[PATCH 0/3] coderepo-related cleanups + sha256

[PATCH 0/3] coderepo-related cleanups + sha256

[Eric Wong]

libgit2 on my system can't handle sha256 git repos, and I need
it for one of my projects.  So stick to git(1) for now and
simplify our code.

Eric Wong (3):
  qspawn: import Scalar::Util::blessed properly
  viewvcs: use git(1) for coderepo access
  coderepo: consolidate git --batch-check users

 lib/PublicInbox/Qspawn.pm       |  6 ++--
 lib/PublicInbox/RepoSnapshot.pm | 51 +++++++++++++--------------------
 lib/PublicInbox/RepoTree.pm     |  8 +-----
 lib/PublicInbox/ViewVCS.pm      | 46 +++++++++++++++--------------
 lib/PublicInbox/WwwCoderepo.pm  |  9 +-----
 5 files changed, 50 insertions(+), 70 deletions(-)

[PATCH 1/3] qspawn: import Scalar::Util::blessed properly

[Eric Wong]

Scalar::Util may not be loaded by other modules in the future.
---
 lib/PublicInbox/Qspawn.pm | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/PublicInbox/Qspawn.pm b/lib/PublicInbox/Qspawn.pm
index c2d8fd73..0044c1f6 100644
--- a/lib/PublicInbox/Qspawn.pm
+++ b/lib/PublicInbox/Qspawn.pm
@@ -28,6 +28,7 @@ package PublicInbox::Qspawn;
 use v5.12;
 use PublicInbox::Spawn qw(popen_rd);
 use PublicInbox::GzipFilter;
+use Scalar::Util qw(blessed);
 
 # n.b.: we get EAGAIN with public-inbox-httpd, and EINTR on other PSGI servers
 use Errno qw(EAGAIN EINTR);
@@ -234,8 +235,9 @@ sub psgi_return_init_cb { # this may be PublicInbox::HTTPD::Async {cb}
 	my $r = rd_hdr($self) or return;
 	my $env = $self->{psgi_env};
 	my $filter;
-	if (ref($r) eq 'ARRAY' && Scalar::Util::blessed($r->[2]) &&
-			$r->[2]->can('attach')) {
+
+	# this is for RepoAtom since that can fire after parse_cgi_headers
+	if (ref($r) eq 'ARRAY' && blessed($r->[2]) && $r->[2]->can('attach')) {
 		$filter = pop @$r;
 	}
 	$filter //= delete($env->{'qspawn.filter'}) // (ref($r) eq 'ARRAY' ?

[PATCH 2/3] viewvcs: use git(1) for coderepo access

[Eric Wong]

libgit2 development has fallen behind git.git and I've been
using objectformat=sha256 somewhere else for over 18 months.

Hoist out do_cat_async() into it's own sub to hide generic PSGI
vs -httpd differences while we're at it to save us some code.
---
 lib/PublicInbox/ViewVCS.pm     | 36 +++++++++++++---------------------
 lib/PublicInbox/WwwCoderepo.pm |  9 +--------
 2 files changed, 15 insertions(+), 30 deletions(-)

diff --git a/lib/PublicInbox/ViewVCS.pm b/lib/PublicInbox/ViewVCS.pm
index e0fdf639..eae5b7f4 100644
--- a/lib/PublicInbox/ViewVCS.pm
+++ b/lib/PublicInbox/ViewVCS.pm
@@ -125,6 +125,17 @@ sub cmt_title { # git->cat_async callback
 		cmt_finalize($ctx);
 }
 
