gcf2: wire up read-only daemons and rm -gcf2 script

It seems easiest to have a singleton Gcf2Client client object
per daemon worker for all inboxes to use.  This reduces overall
FD usage from pipes.

The `public-inbox-gcf2' command + manpage are gone and a `$^X'
one-liner is used, instead.  This saves inodes for internal
commands and hopefully makes it easier to avoid mismatched
PERL5LIB include paths (as noticed during development :x).

We'll also make the existing cat-file process management
infrastructure more resilient to BOFHs on process killing
sprees (or in case our libgit2-based code fails on us).

(Rare) PublicInbox::WWW PSGI users NOT using public-inbox-httpd
won't automatically benefit from this change, and extra
configuration will be required (to be documented later).

1 files changed, 34 insertions, 2 deletions

diff --git a/lib/PublicInbox/Gcf2.pm b/lib/PublicInbox/Gcf2.pm
index fe76b1fd..7983c841 100644
--- a/lib/PublicInbox/Gcf2.pm
+++ b/lib/PublicInbox/Gcf2.pm
@@ -1,12 +1,13 @@
 # Copyright (C) 2020 all contributors <meta@public-inbox.org>
 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 
-# backend for public-inbox-gcf2(1) (git-cat-file based on libgit2,
-# other libgit2 stuff may go here, too)
+# backend for a git-cat-file-workalike based on libgit2,
+# other libgit2 stuff may go here, too.
 package PublicInbox::Gcf2;
 use strict;
 use PublicInbox::Spawn qw(which popen_rd);
 use Fcntl qw(LOCK_EX);
+use IO::Handle; # autoflush
 my (%CFG, $c_src, $lockfh);
 BEGIN {
         # PublicInbox::Spawn will set PERL_INLINE_DIRECTORY
@@ -54,4 +55,35 @@ use Inline C => $c_src;
 undef $c_src;
 undef %CFG;
 undef $lockfh;
+
+# Usage: $^X -MPublicInbox::Gcf2 -e 'PublicInbox::Gcf2::loop()'
+# (see lib/PublicInbox/Gcf2Client.pm)
+sub loop {
+        my $gcf2 = new();
+        STDERR->autoflush(1);
+        STDOUT->autoflush(1);
+
+        while (<STDIN>) {
+                chomp;
+                my ($oid, $git_dir) = split(/ /, $_, 2);
+                $gcf2->add_alternate("$git_dir/objects");
+                if (!$gcf2->cat_oid(1, $oid)) {
+                        # retry once if missing.  We only get unabbreviated OIDs
+                        # from SQLite or Xapian DBs, here, so malicious clients
+                        # can't trigger excessive retries:
+                        warn "I: $$ $oid missing, retrying in $git_dir\n";
+
+                        $gcf2 = new();
+                        $gcf2->add_alternate("$git_dir/objects");
+
+                        if ($gcf2->cat_oid(1, $oid)) {
+                                warn "I: $$ $oid found after retry\n";
+                        } else {
+                                warn "W: $$ $oid missing after retry\n";
+                                print "$oid missing\n"; # mimic git-cat-file
+                        }
+                }
+        }
+}
+
 1;