Documentation/design_www.txt | 2 ++
lib/PublicInbox/GitHTTPBackend.pm | 214 +++++++++++++++++++++++++++++++++++++++++++++++++++++
lib/PublicInbox/GitHTTPDumb.pm | 121 -----------------------------------------------------
lib/PublicInbox/WWW.pm | 17 ++++++++++++-----
diff --git a/Documentation/design_www.txt b/Documentation/design_www.txt
index 760152333e651b916db7d62ae5d70c486da39cf8..39b124143fd313baa17b6ea80bd94d7e85fb7854 100644
--- a/Documentation/design_www.txt
+++ b/Documentation/design_www.txt
@@ -25,6 +25,8 @@ /$LISTNAME/f/$MESSAGE_ID.txt [1] -> 301 to /$LISTNAME/$MESSAGE_ID/raw
/$LISTNAME/atom.xml [2] -> identical to /$LISTNAME/new.atom
+Additionally, we support "git clone" pointed to http://$HOST/$LISTNAME
+
FIXME: we must refactor/cleanup/add tests for most of our CGI before
adding more endpoints and features.
diff --git a/lib/PublicInbox/GitHTTPBackend.pm b/lib/PublicInbox/GitHTTPBackend.pm
new file mode 100644
index 0000000000000000000000000000000000000000..71b7a8f1b1c12278730dd8c8cd64cf6f3a93d415
--- /dev/null
+++ b/lib/PublicInbox/GitHTTPBackend.pm
@@ -0,0 +1,214 @@
+# Copyright (C) 2016 all contributors
+# License: AGPL-3.0+
+
+# when no endpoints match, fallback to this and serve a static file
+# or smart HTTP
+package PublicInbox::GitHTTPBackend;
+use strict;
+use warnings;
+use Fcntl qw(:seek);
+use POSIX qw(dup2);
+
+# n.b. serving "description" and "cloneurl" should be innocuous enough to
+# not cause problems. serving "config" might...
+my @text = qw[HEAD info/refs
+ objects/info/(?:http-alternates|alternates|packs)
+ cloneurl description];
+
+my @binary = qw!
+ objects/[a-f0-9]{2}/[a-f0-9]{38}
+ objects/pack/pack-[a-f0-9]{40}\.(?:pack|idx)
+ !;
+
+our $ANY = join('|', @binary, @text);
+my $BIN = join('|', @binary);
+my $TEXT = join('|', @text);
+
+sub r {
+ [ $_[0] , [qw(Content-Type text/plain Content-Length 0) ], [] ]
+}
+
+sub serve {
+ my ($cgi, $git, $path) = @_;
+ my $service = $cgi->param('service') || '';
+ if ($service =~ /\Agit-\w+-pack\z/ || $path =~ /\Agit-\w+-pack\z/) {
+ my $ok = serve_smart($cgi, $git, $path);
+ return $ok if $ok;
+ }
+
+ my $type;
+ if ($path =~ /\A(?:$BIN)\z/o) {
+ $type = 'application/octet-stream';
+ } elsif ($path =~ /\A(?:$TEXT)\z/o) {
+ $type = 'text/plain';
+ } else {
+ return r(404);
+ }
+ my $f = "$git->{git_dir}/$path";
+ return r(404) unless -f $f && -r _;
+ my @st = stat(_);
+ my $size = $st[7];
+
+ # TODO: If-Modified-Since and Last-Modified
+ open my $in, '<', $f or return r(404);
+ my $code = 200;
+ my $len = $size;
+ my @h;
+
+ my $env = $cgi->{env} || \%ENV;
+ my $range = $env->{HTTP_RANGE};
+ if (defined $range && $range =~ /\bbytes=(\d*)-(\d*)\z/) {
+ ($code, $len) = prepare_range($cgi, $in, \@h, $1, $2, $size);
+ if ($code == 416) {
+ push @h, 'Content-Range', "bytes */$size";
+ return [ 416, \@h, [] ];
+ }
+ }
+
+ push @h, 'Content-Type', $type, 'Content-Length', $len;
+ sub {
+ my ($res) = @_; # Plack callback
+ my $fh = $res->([ $code, \@h ]);
+ my $buf;
+ my $n = 8192;
+ while ($len > 0) {
