user/dev discussion of public-inbox itself
 help / color / mirror / code / Atom feed
From: Eric Wong <e@yhbt.net>
To: meta@public-inbox.org
Subject: [PATCH 80/82] imap: remove non-UID SEARCH for now
Date: Wed, 10 Jun 2020 07:05:17 +0000	[thread overview]
Message-ID: <20200610070519.18252-81-e@yhbt.net> (raw)
In-Reply-To: <20200610070519.18252-1-e@yhbt.net>

Supporting MSNs in long-lived connections beyond the lifetime of
a single request/response cycle is not scalable to a C10K
scenario.  It's probably not needed, since most clients seem to
use UIDs.

A somewhat efficient implementation I can come up uses
pack("S*" ...) (AKA "uint16_t mapping[50000]") has an overhead
of 100K per-client socket on a mailbox with 50K messages.  The
100K is a contiguous scalar, so it could be swapped out for
idle clients on most architectures if THP is disabled.

An alternative could be to use a tempfile as an allocator
partitioned into 100K chunks (or SQLite); but I'll only do that
if somebody presents a compelling case to support MSN SEARCH.
---
 lib/PublicInbox/IMAP.pm | 33 +++++++++------------------------
 t/imapd.t               | 37 -------------------------------------
 2 files changed, 9 insertions(+), 61 deletions(-)

diff --git a/lib/PublicInbox/IMAP.pm b/lib/PublicInbox/IMAP.pm
index 1ba3a3ff1bb..7efe5387646 100644
--- a/lib/PublicInbox/IMAP.pm
+++ b/lib/PublicInbox/IMAP.pm
@@ -986,21 +986,14 @@ sub parse_date ($) { # 02-Oct-1993
 	timegm(0, 0, 0, $dd, $mm, $yyyy);
 }
 
-sub msn_convert ($$) {
-	my ($self, $uids) = @_;
-	my $adj = $self->{uid_base};
-	$_ -= $adj for @$uids;
-}
-
 sub search_uid_range { # long_response
-	my ($self, $tag, $sql, $range_info, $want_msn) = @_;
+	my ($self, $tag, $sql, $range_info) = @_;
 	my $uids = [];
 	if (defined(my $err = refill_uids($self, $uids, $range_info, $sql))) {
 		$err ||= 'OK Search done';
 		$self->write("\r\n$tag $err\r\n");
 		return;
 	}
-	msn_convert($self, $uids) if $want_msn;
 	$self->msg_more(join(' ', '', @$uids));
 	1; # more
 }
@@ -1162,46 +1155,38 @@ sub refill_xap ($$$$) {
 }
 
 sub search_xap_range { # long_response
-	my ($self, $tag, $q, $range_info, $want_msn) = @_;
+	my ($self, $tag, $q, $range_info) = @_;
 	my $uids = [];
 	if (defined(my $err = refill_xap($self, $uids, $range_info, $q))) {
 		$err ||= 'OK Search done';
 		$self->write("\r\n$tag $err\r\n");
 		return;
 	}
-	msn_convert($self, $uids) if $want_msn;
 	$self->msg_more(join(' ', '', @$uids));
 	1; # more
 }
 
-sub search_common {
-	my ($self, $tag, $rest, $want_msn) = @_;
+sub cmd_uid_search ($$$;) {
+	my ($self, $tag) = splice(@_, 0, 2);
 	my $ibx = $self->{ibx} or return "$tag BAD No mailbox selected\r\n";
-	my $q = parse_query($self, $rest);
+	my $q = parse_query($self, \@_);
 	return "$tag $q\r\n" if !ref($q);
 	my ($sql, $range_info) = delete @$q{qw(sql range_info)};
 	if (!scalar(keys %$q)) { # overview.sqlite3
 		$self->msg_more('* SEARCH');
 		long_response($self, \&search_uid_range,
-				$tag, $sql, $range_info, $want_msn);
+				$tag, $sql, $range_info);
 	} elsif ($q = $q->{xap}) {
 		$self->msg_more('* SEARCH');
 		long_response($self, \&search_xap_range,
-				$tag, $q, $range_info, $want_msn);
+				$tag, $q, $range_info);
 	} else {
 		"$tag BAD Error\r\n";
 	}
 }
 
