implement our own cat-file --batch wrapper

We use --git-dir=... instead of $ENV{GIT_DIR} because ENV changes
do not propagate easily with mod_perl.

4 files changed, 79 insertions, 32 deletions

diff --git a/MANIFEST b/MANIFEST
index b212c76b..ecdc8b5d 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -10,6 +10,7 @@ lib/PublicInbox/MDA.pm
 lib/PublicInbox/Config.pm
 lib/PublicInbox/Feed.pm
 lib/PublicInbox/Filter.pm
+lib/PublicInbox/GitCatFile.pm
 lib/PublicInbox/Hval.pm
 lib/PublicInbox/View.pm
 public-inbox-mda
diff --git a/lib/PublicInbox/Feed.pm b/lib/PublicInbox/Feed.pm
index 87b5b2a9..ad058395 100644
--- a/lib/PublicInbox/Feed.pm
+++ b/lib/PublicInbox/Feed.pm
@@ -7,7 +7,7 @@ use Email::Address;
 use Email::MIME;
 use Date::Parse qw(strptime str2time);
 use PublicInbox::Hval;
-eval { require Git }; # this is GPLv2+, so we are OK to use it
+use PublicInbox::GitCatFile;
 use constant {
         DATEFMT => '%Y-%m-%dT%H:%M:%SZ',
         MAX_PER_PAGE => 25,
@@ -39,7 +39,7 @@ sub generate {
                 updated => POSIX::strftime(DATEFMT, gmtime),
         );
 
-        my $git = try_git_pm($args->{git_dir});
+        my $git = PublicInbox::GitCatFile->new($args->{git_dir});
         each_recent_blob($args, sub {
                 my ($add) = @_;
                 add_to_feed($feed_opts, $feed, $add, $git);
@@ -59,7 +59,7 @@ sub generate_html_index {
         $title = PublicInbox::Hval->new_oneline($title)->as_html;
 
         my @messages;
-        my $git = try_git_pm($args->{git_dir});
+        my $git = PublicInbox::GitCatFile->new($args->{git_dir});
         my $last = each_recent_blob($args, sub {
                 my $mime = do_cat_mail($git, $_[0]) or return 0;
                 $mime->body_set(''); # save some memory
@@ -294,36 +294,10 @@ sub dump_html_line {
         dump_html_line($self->next, $level, $html) if $self->next;
 }
 
-sub try_git_pm {
-        my ($dir) = @_;
-        eval { Git->repository(Directory => $dir) };
-};
-
 sub do_cat_mail {
         my ($git, $path) = @_;
-        my $str;
-        if ($git) {
-                open my $fh, '>', \$str or
-                                die "failed to setup string handle: $!\n";
-                binmode $fh;
-                my $err = '';
-                my $bytes;
-                {
-                        local $SIG{__WARN__} = sub { $err .= $_[0] };
-                        $bytes = $git->cat_blob("HEAD:$path", $fh);
-                }
-                close $fh or die "failed to close string handle: $!\n";
-
-                if ($bytes < 0 && $err &&
-                                $err !~ /doesn't exist in the repository/) {
-                        warn $err;
-                }
-                return if $bytes <= 0;
-        } else {
-                $str = `git cat-file blob HEAD:$path`;
-                return if $? != 0 || length($str) == 0;
-        }
-        Email::MIME->new($str);
+        my $str = $git->cat_file("HEAD:$path");
+        Email::MIME->new($$str);
 }
 
 1;
diff --git a/lib/PublicInbox/GitCatFile.pm b/lib/PublicInbox/GitCatFile.pm
new file mode 100644
index 00000000..bdad2b56
--- /dev/null
+++ b/lib/PublicInbox/GitCatFile.pm
@@ -0,0 +1,73 @@
+# Copyright (C) 2014, Eric Wong <normalperson@yhbt.net> and all contributors
+# License: GPLv2 or later (https://www.gnu.org/licenses/gpl-2.0.txt)
+# This is based on code in Git.pm which is GPLv2, but modified to avoid
+# dependence on environment variables for compatibility with mod_perl.
+# There are also API changes to simplify our usage and data set.
+package PublicInbox::GitCatFile;
+use strict;
+use warnings;
+use IPC::Open2 qw(open2);
+
+sub new {
+        my ($class, $git_dir) = @_;
+        bless { git_dir => $git_dir }, $class;
+}
+
+sub _cat_file_begin {
+        my ($self) = @_;
+        return if $self->{pid};
+        my ($in, $out);
+        my $pid = open2($in, $out, 'git', '--git-dir', $self->{git_dir},
+                        'cat-file', '--batch');
+
+        $self->{pid} = $pid;
+        $self->{in} = $in;
+        $self->{out} = $out;
+}
+
+sub cat_file {
+        my ($self, $object) = @_;
+
+        $self->_cat_file_begin;
+        print { $self->{out} } $object, "\n" or die "write error: $!\n";
+
+        my $in = $self->{in};
+        my $head = <$in>;
+        $head =~ / missing$/ and return undef;
+        $head =~ /^[0-9a-f]{40} \S+ (\d+)$/ or
+                die "Unexpected result from git cat-file: $head\n";
+
+        my $size = $1;
+        my $bytes_left = $size;
+        my $buf;
+        my $rv = '';
+
+        while ($bytes_left) {
+                my $read = read($in, $buf, $bytes_left);
+                defined($read) or die "read pipe failed: $!\n";
+                $rv .= $buf;
+                $bytes_left -= $read;
+        }
+
+        my $read = read($in, $buf, 1);
+        defined($read) or die "read pipe failed: $!\n";
+        if ($read != 1 || $buf ne "\n") {
+                die "newline missing after blob\n";
+        }
+        \$rv;
+}
+
+sub DESTROY {
+        my ($self) = @_;
+        my $pid = $self->{pid} or return;
+        $self->{pid} = undef;
+        foreach my $f (qw(in out)) {
+                my $fh = $self->{$f};
+                defined $fh or next;
+                close $fh;
+                $self->{$f} = undef;
+        }
+        waitpid $pid, 0;
+}
+
+1;
diff --git a/public-inbox.cgi b/public-inbox.cgi
index 93147124..ffd6ec08 100755
--- a/public-inbox.cgi
+++ b/public-inbox.cgi
@@ -218,7 +218,6 @@ sub psgi_app {
         require POSIX;
         require XML::Atom::SimpleFeed;
         require Plack::Request;
-        eval { require Git }; # optional
         sub {
                 my $req = Plack::Request->new(@_);
                 main($req, $req->method);