+ $n = $len if $len < $n;
+ my $r = read($in, $buf, $n);
+ last if (!defined($r) || $r <= 0);
+ $len -= $r;
+ $fh->write($buf);
+ }
+ $fh->close;
+ }
+}
+
+sub prepare_range {
+ my ($cgi, $in, $h, $beg, $end, $size) = @_;
+ my $code = 200;
+ my $len = $size;
+ if ($beg eq '') {
+ if ($end ne '') { # "bytes=-$end" => last N bytes
+ $beg = $size - $end;
+ $beg = 0 if $beg < 0;
+ $end = $size - 1;
+ $code = 206;
+ } else {
+ $code = 416;
+ }
+ } else {
+ if ($beg > $size) {
+ $code = 416;
+ } elsif ($end eq '' || $end >= $size) {
+ $end = $size - 1;
+ $code = 206;
+ } elsif ($end < $size) {
+ $code = 206;
+ } else {
+ $code = 416;
+ }
+ }
+ if ($code == 206) {
+ $len = $end - $beg + 1;
+ if ($len <= 0) {
+ $code = 416;
+ } else {
+ seek($in, $beg, SEEK_SET) or return [ 500, [], [] ];
+ push @$h, qw(Accept-Ranges bytes Content-Range);
+ push @$h, "bytes $beg-$end/$size";
+
+ # FIXME: Plack::Middleware::Deflater bug?
+ if (my $env = $cgi->{env}) {
+ $env->{'psgix.no-compress'} = 1;
+ }
+ }
+ }
+ ($code, $len);
+}
+
+# returns undef if 403 so it falls back to dumb HTTP
+sub serve_smart {
+ my ($cgi, $git, $path) = @_;
+ my $env = $cgi->{env};
+
+ my $input = $env->{'psgi.input'};
+ my $buf;
+ my $in;
+ my $err = $env->{'psgi.errors'};
+ if (fileno($input) >= 0) { # FIXME untested
+ $in = $input;
+ } else {
+ $in = IO::File->new_tmpfile;
+ while (1) {
+ my $r = $input->read($buf, 8192);
+ unless (defined $r) {
+ $err->print('error reading input: ', $!, "\n");
+ return r(500);
+ }
+ last if ($r == 0);
+ $in->write($buf);
+ }
+ $in->flush;
+ $in->sysseek(0, SEEK_SET);
+ }
+ my $out = IO::File->new_tmpfile;
+ my $pid = fork; # TODO: vfork under Linux...
+ unless (defined $pid) {
+ $err->print('error forking: ', $!, "\n");
+ return r(500);
+ }
+ if ($pid == 0) {
+ # GIT_HTTP_EXPORT_ALL, GIT_COMMITTER_NAME, GIT_COMMITTER_EMAIL
+ # may be set in the server-process and are passed as-is
+ foreach my $name (qw(QUERY_STRING
+ REMOTE_USER REMOTE_ADDR
+ HTTP_CONTENT_ENCODING
+ CONTENT_TYPE
+ SERVER_PROTOCOL
+ REQUEST_METHOD)) {
+ my $val = $env->{$name};
+ $ENV{$name} = $val if defined $val;
+ }
+ # $ENV{GIT_PROJECT_ROOT} = $git->{git_dir};
+ $ENV{GIT_HTTP_EXPORT_ALL} = '1';
+ $ENV{PATH_TRANSLATED} = "$git->{git_dir}/$path";
+ dup2(fileno($in), 0) or die "redirect stdin failed: $!\n";
+ dup2(fileno($out), 1) or die "redirect stdout failed: $!\n";
+ my @cmd = qw(git http-backend);
+ exec(@cmd) or die 'exec `' . join(' ', @cmd). "' failed: $!\n";
+ }
+
+ if (waitpid($pid, 0) != $pid) {
+ $err->print("git http-backend ($git->{git_dir}): ", $?, "\n");
+ return r(500);
+ }
+ $in = undef;
+ $out->seek(0, SEEK_SET);
+ my @h;
+ my $code = 200;
+ {
+ local $/ = "\r\n";
+ while (defined(my $line = <$out>)) {
+ if ($line =~ /\AStatus:\s*(\d+)/) {
+ $code = $1;
+ } else {
+ chomp $line;
+ last if $line eq '';
+ push @h, split(/:\s*/, $line, 2);
+ }
+ }
+ }
+ return if $code == 403;
+ sub {
+ my ($cb) = @_;
+ my $fh = $cb->([ $code, \@h ]);
+ while (1) {
+ my $r = $out->read($buf, 8192);