-sub cmd_uid_search ($$$;) {
-	my ($self, $tag) = splice(@_, 0, 2);
-	search_common($self, $tag, \@_);
-}
-
-sub cmd_search ($$$;) {
-	my ($self, $tag) = splice(@_, 0, 2);
-	search_common($self, $tag, \@_, 1);
-}
+# note: MSN SEARCH is NOT supported.  Do any widely-used MUAs
+# rely on MSNs from SEARCH results?  Let us know at meta@public-inbox.org
 
 sub args_ok ($$) { # duplicated from PublicInbox::NNTP
 	my ($cb, $argc) = @_;
diff --git a/t/imapd.t b/t/imapd.t
index c7e0baaa10b..f5ca8b7eb03 100644
--- a/t/imapd.t
+++ b/t/imapd.t
@@ -360,43 +360,6 @@ is(scalar keys %$ret, 3, 'got all 3 messages with comma-separated sequence');
 $ret = $mic->fetch_hash('1:*', 'RFC822') or BAIL_OUT "FETCH $@";
 is(scalar keys %$ret, 3, 'got all 3 messages');
 
-SKIP: {
-	# do any clients use non-UID IMAP SEARCH?
-	skip 'Xapian missing', 2 if $level eq 'basic';
-	my $x = $mic->search('all');
-	is_deeply($x, [1, 2, 3], 'MSN SEARCH works before rm');
-	$x = $mic->search(qw(header subject embedded));
-	is_deeply($x, [2], 'MSN SEARCH on Subject works before rm');
-}
-
-{
-	my $rdr = { 0 => \($ret->{1}->{RFC822}) };
-	my $env = { HOME => $ENV{HOME} };
-	my @cmd = qw(-learn rm --all);
-	run_script(\@cmd, $env, $rdr) or BAIL_OUT('-learn rm');
-}
-
-SKIP: {
-	# do any clients use non-UID IMAP SEARCH?  We only ensure
-	# MSN "SEARCH" can return a result which can be retrieved
-	# via MSN "FETCH"
-	skip 'Xapian missing', 3 if $level eq 'basic';
-	my $x = $mic->search(qw(header subject embedded));
-	is(scalar(@$x), 1, 'MSN SEARCH on Subject works after rm');
-	$x = $mic->message_string($x->[0]);
-	is($x, $ret->{2}->{RFC822}, 'message 2 unchanged');
-}
-
-my $r2 = $mic->fetch_hash('1:*', 'BODY.PEEK[]') or BAIL_OUT "FETCH $@";
-is(scalar keys %$r2, 2, 'did not get all 3 messages');
-is($r2->{1}->{'BODY[]'}, $ret->{2}->{RFC822}, 'message 2 unchanged');
-is($r2->{2}->{'BODY[]'}, $ret->{3}->{RFC822}, 'message 3 unchanged');
-$r2 = $mic->fetch_hash(2, 'BODY.PEEK[HEADER.FIELDS (message-id)]')
-			or BAIL_OUT "FETCH $@";
-is($r2->{2}->{'BODY[HEADER.FIELDS (MESSAGE-ID)]'},
-	'Message-ID: <20200418222508.GA13918@dcvr>'."\r\n\r\n",
-	'BODY.PEEK[HEADER.FIELDS ...] drops .PEEK');
-
 {
 	my @new_list = $mic->list;
 	# tag differs in [-1]

  parent reply	other threads:[~2020-06-10  7:08 UTC|newest]

