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 # "Solve" blobs which don't exist in git code repositories by
5 # searching inboxes for post-image blobs.
7 # this emits a lot of debugging/tracing information which may be
8 # publically viewed over HTTP(S). Be careful not to expose
9 # local filesystem layouts in the process.
10 package PublicInbox::SolverGit;
14 use Fcntl qw(SEEK_SET);
15 use PublicInbox::Git qw(git_unquote git_quote);
16 use PublicInbox::Spawn qw(spawn popen_rd);
17 use PublicInbox::MsgIter qw(msg_iter msg_part_text);
18 use URI::Escape qw(uri_escape_utf8);
20 # don't bother if somebody sends us a patch with these path components,
21 # it's junk at best, an attack attempt at worse:
22 my %bad_component = map { $_ => 1 } ('', '.', '..');
25 my ($class, $gits, $inboxes) = @_;
32 # look for existing blobs already in git repos
33 sub solve_existing ($$$) {
34 my ($self, $out, $want) = @_;
35 my $oid_b = $want->{oid_b};
36 my @ambiguous; # Array of [ git, $oids]
37 foreach my $git (@{$self->{gits}}) {
38 my ($oid_full, $type, $size) = $git->check($oid_b);
39 if (defined($type) && $type eq 'blob') {
40 return [ $git, $oid_full, $type, int($size) ];
43 next if length($oid_b) == 40;
45 # parse stderr of "git cat-file --batch-check"
46 my $err = $git->last_check_err;
47 my (@oids) = ($err =~ /\b([a-f0-9]{40})\s+blob\b/g);
48 next unless scalar(@oids);
50 # TODO: do something with the ambiguous array?
51 # push @ambiguous, [ $git, @oids ];
53 print $out "`$oid_b' ambiguous in ",
54 join("\n", $git->pub_urls), "\n",
55 join('', map { "$_ blob\n" } @oids), "\n";
57 scalar(@ambiguous) ? \@ambiguous : undef;
60 # returns a hashref with information about a diff ($di):
62 # oid_a => abbreviated pre-image oid,
63 # oid_b => abbreviated post-image oid,
64 # tmp => anonymous file handle with the diff,
65 # hdr_lines => arrayref of various header lines for mode information
66 # mode_a => original mode of oid_a (string, not integer),
67 # ibx => PublicInbox::Inbox object containing the diff
68 # smsg => PublicInbox::SearchMsg object containing diff
69 # path_a => pre-image path
70 # path_b => post-image path
72 sub extract_diff ($$$$) {
73 my ($p, $re, $ibx, $smsg) = @_;
74 my ($part) = @$p; # ignore $depth and @idx;
75 my $hdr_lines; # diff --git a/... b/...
77 my $ct = $part->content_type || 'text/plain';
78 my ($s, undef) = msg_part_text($part, $ct);
82 # Email::MIME::Encodings forces QP to be CRLF upon decoding,
83 # change it back to LF:
84 my $cte = $part->header('Content-Transfer-Encoding') || '';
85 if ($cte =~ /\bquoted-printable\b/i && $part->crlf eq "\n") {
89 foreach my $l (split(/^/m, $s)) {
95 if ($mode_a =~ /\A(?:100644|120000|100755)\z/) {
96 $di->{mode_a} = $mode_a;
100 # start writing the diff out to a tempfile
101 open($tmp, '+>', undef) or die "open(tmp): $!";
104 push @$hdr_lines, $l;
105 $di->{hdr_lines} = $hdr_lines;
106 print $tmp @$hdr_lines, $l or die "print(tmp): $!";
108 # for debugging/diagnostics:
111 } elsif ($l =~ m!\Adiff --git ("?a/.+) ("?b/.+)$!) {
112 return $di if $tmp; # got our blob, done!
