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::Tmpfile;
24 use PublicInbox::Hval qw(ascii_html to_filename);
26 require PublicInbox::HlMod;
27 PublicInbox::HlMod->new;
30 my %QP_MAP = ( A => 'oid_a', B => 'oid_b', a => 'path_a', b => 'path_b' );
31 my $max_size = 1024 * 1024; # TODO: configurable
32 my $BIN_DETECT = 8000; # same as git
35 my ($ctx, $code, $strref) = @_;
36 my $wcb = delete $ctx->{-wcb};
37 $ctx->{-upfx} = '../../'; # from "/$INBOX/$OID/s/"
38 my $res = PublicInbox::WwwStream->response($ctx, $code, sub {
40 $nr == 1 ? $$strref : undef;
42 $wcb ? $wcb->($res) : $res;
45 sub stream_large_blob ($$$$) {
46 my ($ctx, $res, $logref, $fn) = @_;
47 my ($git, $oid, $type, $size, $di) = @$res;
48 my $cmd = ['git', "--git-dir=$git->{git_dir}", 'cat-file', $type, $oid];
49 my $qsp = PublicInbox::Qspawn->new($cmd);
50 my @cl = ('Content-Length', $size);
51 my $env = $ctx->{env};
52 $env->{'public-inbox.tmpgit'} = $git; # for {-tmp}/File::Temp::Dir
53 $env->{'qspawn.wcb'} = delete $ctx->{-wcb};
54 $qsp->psgi_return($env, undef, sub {
56 if (!defined $r) { # error
57 html_page($ctx, 500, $logref);
58 } elsif (index($$bref, "\0") >= 0) {
59 my $ct = 'application/octet-stream';
60 [200, ['Content-Type', $ct, @cl ] ];
62 my $n = bytes::length($$bref);
63 if ($n >= $BIN_DETECT || $n == $size) {
64 my $ct = 'text/plain; charset=UTF-8';
65 return [200, ['Content-Type', $ct, @cl] ];
68 warn "premature EOF on $oid $$logref\n";
69 return html_page($ctx, 500, $logref);
71 undef; # bref keeps growing
76 sub show_other ($$$$) {
77 my ($ctx, $res, $logref, $fn) = @_;
78 my ($git, $oid, $type, $size) = @$res;
79 if ($size > $max_size) {
80 $$logref = "$oid is too big to show\n" . $$logref;
81 return html_page($ctx, 200, $logref);
83 my $cmd = ['git', "--git-dir=$git->{git_dir}",
84 qw(show --encoding=UTF-8 --no-color --no-abbrev), $oid ];
85 my $qsp = PublicInbox::Qspawn->new($cmd);
86 my $env = $ctx->{env};
87 $qsp->psgi_qx($env, undef, sub {
89 if (my $err = $qsp->{err}) {
91 $$logref .= "git show error: $err";
92 return html_page($ctx, 500, $logref);
94 my $l = PublicInbox::Linkify->new;
96 $l->linkify_1($$bref);
97 $$bref = '<pre>'. $l->linkify_2(ascii_html($$bref));
98 $$bref .= '</pre><hr>' . $$logref;
99 html_page($ctx, 200, $bref);
104 my ($ctx, $res, $log, $hints, $fn) = @_;
106 unless (seek($log, 0, 0)) {
107 $ctx->{env}->{'psgi.errors'}->print("seek(log): $!\n");
108 return html_page($ctx, 500, \'seek error');
110 $log = do { local $/; <$log> };
113 my $l = PublicInbox::Linkify->new;
115 $log = '<pre>debug log:</pre><hr /><pre>' .
116 $l->linkify_2(ascii_html($log)) . '</pre>';
118 $res or return html_page($ctx, 404, \$log);
119 $ref eq 'ARRAY' or return html_page($ctx, 500, \$log);
121 my ($git, $oid, $type, $size, $di) = @$res;
122 return show_other($ctx, $res, \$log, $fn) if $type ne 'blob';
123 my $path = to_filename($di->{path_b} || $hints->{path_b} || 'blob');
124 my $raw_link = "(<a\nhref=$path>raw</a>)";
125 if ($size > $max_size) {
126 return stream_large_blob($ctx, $res, \$log, $fn) if defined $fn;
127 $log = "<pre><b>Too big to show, download available</b>\n" .
128 "$oid $type $size bytes $raw_link</pre>" . $log;
129 return html_page($ctx, 500, \$log);
132 my $blob = $git->cat_file($oid);
134 my $e = "Failed to retrieve generated blob ($oid)";
135 $ctx->{env}->{'psgi.errors'}->print("$e ($git->{git_dir})\n");
136 $log = "<pre><b>$e</b></pre>" . $log;
137 return html_page($ctx, 500, \$log);
140 my $bin = index(substr($$blob, 0, $BIN_DETECT), "\0") >= 0;
142 my $h = [ 'Content-Length', $size, 'Content-Type' ];
143 push(@$h, ($bin ? 'application/octet-stream' : 'text/plain'));
144 return delete($ctx->{-wcb})->([200, $h, [ $$blob ]]);
148 $log = "<pre>$oid $type $size bytes (binary)" .
149 " $raw_link</pre>" . $log;
150 return html_page($ctx, 200, \$log);
153 # TODO: detect + convert to ensure validity
154 utf8::decode($$blob);
155 my $nl = ($$blob =~ tr/\n/\n/);
156 my $pad = length($nl);
158 $l->linkify_1($$blob);
159 my $ok = $hl->do_hl($blob, $path) if $hl;
163 $$blob = ascii_html($$blob);
166 # using some of the same CSS class names and ids as cgit
167 $log = "<pre>$oid $type $size bytes $raw_link</pre>" .
168 "<hr /><table\nclass=blob>".
169 "<tr><td\nclass=linenumbers><pre>" . join('', map {
170 sprintf("<a id=n$_ href=#n$_>% ${pad}u</a>\n", $_)
171 } (1..$nl)) . '</pre></td>' .
172 '<td><pre> </pre></td>'. # pad for non-CSS users
173 "<td\nclass=lines><pre\nstyle='white-space:pre'><code>" .
174 $l->linkify_2($$blob) .
175 '</code></pre></td></tr></table>' . $log;
177 html_page($ctx, 200, \$log);
181 my ($ctx, $oid_b, $fn) = @_;
184 while (my ($from, $to) = each %QP_MAP) {
185 defined(my $v = $qp->{$from}) or next;
189 my $log = tmpfile("solve.$oid_b");
190 my $solver = PublicInbox::SolverGit->new($ctx->{-inbox}, sub {
191 solve_result($ctx, $_[0], $log, $hints, $fn);
194 # PSGI server will call this and give us a callback
196 $ctx->{-wcb} = $_[0]; # HTTP write callback
197 $solver->solve($ctx->{env}, $log, $oid_b, $hints);