1 # Copyright (C) 2019 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # show any VCS object, similar to "git show"
5 # FIXME: we only show blobs for now
7 # This can use a "solver" to reconstruct blobs based on git
8 # patches (with abbreviated OIDs in the header). However, the
9 # abbreviated OIDs must match exactly what's in the original
10 # email (unless a normal code repo already has the blob).
12 # In other words, we can only reliably reconstruct blobs based
13 # on links generated by ViewDiff (and only if the emailed
14 # patches apply 100% cleanly to published blobs).
16 package PublicInbox::ViewVCS;
19 use bytes (); # only for bytes::length
20 use PublicInbox::SolverGit;
21 use PublicInbox::WwwStream;
22 use PublicInbox::Linkify;
23 use PublicInbox::Hval qw(ascii_html to_filename);
25 require PublicInbox::HlMod;
26 PublicInbox::HlMod->new;
29 my %QP_MAP = ( A => 'oid_a', B => 'oid_b', a => 'path_a', b => 'path_b' );
30 my $max_size = 1024 * 1024; # TODO: configurable
31 my $BIN_DETECT = 8000; # same as git
34 my ($ctx, $code, $strref) = @_;
35 my $wcb = delete $ctx->{-wcb};
36 $ctx->{-upfx} = '../../'; # from "/$INBOX/$OID/s/"
37 my $res = PublicInbox::WwwStream->response($ctx, $code, sub {
39 $nr == 1 ? $$strref : undef;
41 $wcb ? $wcb->($res) : $res;
44 sub stream_large_blob ($$$$) {
45 my ($ctx, $res, $logref, $fn) = @_;
46 my ($git, $oid, $type, $size, $di) = @$res;
47 my $cmd = ['git', "--git-dir=$git->{git_dir}", 'cat-file', $type, $oid];
48 my $qsp = PublicInbox::Qspawn->new($cmd);
49 my @cl = ('Content-Length', $size);
50 my $env = $ctx->{env};
51 $env->{'public-inbox.tmpgit'} = $git; # for {-tmp}/File::Temp::Dir
52 $env->{'qspawn.wcb'} = delete $ctx->{-wcb};
53 $qsp->psgi_return($env, undef, sub {
55 if (!defined $r) { # error
56 html_page($ctx, 500, $logref);
57 } elsif (index($$bref, "\0") >= 0) {
58 my $ct = 'application/octet-stream';
59 [200, ['Content-Type', $ct, @cl ] ];
61 my $n = bytes::length($$bref);
62 if ($n >= $BIN_DETECT || $n == $size) {
63 my $ct = 'text/plain; charset=UTF-8';
64 return [200, ['Content-Type', $ct, @cl] ];
67 warn "premature EOF on $oid $$logref\n";
68 return html_page($ctx, 500, $logref);
70 undef; # bref keeps growing
75 sub show_other ($$$$) {
76 my ($ctx, $res, $logref, $fn) = @_;
77 my ($git, $oid, $type, $size) = @$res;
78 if ($size > $max_size) {
79 $$logref = "$oid is too big to show\n" . $$logref;
80 return html_page($ctx, 200, $logref);
82 my $cmd = ['git', "--git-dir=$git->{git_dir}",
83 qw(show --encoding=UTF-8 --no-color --no-abbrev), $oid ];
84 my $qsp = PublicInbox::Qspawn->new($cmd);
85 my $env = $ctx->{env};
86 $qsp->psgi_qx($env, undef, sub {
88 if (my $err = $qsp->{err}) {
90 $$logref .= "git show error: $err";
91 return html_page($ctx, 500, $logref);
93 my $l = PublicInbox::Linkify->new;
95 $l->linkify_1($$bref);
96 $$bref = '<pre>'. $l->linkify_2(ascii_html($$bref));
97 $$bref .= '</pre><hr>' . $$logref;
98 html_page($ctx, 200, $bref);
103 my ($ctx, $res, $log, $hints, $fn) = @_;
105 unless (seek($log, 0, 0)) {
106 $ctx->{env}->{'psgi.errors'}->print("seek(log): $!\n");
107 return html_page($ctx, 500, \'seek error');
109 $log = do { local $/; <$log> };
112 my $l = PublicInbox::Linkify->new;
114 $log = '<pre>debug log:</pre><hr /><pre>' .
115 $l->linkify_2(ascii_html($log)) . '</pre>';
117 $res or return html_page($ctx, 404, \$log);
118 $ref eq 'ARRAY' or return html_page($ctx, 500, \$log);
120 my ($git, $oid, $type, $size, $di) = @$res;
121 return show_other($ctx, $res, \$log, $fn) if $type ne 'blob';
122 my $path = to_filename($di->{path_b} || $hints->{path_b} || 'blob');
123 my $raw_link = "(<a\nhref=$path>raw</a>)";
124 if ($size > $max_size) {
125 return stream_large_blob($ctx, $res, \$log, $fn) if defined $fn;
126 $log = "<pre><b>Too big to show, download available</b>\n" .
127 "$oid $type $size bytes $raw_link</pre>" . $log;
128 return html_page($ctx, 500, \$log);
131 my $blob = $git->cat_file($oid);
133 my $e = "Failed to retrieve generated blob ($oid)";
134 $ctx->{env}->{'psgi.errors'}->print("$e ($git->{git_dir})\n");
135 $log = "<pre><b>$e</b></pre>" . $log;
136 return html_page($ctx, 500, \$log);
139 my $bin = index(substr($$blob, 0, $BIN_DETECT), "\0") >= 0;
141 my $h = [ 'Content-Length', $size, 'Content-Type' ];
142 push(@$h, ($bin ? 'application/octet-stream' : 'text/plain'));
143 return delete($ctx->{-wcb})->([200, $h, [ $$blob ]]);
147 $log = "<pre>$oid $type $size bytes (binary)" .
148 " $raw_link</pre>" . $log;
149 return html_page($ctx, 200, \$log);
152 # TODO: detect + convert to ensure validity
153 utf8::decode($$blob);
154 my $nl = ($$blob =~ tr/\n/\n/);
155 my $pad = length($nl);
157 $l->linkify_1($$blob);
158 my $ok = $hl->do_hl($blob, $path) if $hl;
162 $$blob = ascii_html($$blob);
165 # using some of the same CSS class names and ids as cgit
166 $log = "<pre>$oid $type $size bytes $raw_link</pre>" .
167 "<hr /><table\nclass=blob>".
168 "<tr><td\nclass=linenumbers><pre>" . join('', map {
169 sprintf("<a id=n$_ href=#n$_>% ${pad}u</a>\n", $_)
170 } (1..$nl)) . '</pre></td>' .
171 '<td><pre> </pre></td>'. # pad for non-CSS users
172 "<td\nclass=lines><pre\nstyle='white-space:pre'><code>" .
173 $l->linkify_2($$blob) .
174 '</code></pre></td></tr></table>' . $log;
176 html_page($ctx, 200, \$log);
180 my ($ctx, $oid_b, $fn) = @_;
183 while (my ($from, $to) = each %QP_MAP) {
184 defined(my $v = $qp->{$from}) or next;
188 open my $log, '+>', undef or die "open: $!";
189 my $solver = PublicInbox::SolverGit->new($ctx->{-inbox}, sub {
190 solve_result($ctx, $_[0], $log, $hints, $fn);
193 # PSGI server will call this and give us a callback
195 $ctx->{-wcb} = $_[0]; # HTTP write callback
196 $solver->solve($ctx->{env}, $log, $oid_b, $hints);