+ die "$!\n" unless defined $r;
+ last if ($r == 0);
+ $fh->write($buf);
+ }
+ $fh->close;
+ }
+}
+
+1;
diff --git a/lib/PublicInbox/GitHTTPDumb.pm b/lib/PublicInbox/GitHTTPDumb.pm
deleted file mode 100644
index c088d8c4980ebf11cd0fe4188e79a1afcae60409..0000000000000000000000000000000000000000
--- a/lib/PublicInbox/GitHTTPDumb.pm
+++ /dev/null
@@ -1,121 +0,0 @@
-# Copyright (C) 2016 all contributors
-# License: AGPL-3.0+
-
-# when no endpoints match, fallback to this and serve a static file
-# This can serve Smart HTTP in the future.
-package PublicInbox::GitHTTPDumb;
-use strict;
-use warnings;
-use Fcntl qw(:seek);
-
-# n.b. serving "description" and "cloneurl" should be innocuous enough to
-# not cause problems. serving "config" might...
-my @text = qw[HEAD info/refs
- objects/info/(?:http-alternates|alternates|packs)
- cloneurl description];
-
-my @binary = qw!
- objects/[a-f0-9]{2}/[a-f0-9]{38}
- objects/pack/pack-[a-f0-9]{40}\.(?:pack|idx)
- !;
-
-our $ANY = join('|', @binary, @text);
-my $BIN = join('|', @binary);
-my $TEXT = join('|', @text);
-
-sub r {
- [ $_[0] , [qw(Content-Type text/plain Content-Length 0) ], [] ]
-}
-
-sub serve {
- my ($cgi, $git, $path) = @_;
- my $type;
- if ($path =~ /\A(?:$BIN)\z/o) {
- $type = 'application/octet-stream';
- } elsif ($path =~ /\A(?:$TEXT)\z/o) {
- $type = 'text/plain';
- } else {
- return r(404);
- }
- my $f = "$git->{git_dir}/$path";
- return r(404) unless -f $f && -r _;
- my @st = stat(_);
- my $size = $st[7];
-
- # TODO: If-Modified-Since and Last-Modified
- open my $in, '<', $f or return r(404);
- my $code = 200;
- my $len = $size;
- my @h;
-
- my $env = $cgi->{env} || \%ENV;
- my $range = $env->{HTTP_RANGE};
- if (defined $range && $range =~ /\bbytes=(\d*)-(\d*)\z/) {
- ($code, $len) = prepare_range($cgi, $in, \@h, $1, $2, $size);
- if ($code == 416) {
- push @h, 'Content-Range', "bytes */$size";
- return [ 416, \@h, [] ];
- }
- }
-
- push @h, 'Content-Type', $type, 'Content-Length', $len;
- sub {
- my ($res) = @_; # Plack callback
- my $fh = $res->([ $code, \@h ]);
- my $buf;
- my $n = 8192;
- while ($len > 0) {
- $n = $len if $len < $n;
- my $r = read($in, $buf, $n);
- last if (!defined($r) || $r <= 0);
- $len -= $r;
- $fh->write($buf);
- }
- $fh->close;
- }
-}
-
-sub prepare_range {
- my ($cgi, $in, $h, $beg, $end, $size) = @_;
- my $code = 200;
- my $len = $size;
- if ($beg eq '') {
- if ($end ne '') { # "bytes=-$end" => last N bytes
- $beg = $size - $end;
- $beg = 0 if $beg < 0;
- $end = $size - 1;
- $code = 206;
- } else {
- $code = 416;
- }
- } else {
- if ($beg > $size) {
- $code = 416;
- } elsif ($end eq '' || $end >= $size) {
- $end = $size - 1;
- $code = 206;
- } elsif ($end < $size) {
- $code = 206;
- } else {
- $code = 416;
- }
- }
- if ($code == 206) {
- $len = $end - $beg + 1;
- if ($len <= 0) {
- $code = 416;
- } else {
- seek($in, $beg, SEEK_SET) or return [ 500, [], [] ];
- push @$h, qw(Accept-Ranges bytes Content-Range);
- push @$h, "bytes $beg-$end/$size";
-
- # FIXME: Plack::Middleware::Deflater bug?