Thread overview: 84+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-06-10  7:03 [PATCH 00/82] public-inbox-imapd: read-only IMAP server Eric Wong
2020-06-10  7:03 ` [PATCH 01/82] doc: add some IMAP standards Eric Wong
2020-06-10  7:03 ` [PATCH 02/82] nntpd: restrict allowed newsgroup names Eric Wong
2020-06-10  7:04 ` [PATCH 03/82] preliminary imap server implementation Eric Wong
2020-06-10  7:04 ` [PATCH 04/82] inboxidle: new class to detect inbox changes Eric Wong
2020-06-10  7:04 ` [PATCH 05/82] imap: support IDLE Eric Wong
2020-06-10  7:04 ` [PATCH 06/82] msgmap: split ->max into its own method Eric Wong
2020-06-10  7:04 ` [PATCH 07/82] imap: delay InboxIdle start, support refresh Eric Wong
2020-06-10  7:04 ` [PATCH 08/82] imap: implement STATUS command Eric Wong
2020-06-10  7:04 ` [PATCH 09/82] imap: use Text::ParseWords::parse_line to handle quoted words Eric Wong
2020-06-10  7:04 ` [PATCH 10/82] imap: support LIST command Eric Wong
2020-06-10  7:04 ` [PATCH 11/82] t/imapd: support FakeInotify and KQNotify Eric Wong
2020-06-10  7:04 ` [PATCH 12/82] imap: support fetch for BODYSTRUCTURE and BODY Eric Wong
2020-06-10  7:04 ` [PATCH 13/82] eml: each_part: single part $idx is 1 Eric Wong
2020-06-10  7:04 ` [PATCH 14/82] imap: allow fetch of partial of BODY[...] and headers Eric Wong
2020-06-10  7:04 ` [PATCH 15/82] imap: always include `resp-text' in responses Eric Wong
2020-06-10  7:04 ` [PATCH 16/82] imap: split out unit tests and benchmarks Eric Wong
2020-06-10  7:04 ` [PATCH 17/82] imap: fix multi-message partial header fetches Eric Wong
2020-06-10  7:04 ` [PATCH 18/82] imap: simplify partial fetch structure Eric Wong
2020-06-10  7:04 ` [PATCH 19/82] imap: support sequence number FETCH Eric Wong
2020-06-10  7:04 ` [PATCH 20/82] imap: do not include ".PEEK" in responses Eric Wong
2020-06-10  7:04 ` [PATCH 21/82] imap: support the CLOSE command Eric Wong
2020-06-10  7:04 ` [PATCH 22/82] imap: speed up HEADER.FIELDS[.NOT] range fetches Eric Wong
2020-06-10  7:04 ` [PATCH 23/82] git: async: flatten the inflight array Eric Wong
2020-06-10  7:04 ` [PATCH 24/82] git: do our own read buffering for cat-file Eric Wong
2020-06-10  7:04 ` [PATCH 25/82] imap: use git-cat-file asynchronously Eric Wong
2020-06-10  7:04 ` [PATCH 26/82] git: idle rbuf for async Eric Wong
2020-06-10  7:04 ` [PATCH 27/82] imap: support LSUB command Eric Wong
2020-06-10  7:04 ` [PATCH 28/82] imap: FETCH: support comma-delimited ranges Eric Wong
2020-06-10  7:04 ` [PATCH 29/82] add imapd compression test Eric Wong
2020-06-10  7:04 ` [PATCH 30/82] testcommon: tcp_(server|connect): BAIL_OUT on failure Eric Wong
2020-06-10  7:04 ` [PATCH 31/82] *deflate: drop invalid comment about rbuf Eric Wong
2020-06-10  7:04 ` [PATCH 32/82] imap: fix pipelining with async git Eric Wong
2020-06-10  7:04 ` [PATCH 33/82] git: cat_async: provide requested OID + "missing" on missing blobs Eric Wong
2020-06-10  7:04 ` [PATCH 34/82] git: move async_cat reference to PublicInbox::Git Eric Wong
2020-06-10  7:04 ` [PATCH 35/82] git: async: automatic retry on alternates change Eric Wong
2020-06-10  7:04 ` [PATCH 36/82] imapclient: wrapper for Mail::IMAPClient Eric Wong
2020-06-10  7:04 ` [PATCH 37/82] xt: add imapd-validate and imapd-mbsync-oimap Eric Wong
2020-06-10  7:04 ` [PATCH 38/82] imap: support out-of-bounds ranges Eric Wong
2020-06-10  7:04 ` [PATCH 39/82] xt/perf-imap-list: time refresh_inboxlist Eric Wong
2020-06-10  7:04 ` [PATCH 40/82] imap: case-insensitive mailbox name comparisons Eric Wong
2020-06-10  7:04 ` [PATCH 41/82] imap: break giant inboxes into sub-inboxes of 50K messages Eric Wong
2020-06-10  7:04 ` [PATCH 42/82] imap: start doing iterative config reloading Eric Wong
2020-06-10  7:04 ` [PATCH 43/82] imap: require ".$UID_MIN-$UID_END" suffix Eric Wong
