dumping ground for random patches and texts
 help / color / mirror / Atom feed
From: Eric Wong <e@80x24.org>
To: spew@80x24.org
Subject: [PATCH 01/16] git: decouple cat_async_retry from POSIX pipe semantics
Date: Sun,  1 Oct 2023 02:43:08 +0000	[thread overview]
Message-ID: <20231001024323.1960491-1-e@80x24.org> (raw)

While pipes guarantee writes of <= 512 bytes to be atomic,
Unix stream sockets (or TCP sockets) have no such guarantees.
Removing the pipe assumption will make it possible for us to
switch to bidirectional Unix stream sockets and save FDs with
`git cat-file' processes as we have with Gcf2Client.  The
performance benefit of larger pipe buffers over stream sockets
isn't irrelevant when interacting with git as it is with
SearchIdx shards.
---
 lib/PublicInbox/Git.pm | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)

diff --git a/lib/PublicInbox/Git.pm b/lib/PublicInbox/Git.pm
index eb88aa48..8ac40d2b 100644
--- a/lib/PublicInbox/Git.pm
+++ b/lib/PublicInbox/Git.pm
@@ -223,25 +223,21 @@ sub my_readline ($$) {
 }
 
 sub cat_async_retry ($$) {
-	my ($self, $inflight) = @_;
+	my ($self, $old_inflight) = @_;
 
 	# {inflight} may be non-existent, but if it isn't we delete it
 	# here to prevent cleanup() from waiting:
 	delete $self->{inflight};
 	cleanup($self);
+	batch_prepare($self, my $new_inflight = []);
 
-	batch_prepare($self, $inflight);
-	my $buf = '';
-	for (my $i = 0; $i < @$inflight; $i += 3) {
-		$buf .= "$inflight->[$i]\n";
+	while (my ($oid, $cb, $arg) = splice(@$old_inflight, 0, 3)) {
+		write_all($self, $self->{out}, $oid."\n",
+				\&cat_async_step, $new_inflight);
+		$oid = \$oid if !@$new_inflight; # to indicate oid retried
+		push @$new_inflight, $oid, $cb, $arg;
 	}
-	$self->{out}->blocking(1); # brand new pipe, should never block
-	print { $self->{out} } $buf or $self->fail("write error: $!");
-	$self->{out}->blocking(0);
-	my $req = shift @$inflight;
-	unshift(@$inflight, \$req); # \$ref to indicate retried
-
-	cat_async_step($self, $inflight); # take one step
+	cat_async_step($self, $new_inflight); # take one step
 }
 
 # returns true if prefetch is successful

             reply	other threads:[~2023-10-01  2:43 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-01  2:43 Eric Wong [this message]
2023-10-01  2:43 ` [PATCH 02/16] git: use Unix stream sockets for `cat-file --batch-*' Eric Wong
2023-10-01  2:43 ` [PATCH 03/16] git+gcf2client: switch to level-triggered wakeups Eric Wong
2023-10-01  2:43 ` [PATCH 04/16] gcf2: Eric Wong
2023-10-01  2:43 ` [PATCH 05/16] t/git: show git_version in diag output Eric Wong
2023-10-01  2:43 ` [PATCH 06/16] process_pipe: do not run `close' perlop unless requested Eric Wong
2023-10-01  2:43 ` [PATCH 07/16] git: improve error reporting Eric Wong
2023-10-01  2:43 ` [PATCH 08/16] git: packed_bytes: account for TOUTTOC between glob and stat Eric Wong
2023-10-01  2:43 ` [PATCH 09/16] gcf2client: warnings Eric Wong
2023-10-01  2:43 ` [PATCH 10/16] lei rediff: order-file support Eric Wong
2023-10-01  2:43 ` [PATCH 11/16] lei: correct exit signal Eric Wong
2023-10-01  2:43 ` [PATCH 12/16] lei mail-diff: remove correct temporary directory Eric Wong
2023-10-01  2:43 ` [PATCH 13/16] lei rediff: perl -e instead of -E Eric Wong
2023-10-01  2:43 ` [PATCH 14/16] lei_store: unlink early Eric Wong
2023-10-01  2:43 ` [PATCH 15/16] overidx: fix version comparison Eric Wong
2023-10-01  2:43 ` [PATCH 16/16] enable warnings globally 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

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20231001024323.1960491-1-e@80x24.org \
    --to=e@80x24.org \
    --cc=spew@80x24.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).