searchidx: all indexers check for bad blobs

This should help us detect bugs in our code or storage
synchronization problems more easily.  This probably won't
detect corrupted git storage, but can detect corrupted SQLite
files.

"Bad blobs, bad blobs, whatcha gonna do when they come for you?"

3 files changed, 21 insertions, 15 deletions

diff --git a/lib/PublicInbox/ExtSearchIdx.pm b/lib/PublicInbox/ExtSearchIdx.pm
index 11f7786d..b0a12bca 100644
--- a/lib/PublicInbox/ExtSearchIdx.pm
+++ b/lib/PublicInbox/ExtSearchIdx.pm
@@ -19,7 +19,8 @@ use v5.10.1;
 use parent qw(PublicInbox::ExtSearch PublicInbox::Lock);
 use Carp qw(croak carp);
 use PublicInbox::Search;
-use PublicInbox::SearchIdx qw(crlf_adjust prepare_stack is_ancestor);
+use PublicInbox::SearchIdx qw(crlf_adjust prepare_stack is_ancestor
+        is_bad_blob);
 use PublicInbox::OverIdx;
 use PublicInbox::MiscIdx;
 use PublicInbox::MID qw(mids);
@@ -91,16 +92,6 @@ sub attach_config {
         $cfg->each_inbox(\&_ibx_attach, $self);
 }
 
-sub is_bad_blob ($$$$) {
-        my ($oid, $type, $size, $expect_oid) = @_;
-        if ($type ne 'blob') {
-                carp "W: $expect_oid is not a blob (type=$type)";
-                return 1;
-        }
-        croak "BUG: $oid != $expect_oid" if $oid ne $expect_oid;
-        $size == 0 ? 1 : 0; # size == 0 means purged
-}
-
 sub check_batch_limit ($) {
         my ($req) = @_;
         my $self = $req->{self};
diff --git a/lib/PublicInbox/SearchIdx.pm b/lib/PublicInbox/SearchIdx.pm
index 0124dd11..0fbe6560 100644
--- a/lib/PublicInbox/SearchIdx.pm
+++ b/lib/PublicInbox/SearchIdx.pm
@@ -15,7 +15,7 @@ use PublicInbox::InboxWritable;
 use PublicInbox::MID qw(mids_for_index mids);
 use PublicInbox::MsgIter;
 use PublicInbox::IdxStack;
-use Carp qw(croak);
+use Carp qw(croak carp);
 use POSIX qw(strftime);
 use Time::Local qw(timegm);
 use PublicInbox::OverIdx;
@@ -23,7 +23,7 @@ use PublicInbox::Spawn qw(spawn nodatacow_dir);
 use PublicInbox::Git qw(git_unquote);
 use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
 our @EXPORT_OK = qw(crlf_adjust log2stack is_ancestor check_size prepare_stack
-        index_text term_generator add_val);
+        index_text term_generator add_val is_bad_blob);
 my $X = \%PublicInbox::Search::X;
 our ($DB_CREATE_OR_OPEN, $DB_OPEN);
 our $DB_NO_SYNC = 0;
@@ -591,8 +591,19 @@ sub crlf_adjust ($) {
         }
 }
 
+sub is_bad_blob ($$$$) {
+        my ($oid, $type, $size, $expect_oid) = @_;
+        if ($type ne 'blob') {
+                carp "W: $expect_oid is not a blob (type=$type)";
+                return 1;
+        }
+        croak "BUG: $oid != $expect_oid" if $oid ne $expect_oid;
+        $size == 0 ? 1 : 0; # size == 0 means purged
+}
+
 sub index_both { # git->cat_async callback
         my ($bref, $oid, $type, $size, $sync) = @_;
+        return if is_bad_blob($oid, $type, $size, $sync->{oid});
         my ($nr, $max) = @$sync{qw(nr max)};
         ++$$nr;
         $$max -= $size;
@@ -609,6 +620,7 @@ sub index_both { # git->cat_async callback
 
 sub unindex_both { # git->cat_async callback
         my ($bref, $oid, $type, $size, $sync) = @_;
+        return if is_bad_blob($oid, $type, $size, $sync->{oid});
         unindex_eml($sync->{sidx}, $oid, PublicInbox::Eml->new($bref));
         # may be undef if leftover
         if (defined(my $cur_cmt = $sync->{cur_cmt})) {
@@ -713,7 +725,7 @@ sub process_stack {
                 $sync->{index_oid} = \&index_both;
         }
         while (my ($f, $at, $ct, $oid, $cur_cmt) = $stk->pop_rec) {
-                my $arg = { %$sync, cur_cmt => $cur_cmt };
+                my $arg = { %$sync, cur_cmt => $cur_cmt, oid => $oid };
                 last if $sync->{quit};
                 if ($f eq 'm') {
                         $arg->{autime} = $at;
diff --git a/lib/PublicInbox/V2Writable.pm b/lib/PublicInbox/V2Writable.pm
index 5aec7561..07a7fa42 100644
--- a/lib/PublicInbox/V2Writable.pm
+++ b/lib/PublicInbox/V2Writable.pm
@@ -17,7 +17,8 @@ use PublicInbox::InboxWritable;
 use PublicInbox::OverIdx;
 use PublicInbox::Msgmap;
 use PublicInbox::Spawn qw(spawn popen_rd);
-use PublicInbox::SearchIdx qw(log2stack crlf_adjust is_ancestor check_size);
+use PublicInbox::SearchIdx qw(log2stack crlf_adjust is_ancestor check_size
+        is_bad_blob);
 use IO::Handle; # ->autoflush
 use File::Temp ();
 
@@ -896,6 +897,7 @@ sub reindex_checkpoint ($$) {
 
 sub index_oid { # cat_async callback
         my ($bref, $oid, $type, $size, $arg) = @_;
+        return if is_bad_blob($oid, $type, $size, $arg->{oid});
         my $self = $arg->{self};
         local $self->{current_info} = "$self->{current_info} $oid";
         return if $size == 0; # purged
@@ -1147,6 +1149,7 @@ sub unindex_oid_aux ($$$) {
 
 sub unindex_oid ($$;$) { # git->cat_async callback
         my ($bref, $oid, $type, $size, $sync) = @_;
+        return if is_bad_blob($oid, $type, $size, $sync->{oid});
         my $self = $sync->{self};
         local $self->{current_info} = "$self->{current_info} $oid";
         my $unindexed = $sync->{in_unindex} ? $sync->{unindexed} : undef;