2020-06-10  7:04 ` [PATCH 44/82] imapd: ensure LIST is sorted alphabetically, for now Eric Wong
2020-06-10  7:04 ` [PATCH 45/82] imap: omit $UID_END from mailbox name, use index Eric Wong
2020-06-10  7:04 ` [PATCH 46/82] t/config.t: always compare against git bool behavior Eric Wong
2020-06-10  7:04 ` [PATCH 47/82] xt/*: show some tunable parameters Eric Wong
2020-06-10  7:04 ` [PATCH 48/82] imap: STATUS and LIST are case-insensitive, too Eric Wong
2020-06-10  7:04 ` [PATCH 49/82] imap: EXAMINE/STATUS: return correct counts Eric Wong
2020-06-10  7:04 ` [PATCH 50/82] imap: avoid uninitialized warnings on incomplete commands Eric Wong
2020-06-10  7:04 ` [PATCH 51/82] imap: start parsing out queries for SQLite and Xapian Eric Wong
2020-06-10  7:04 ` [PATCH 52/82] imap: SEARCH: clamp results to the 50K UID range Eric Wong
2020-06-10  7:04 ` [PATCH 53/82] imap: allow UID range search on timestamps Eric Wong
2020-06-10  7:04 ` [PATCH 54/82] over: get_art: use dbh->prepare_cached Eric Wong
2020-06-10  7:04 ` [PATCH 55/82] search: index byte size of a message for IMAP search Eric Wong
2020-06-10  7:04 ` [PATCH 56/82] search: index UID for IMAP search, too Eric Wong
2020-06-10  7:04 ` [PATCH 57/82] imap: remove dummies from sequence number FETCH Eric Wong
2020-06-10  7:04 ` [PATCH 58/82] imap: compile UID FETCH to opcodes Eric Wong
2020-06-10  7:04 ` [PATCH 59/82] imap: UID FETCH: optimize for smsg-only case Eric Wong
2020-06-10  7:04 ` [PATCH 60/82] imap: UID FETCH: optimize (UID FLAGS) harder Eric Wong
2020-06-10  7:04 ` [PATCH 61/82] imap: IDLE: avoid extraneous wakeups, keep-alive Eric Wong
2020-06-10  7:04 ` [PATCH 62/82] imap: 30 minute auto-logout timer Eric Wong
2020-06-10  7:05 ` [PATCH 63/82] imap: split ->logged_in attribute into a separate class Eric Wong
2020-06-10  7:05 ` [PATCH 64/82] searchidx: v1 (re)-index uses git asynchronously Eric Wong
2020-06-10  7:05 ` [PATCH 65/82] index: account for CRLF conversion when storing bytes Eric Wong
2020-06-10  7:05 ` [PATCH 66/82] imap: rely on smsg->{bytes} for RFC822.SIZE Eric Wong
2020-06-10  7:05 ` [PATCH 67/82] imap: UID FETCH requires at least one data item Eric Wong
2020-06-10  7:05 ` [PATCH 68/82] imap: LIST shows "INBOX" in all caps Eric Wong
2020-06-10  7:05 ` [PATCH 69/82] imap: support 8000 octet lines Eric Wong
2020-06-10  7:05 ` [PATCH 70/82] imap: reinstate some message sequence number support Eric Wong
2020-06-10  7:05 ` [PATCH 71/82] imap: cleanup ->{uid_base} usage Eric Wong
2020-06-10  7:05 ` [PATCH 72/82] imap: FETCH: more granular CRLF conversion Eric Wong
2020-06-10  7:05 ` [PATCH 73/82] imap: further speed up HEADER.FIELDS FETCH requests Eric Wong
2020-06-10  7:05 ` [PATCH 74/82] imap: FETCH: try to make fake MSNs sequentially Eric Wong
2020-06-10  7:05 ` [PATCH 75/82] imap: STATUS/EXAMINE: rely on SQLite overview Eric Wong
2020-06-10  7:05 ` [PATCH 76/82] imap: UID SEARCH: support multiple ranges Eric Wong
2020-06-10  7:05 ` [PATCH 77/82] imap: wire up Xapian, MSN SEARCH and multi sequence-sets Eric Wong
2020-06-10  7:05 ` [PATCH 78/82] imap: misc cleanups and notes Eric Wong
2020-06-10  7:05 ` [PATCH 79/82] imapd: don't bother sorting LIST output Eric Wong
2020-06-10  7:05 ` Eric Wong [this message]
2020-06-10  7:05 ` [PATCH 81/82] over: uid_range: remove LIMIT Eric Wong
2020-06-10  7:05 ` [PATCH 82/82] imap: FETCH: proper MSN => UID mapping for requests Eric Wong
2020-06-12 23:49 ` [PATCH 83/82] imap: introduce memory-efficient uo2m mapping 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=20200610070519.18252-81-e@yhbt.net \
    --to=e@yhbt.net \
    --cc=meta@public-inbox.org \
    --subject='Re: [PATCH 80/82] imap: remove non-UID SEARCH for now' \
    /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

Code repositories for project(s) associated with this 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).