From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <e@80x24.org>
X-Spam-Checker-Version: SpamAssassin 3.4.2 (2018-09-13) on dcvr.yhbt.net
X-Spam-Level: 
X-Spam-Status: No, score=-4.0 required=3.0 tests=ALL_TRUSTED,BAYES_00
	shortcircuit=no autolearn=ham autolearn_force=no version=3.4.2
Received: from localhost (dcvr.yhbt.net [127.0.0.1])
	by dcvr.yhbt.net (Postfix) with ESMTP id 861681F4B4;
	Mon, 18 Jan 2021 21:19:54 +0000 (UTC)
Date: Mon, 18 Jan 2021 21:19:54 +0000
From: Eric Wong <e@80x24.org>
To: meta@public-inbox.org
Subject: Re: [PATCH 1/2] lei q: parallelize Maildir and mbox writing
Message-ID: <20210118211954.GA28437@dcvr>
References: <20210118103032.1430-1-e@80x24.org>
 <20210118103032.1430-2-e@80x24.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <20210118103032.1430-2-e@80x24.org>
List-Id: <meta.public-inbox.org>

Eric Wong <e@80x24.org> wrote:
> diff --git a/lib/PublicInbox/LeiToMail.pm b/lib/PublicInbox/LeiToMail.pm
<snip>
> @@ -14,6 +15,8 @@ use Symbol qw(gensym);
>  use IO::Handle; # ->autoflush
>  use Fcntl qw(SEEK_SET SEEK_END O_CREAT O_EXCL O_WRONLY);
>  use Errno qw(EEXIST ESPIPE ENOENT);
> +use File::Temp 0.19 (); # 0.19 for ->newdir

File::Temp is unnecessary, will squash this in:

diff --git a/lib/PublicInbox/LeiToMail.pm b/lib/PublicInbox/LeiToMail.pm
index 17d48a90..8d030227 100644
--- a/lib/PublicInbox/LeiToMail.pm
+++ b/lib/PublicInbox/LeiToMail.pm
@@ -15,7 +15,6 @@ use Symbol qw(gensym);
 use IO::Handle; # ->autoflush
 use Fcntl qw(SEEK_SET SEEK_END O_CREAT O_EXCL O_WRONLY);
 use Errno qw(EEXIST ESPIPE ENOENT);
-use File::Temp 0.19 (); # 0.19 for ->newdir
 use PublicInbox::Git;
 
 my %kw2char = ( # Maildir characters

My initial attempt at making this change didn't have the
{each_smsg_done} pipe performing keepalive.  Instead it created
another tmpdir copy of the tmpdir created by LeiXSearch->git.

That was ugly and didn't work, since it was possible for for the
original LeiXSearch->git tmpdir to go out-of-scope before the
lei2mail worker even got a chance to copy the bare directory
and its alternates file.

The {each_smsg_done} pipe is much nicer since the kernel can
keep track of in-flight pipes and doesn't inflict extra FS
activity.