+sub do_cat_async {
+	my ($ctx, $cb, @oids) = @_;
+	# favor git(1) over Gcf2 (libgit2) for SHA-256 support
+	$ctx->{git}->cat_async($_, $cb, $ctx) for @oids;
+	if ($ctx->{env}->{'pi-httpd.async'}) {
+		PublicInbox::GitAsyncCat::watch_cat($ctx->{git});
+	} else { # synchronous, generic PSGI
+		$ctx->{git}->cat_async_wait;
+	}
+}
+
 sub show_commit_start { # ->psgi_qx callback
 	my ($bref, $ctx) = @_;
 	if (my $qsp_err = delete $ctx->{-qsp_err}) {
@@ -142,16 +153,7 @@ sub show_commit_start { # ->psgi_qx callback
 	return cmt_finalize($ctx) if !$P;
 	@{$ctx->{-cmt_P}} = split(/ /, $P);
 	@{$ctx->{-cmt_p}} = split(/ /, $p); # abbreviated
-	if ($ctx->{env}->{'pi-httpd.async'}) {
-		for (@{$ctx->{-cmt_P}}) {
-			ibx_async_cat($ctx, $_, \&cmt_title, $ctx);
-		}
-	} else { # synchronous
-		for (@{$ctx->{-cmt_P}}) {
-			$ctx->{git}->cat_async($_, \&cmt_title, $ctx);
-		}
-		$ctx->{git}->cat_async_wait;
-	}
+	do_cat_async($ctx, \&cmt_title, @{$ctx->{-cmt_P}});
 }
 
 sub ibx_url_for {
@@ -473,12 +475,7 @@ sub show_tag ($$) {
 	my ($ctx, $res) = @_;
 	my ($git, $oid) = @$res;
 	$ctx->{git} = $git;
-	if ($ctx->{env}->{'pi-httpd.async'}) {
-		ibx_async_cat($ctx, $oid, \&show_tag_result, $ctx);
-	} else { # synchronous (generic PSGI)
-		$git->cat_async($oid, \&show_tag_result, $ctx);
-		$git->cat_async_wait;
-	}
+	do_cat_async($ctx, \&show_tag_result, $oid);
 }
 
 # user_cb for SolverGit, called as: user_cb->($result_or_error, $uarg)
@@ -508,12 +505,7 @@ EOM
 	}
 	bless $ctx, 'PublicInbox::WwwStream'; # for DESTROY
 	$ctx->{git} = $git;
-	if ($ctx->{env}->{'pi-httpd.async'}) {
-		ibx_async_cat($ctx, $oid, \&show_blob, $ctx);
-	} else { # synchronous
-		$git->cat_async($oid, \&show_blob, $ctx);
-		$git->cat_async_wait;
-	}
+	do_cat_async($ctx, \&show_blob, $oid);
 }
 
 sub show_blob { # git->cat_async callback
diff --git a/lib/PublicInbox/WwwCoderepo.pm b/lib/PublicInbox/WwwCoderepo.pm
index f9c150c0..5ca8ef55 100644
--- a/lib/PublicInbox/WwwCoderepo.pm
+++ b/lib/PublicInbox/WwwCoderepo.pm
@@ -11,8 +11,6 @@ use File::Temp 0.19 (); # newdir
 use PublicInbox::ViewVCS;
 use PublicInbox::WwwStatic qw(r);
 use PublicInbox::GitHTTPBackend;
-use PublicInbox::Git;
-use PublicInbox::GitAsyncCat;
 use PublicInbox::WwwStream;
 use PublicInbox::Hval qw(ascii_html);
 use PublicInbox::ViewDiff qw(uri_escape_path);
@@ -199,12 +197,7 @@ sub summary {
 	$tip //= 'HEAD';
 	my @try = ("$tip:README", "$tip:README.md"); # TODO: configurable
 	$ctx->{-nr_readme_tries} = [ @try ];
-	$ctx->{git}->cat_async($_, \&set_readme, $ctx) for @try;
-	if ($ctx->{env}->{'pi-httpd.async'}) {
-		PublicInbox::GitAsyncCat::watch_cat($ctx->{git});
-	} else { # synchronous
-		$ctx->{git}->cat_async_wait;
-	}
+	PublicInbox::ViewVCS::do_cat_async($ctx, \&set_readme, @try);
 	sub { # $_[0] => PublicInbox::HTTP::{Identity,Chunked}
 		$ctx->{env}->{'qspawn.wcb'} = $_[0];
 		$qsp->psgi_qx($ctx->{env}, undef, \&capture_refs, $ctx);