114 my ($path_a, $path_b) = ($1, $2);
116 # diff header lines won't have \r because git
117 # will quote them, but Email::MIME gives CRLF
118 # for quoted-printable:
121 # don't care for leading 'a/' and 'b/'
122 my (undef, @a) = split(m{/}, git_unquote($path_a));
123 my (undef, @b) = split(m{/}, git_unquote($path_b));
125 # get rid of path-traversal attempts and junk patches:
127 return if $bad_component{$_};
130 $di->{path_a} = join('/', @a);
131 $di->{path_b} = join('/', @b);
134 print $tmp $l or die "print(tmp): $!";
135 } elsif ($hdr_lines) {
136 push @$hdr_lines, $l;
137 if ($l =~ /\Anew file mode (100644|120000|100755)$/) {
145 sub path_searchable ($) { defined($_[0]) && $_[0] =~ m!\A[\w/\. \-]+\z! }
147 sub find_extract_diff ($$$) {
148 my ($self, $ibx, $want) = @_;
149 my $srch = $ibx->search or return;
151 my $post = $want->{oid_b} or die 'BUG: no {oid_b}';
152 $post =~ /\A[a-f0-9]+\z/ or die "BUG: oid_b not hex: $post";
154 my $q = "dfpost:$post";
155 my $pre = $want->{oid_a};
156 if (defined $pre && $pre =~ /\A[a-f0-9]+\z/) {
159 $pre = '[a-f0-9]{7}'; # for $re below
162 my $path_b = $want->{path_b};
163 if (path_searchable($path_b)) {
164 $q .= qq{ dfn:"$path_b"};
166 my $path_a = $want->{path_a};
167 if (path_searchable($path_a) && $path_a ne $path_b) {
168 $q .= qq{ dfn:"$path_a"};
172 my $msgs = $srch->query($q, { relevance => 1 });
173 my $re = qr/\Aindex ($pre[a-f0-9]*)\.\.($post[a-f0-9]*)(?: (\d+))?/;
176 foreach my $smsg (@$msgs) {
177 $ibx->smsg_mime($smsg) or next;
178 msg_iter(delete($smsg->{mime}), sub {
179 $di ||= extract_diff($_[0], $re, $ibx, $smsg);
185 # pure Perl "git init"
186 sub do_git_init_wt ($) {
188 my $wt = File::Temp->newdir('solver.wt-XXXXXXXX', TMPDIR => 1);
189 my $dir = $wt->dirname;
191 foreach ('', qw(objects refs objects/info refs/heads)) {
192 mkdir("$dir/.git/$_") or die "mkdir $_: $!";
194 open my $fh, '>', "$dir/.git/config" or die "open .git/config: $!";
195 print $fh <<'EOF' or die "print .git/config $!";
197 repositoryFormatVersion = 0
200 fsyncObjectfiles = false
201 logAllRefUpdates = false
203 close $fh or die "close .git/config: $!";
205 open $fh, '>', "$dir/.git/HEAD" or die "open .git/HEAD: $!";
206 print $fh "ref: refs/heads/master\n" or die "print .git/HEAD: $!";
207 close $fh or die "close .git/HEAD: $!";
209 my $f = '.git/objects/info/alternates';
210 open $fh, '>', "$dir/$f" or die "open: $f: $!";
211 print($fh (map { "$_->{git_dir}/objects\n" } @{$self->{gits}})) or
213 close $fh or die "close: $f: $!";
217 sub extract_old_mode ($) {
219 if (grep(/\Aold mode (100644|100755|120000)$/, @{$di->{hdr_lines}})) {
226 my ($pid, $msg) = @_;
227 waitpid($pid, 0) == $pid or die "waitpid($msg): $!";
228 $? == 0 or die "$msg failed: $?";
231 sub prepare_index ($$$$) {
232 my ($out, $wt_dir, $existing, $di) = @_;
233 my $oid_full = $existing->[1];
235 my $path_a = $di->{path_a} or die "BUG: path_a missing for $oid_full";
236 my $mode_a = $di->{mode_a} || extract_old_mode($di);
238 # unlike git-apply(1), this only gets called once in a patch
239 # series and happens too quickly to be worth making async:
240 pipe($r, $w) or die "pipe: $!";
241 my $rdr = { 0 => fileno($r) };
242 my $pid = spawn([qw(git -C), $wt_dir,
243 qw(update-index -z --index-info)], undef, $rdr);
244 close $r or die "close pipe(r): $!";
245 print $w "$mode_a $oid_full\t$path_a\0" or die "print update-index: $!";
247 close $w or die "close update-index: $!";
248 reap($pid, 'update-index -z --index-info');
250 print $out "index prepared:\n",
251 "$mode_a $oid_full\t", git_quote($path_a), "\n";
254 sub do_apply_begin ($$$) {
255 my ($out, $wt_dir, $di) = @_;
257 my $tmp = delete $di->{tmp} or die "BUG: no tmp ", di_url($di);
258 $tmp->flush or die "tmp->flush failed: $!";
259 $out->flush or die "err->flush failed: $!";
260 sysseek($tmp, 0, SEEK_SET) or die "sysseek(tmp) failed: $!";
262 defined(my $err_fd = fileno($out)) or die "fileno(out): $!";
263 my $rdr = { 0 => fileno($tmp), 1 => $err_fd, 2 => $err_fd };
265 # we need --ignore-whitespace because some patches are CRLF
266 my $cmd = [ qw(git -C), $wt_dir,
267 qw(apply --cached --ignore-whitespace
268 --whitespace=warn --verbose) ];
269 spawn($cmd, undef, $rdr);
272 sub do_apply_continue ($$) {
273 my ($wt_dir, $apply_pid) = @_;
274 reap($apply_pid, 'apply');
275 popen_rd([qw(git -C), $wt_dir, qw(ls-files -s -z)]);
278 sub do_apply_end ($$$$) {
279 my ($out, $wt_git, $rd, $di) = @_;
282 defined(my $line = <$rd>) or die "failed to read ls-files: $!";
283 chomp $line or die "no trailing \\0 in [$line] from ls-files";
285 my ($info, $file) = split(/\t/, $line, 2);
286 my ($mode_b, $oid_b_full, $stage) = split(/ /, $info);
288 defined($line = <$rd>) and die "extra files in index: $line";
289 close $rd or die "close ls-files: $?";
291 $file eq $di->{path_b} or
292 die "index mismatch: file=$file != path_b=$di->{path_b}";
294 my (undef, undef, $size) = $wt_git->check($oid_b_full);
296 defined($size) or die "failed to read_size from $oid_b_full";
298 print $out "$mode_b $oid_b_full\t$file\n";
299 [ $wt_git, $oid_b_full, 'blob', $size, $di ];
304 # note: we don't pass the PSGI env here, different inboxes
305 # can have different HTTP_HOST on the same instance.