- if (my $env = $cgi->{env}) {
- $env->{'psgix.no-compress'} = 1;
- }
- }
- }
- ($code, $len);
-}
-
-1;
diff --git a/lib/PublicInbox/WWW.pm b/lib/PublicInbox/WWW.pm
index 1c6936f76bd2dcf6f3a4c4e64e8f4c8f90595523..b4b012f962fb81ce0f029f59231fa162b0b5db57 100644
--- a/lib/PublicInbox/WWW.pm
+++ b/lib/PublicInbox/WWW.pm
@@ -18,7 +18,7 @@ use URI::Escape qw(uri_escape_utf8 uri_unescape);
use constant SSOMA_URL => 'http://ssoma.public-inbox.org/';
use constant PI_URL => 'http://public-inbox.org/';
require PublicInbox::Git;
-use PublicInbox::GitHTTPDumb;
+use PublicInbox::GitHTTPBackend;
our $LISTNAME_RE = qr!\A/([\w\.\-]+)!;
our $MID_RE = qr!([^/]+)!;
our $END_RE = qr!(f/|T/|t/|t\.mbox(?:\.gz)?|t\.atom|raw|)!;
@@ -28,10 +28,17 @@ sub run {
my ($cgi, $method) = @_;
$pi_config ||= PublicInbox::Config->new;
my $ctx = { cgi => $cgi, pi_config => $pi_config };
- if ($method !~ /\AGET|HEAD\z/) {
+ my $path_info = $cgi->path_info;
+
+ if ($method eq 'POST' &&
+ $path_info =~ m!$LISTNAME_RE/(git-upload-pack)\z!) {
+ my $path = $2;
+ return (invalid_list($ctx, $1) ||
+ serve_git($cgi, $ctx->{git}, $path));
+ }
+ elsif ($method !~ /\AGET|HEAD\z/) {
return r(405, 'Method Not Allowed');
}
- my $path_info = $cgi->path_info;
# top-level indices and feeds
if ($path_info eq '/') {
@@ -44,7 +51,7 @@ } elsif ($path_info =~ m!$LISTNAME_RE/(?:atom\.xml|new\.atom)\z!o) {
invalid_list($ctx, $1) || get_atom($ctx);
} elsif ($path_info =~ m!$LISTNAME_RE/
- ($PublicInbox::GitHTTPDumb::ANY)\z!ox) {
+ ($PublicInbox::GitHTTPBackend::ANY)\z!ox) {
my $path = $2;
invalid_list($ctx, $1) || serve_git($cgi, $ctx->{git}, $path);
} elsif ($path_info =~ m!$LISTNAME_RE/$MID_RE/$END_RE\z!o) {
@@ -402,7 +409,7 @@ }
sub serve_git {
my ($cgi, $git, $path) = @_;
- PublicInbox::GitHTTPDumb::serve($cgi, $git, $path);
+ PublicInbox::GitHTTPBackend::serve($cgi, $git, $path);
}
1;