[PATCH 3/3] coderepo: consolidate git --batch-check users

[Eric Wong]

And another opportunity to simplify our code between different
PSGI-ish implementations.  The snapshot retrieval is simpler,
but potentially slower since we waste cycles scanning for tags
even after we've found one.  It's probably not a big deal since
it's only short info lines and we can utilize pipelining.
---
 lib/PublicInbox/RepoSnapshot.pm | 51 +++++++++++++--------------------
 lib/PublicInbox/RepoTree.pm     |  8 +-----
 lib/PublicInbox/ViewVCS.pm      | 14 +++++++--
 3 files changed, 33 insertions(+), 40 deletions(-)

diff --git a/lib/PublicInbox/RepoSnapshot.pm b/lib/PublicInbox/RepoSnapshot.pm
index 826392a8..93ba4db6 100644
--- a/lib/PublicInbox/RepoSnapshot.pm
+++ b/lib/PublicInbox/RepoSnapshot.pm
@@ -4,9 +4,8 @@
 # cgit-compatible /snapshot/ endpoint for WWW coderepos
 package PublicInbox::RepoSnapshot;
 use v5.12;
-use PublicInbox::Git;
 use PublicInbox::Qspawn;
-use PublicInbox::GitAsyncCat;
+use PublicInbox::ViewVCS;
 use PublicInbox::WwwStatic qw(r);
 
 # Not using standard mime types since the compressed tarballs are
@@ -42,31 +41,25 @@ sub archive_hdr { # parse_hdr for Qspawn
 		'ETag', qq("$ctx->{etag}") ] ];
 }
 
-sub archive_cb {
-	my ($ctx) = @_;
-	my @cfg;
-	if (my $cmd = $FMT_CFG{$ctx->{snap_fmt}}) {
-		@cfg = ('-c', "tar.$ctx->{snap_fmt}.command=$cmd");
-	}
-	my $qsp = PublicInbox::Qspawn->new(['git', @cfg,
-			"--git-dir=$ctx->{git}->{git_dir}", 'archive',
-			"--prefix=$ctx->{snap_pfx}/",
-			"--format=$ctx->{snap_fmt}", $ctx->{treeish}]);
-	$qsp->psgi_return($ctx->{env}, undef, \&archive_hdr, $ctx);
-}
-
 sub ver_check { # git->check_async callback
 	my ($oid, $type, $size, $ctx) = @_;
-	if ($type eq 'missing') { # try 'v' and 'V' prefixes
-		my $pfx = shift @{$ctx->{try_pfx}} or return
+	return if defined $ctx->{etag};
+	my $treeish = shift @{$ctx->{-try}} // die 'BUG: no {-try}';
+	if ($type eq 'missing') {
+		scalar(@{$ctx->{-try}}) or
 			delete($ctx->{env}->{'qspawn.wcb'})->(r(404));
-		my $v = $ctx->{treeish} = $pfx.$ctx->{snap_ver};
-		return $ctx->{env}->{'pi-httpd.async'} ?
-			async_check($ctx, $v, \&ver_check, $ctx) :
-			$ctx->{git}->check_async($v, \&ver_check, $ctx);
+	} else { # found, done:
+		$ctx->{etag} = $oid;
+		my @cfg;
+		if (my $cmd = $FMT_CFG{$ctx->{snap_fmt}}) {
+			@cfg = ('-c', "tar.$ctx->{snap_fmt}.command=$cmd");
+		}
+		my $qsp = PublicInbox::Qspawn->new(['git', @cfg,
+				"--git-dir=$ctx->{git}->{git_dir}", 'archive',
+				"--prefix=$ctx->{snap_pfx}/",
+				"--format=$ctx->{snap_fmt}", $treeish]);
+		$qsp->psgi_return($ctx->{env}, undef, \&archive_hdr, $ctx);
 	}
-	$ctx->{etag} = $oid;
-	archive_cb($ctx);
 }
 