306 my $url = $di->{ibx}->base_url;
307 my $mid = $di->{smsg}->{mid};
308 defined($url) ? "$url$mid/" : "<$mid>";
311 # reconstruct the oid_b blob using patches we found:
312 sub apply_patches_cb ($$$$$) {
313 my ($self, $out, $found, $patches, $oid_b) = @_;
315 my $tot = scalar(@$patches) or return sub {
316 print $out "no patch(es) for $oid_b\n";
320 my $wt = do_git_init_wt($self);
321 my $wt_dir = $wt->dirname;
322 my $wt_git = PublicInbox::Git->new("$wt_dir/.git");
323 $wt_git->{-wt} = $wt;
326 my ($apply_pid, $rd, $di);
328 # returns an empty string if in progress, undef if not found,
329 # or the final [ ::Git, oid_full, type, size, $di ] arrayref
333 $found->{$di->{oid_b}} =
334 do_apply_end($out, $wt_git, $rd, $di);
336 # continue to shift @$patches
337 } elsif ($apply_pid) {
338 $rd = do_apply_continue($wt_dir, $apply_pid);
340 return ''; # $rd => do_apply_ned
343 # may return undef here
344 $di = shift @$patches or return $found->{$oid_b};
347 my $oid_a = $di->{oid_a};
348 my $existing = $found->{$oid_a};
349 my $empty_oid = $oid_a =~ /\A0+\z/;
351 if ($empty_oid && $i != 1) {
352 die "empty oid at [$i/$tot] ", di_url($di);
354 if (!$existing && !$empty_oid) {
355 die "missing $oid_a at [$i/$tot] ", di_url($di);
358 # prepare the worktree for patch application:
359 if ($i == 1 && $existing) {
360 prepare_index($out, $wt_dir, $existing, $di);
363 print $out "\napplying [$i/$tot] ", di_url($di), "\n",
364 join('', @{$di->{hdr_lines}}), "\n"
365 or die "print \$out failed: $!";
367 # begin the patch application patch!
368 $apply_pid = do_apply_begin($out, $wt_dir, $di);
369 # next call to this callback will call do_apply_continue
375 # Returns an array ref: [ ::Git object, oid_full, type, size, di ]
376 # or undef if nothing was found.
378 # TODO: complete the migration of this and ViewVCS into an evented
381 my ($self, $out, $oid_b, $hints) = @_;
383 # should we even get here? Probably not, but somebody
384 # could be manually typing URLs:
385 return if $oid_b =~ /\A0+\z/;
387 my $req = { %$hints, oid_b => $oid_b };
389 my $found = {}; # { abbrev => [ ::Git, oid_full, type, size, $di ] }
390 my $patches = []; # [ array of $di hashes ]
391 my $max = $self->{max_patches} || 200;
394 my $want = pop @todo;
396 $apply_cb ||= apply_patches_cb($self, $out, $found,
398 return $apply_cb->();
401 if (scalar(@$patches) > $max) {
402 print $out "Aborting, too many steps to $oid_b\n";
405 # see if we can find the blob in an existing git repo:
406 my $want_oid = $want->{oid_b};
407 if (my $existing = solve_existing($self, $out, $want)) {
408 print $out "found $want_oid in ",
409 join("\n", $existing->[0]->pub_urls), "\n";
411 return $existing if $want_oid eq $oid_b; # DONE!
412 $found->{$want_oid} = $existing;
413 return ''; # ok, one blob resolved, more to go?
416 # scan through inboxes to look for emails which results in
419 foreach my $ibx (@{$self->{inboxes}}) {
420 $di = find_extract_diff($self, $ibx, $want) or next;
422 unshift @$patches, $di;
423 print $out "found $want_oid in ",di_url($di),"\n";
425 # good, we can find a path to the oid we $want, now
426 # lets see if we need to apply more patches:
427 my $src = $di->{oid_a};
429 last if $src =~ /\A0+\z/;
431 # we have to solve it using another oid, fine:
432 my $job = { oid_b => $src, path_b => $di->{path_a} };
434 last; # onto the next @todo item
437 print $out "$want_oid could not be found\n";
440 ''; # continue onto next @todo item;
444 my $ret = eval { $cb->() };
445 unless (defined($ret)) {
446 print $out "E: $@\n" if $@;
449 return $ret if ref($ret);
450 # $ret == ''; so continue looping here