PATCH 08/15] watch: check for {quit} before IDLE Eric Wong
2021-10-21 21:10 ` [PATCH 09/15] watch: remove redundant signal mask manipulation Eric Wong
2021-10-21 21:10 ` [PATCH 10/15] doc: lei-overview: add CAVEATS section Eric Wong
2021-10-21 21:10 ` [PATCH 11/15] lei note-event: clear_src on ENOENT Eric Wong
2021-10-21 21:10 ` [PATCH 12/15] dir_idle: treat IN_MOVED_FROM as a gone event Eric Wong
2021-10-21 21:10 ` [PATCH 13/15] lei: no Perl FileHandle for `undef' w/ ECONNRESET Eric Wong
2021-10-21 21:10 ` Eric Wong [this message]
2021-10-21 21:10 ` [PATCH 15/15] lei: use RENAME_NOREPLACE on Linux 3.15+ Eric Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://public-inbox.org/README
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211021211032.22666-15-e@80x24.org \
--to=e@80x24.org \
--cc=meta@public-inbox.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).