 sub srv {
@@ -81,16 +74,12 @@ sub srv {
 	substr($fn, 0, length($pfx)) eq $pfx or return;
 	$ctx->{snap_pfx} = $fn;
 	my $v = $ctx->{snap_ver} = substr($fn, length($pfx), length($fn));
-	$ctx->{treeish} = $v; # try without [vV] prefix, first
-	@{$ctx->{try_pfx}} = qw(v V); # cf. cgit:ui-snapshot.c
+	# try without [vV] prefix, first
+	my @try = map { "$_$v" } ('', 'v', 'V'); # cf. cgit:ui-snapshot.c
+	@{$ctx->{-try}} = @try;
 	sub {
 		$ctx->{env}->{'qspawn.wcb'} = $_[0];
-		if ($ctx->{env}->{'pi-httpd.async'}) {
-			async_check($ctx, $v, \&ver_check, $ctx);
-		} else {
-			$ctx->{git}->check_async($v, \&ver_check, $ctx);
-			$ctx->{git}->check_async_wait;
-		}
+		PublicInbox::ViewVCS::do_check_async($ctx, \&ver_check, @try);
 	}
 }
 
diff --git a/lib/PublicInbox/RepoTree.pm b/lib/PublicInbox/RepoTree.pm
index 84e20589..7434e9b2 100644
--- a/lib/PublicInbox/RepoTree.pm
+++ b/lib/PublicInbox/RepoTree.pm
@@ -5,7 +5,6 @@
 package PublicInbox::RepoTree;
 use v5.12;
 use PublicInbox::ViewDiff qw(uri_escape_path);
-use PublicInbox::GitAsyncCat;
 use PublicInbox::WwwStatic qw(r);
 use PublicInbox::Qspawn;
 use PublicInbox::WwwStream qw(html_oneshot);
@@ -78,12 +77,7 @@ sub srv_tree {
 	return if index($obj, "\n") >= 0;
 	sub {
 		$ctx->{-wcb} = $_[0]; # HTTP::{Chunked,Identity}
-		if ($ctx->{env}->{'pi-httpd.async'}) {
-			async_check($ctx, $obj, \&tree_show, $ctx);
-		} else {
-			$ctx->{git}->check_async($obj, \&tree_show, $ctx);
-			$ctx->{git}->async_wait_all;
-		}
+		PublicInbox::ViewVCS::do_check_async($ctx, \&tree_show, $obj);
 	};
 }
 
diff --git a/lib/PublicInbox/ViewVCS.pm b/lib/PublicInbox/ViewVCS.pm
index eae5b7f4..37b688ed 100644
--- a/lib/PublicInbox/ViewVCS.pm
+++ b/lib/PublicInbox/ViewVCS.pm
@@ -126,9 +126,9 @@ sub cmt_title { # git->cat_async callback
 }
 
 sub do_cat_async {
-	my ($ctx, $cb, @oids) = @_;
+	my ($ctx, $cb, @req) = @_;
 	# favor git(1) over Gcf2 (libgit2) for SHA-256 support
-	$ctx->{git}->cat_async($_, $cb, $ctx) for @oids;
+	$ctx->{git}->cat_async($_, $cb, $ctx) for @req;
 	if ($ctx->{env}->{'pi-httpd.async'}) {
 		PublicInbox::GitAsyncCat::watch_cat($ctx->{git});
 	} else { # synchronous, generic PSGI
@@ -136,6 +136,16 @@ sub do_cat_async {
 	}
 }
 
+sub do_check_async {
+	my ($ctx, $cb, @req) = @_;
+	if ($ctx->{env}->{'pi-httpd.async'}) {
+		async_check($ctx, $_, $cb, $ctx) for @req;
+	} else { # synchronous, generic PSGI
+		$ctx->{git}->check_async($_, $cb, $ctx) for @req;
+		$ctx->{git}->check_async_wait;
+	}
+}
+
 sub show_commit_start { # ->psgi_qx callback
 	my ($bref, $ctx) = @_;
 	if (my $qsp_err = delete $ctx->{-qsp_err}) {