use Test::Exception;
use Test::Warn;
use Time::HiRes 'time';
-use Config;
+use Math::BigInt;
use lib qw(t/lib);
use DBICTest;
-
-# savepoints test
-{
- my $schema = DBICTest->init_schema(auto_savepoint => 1);
-
- my $ars = $schema->resultset('Artist');
-
- # test two-phase commit and inner transaction rollback from nested transactions
- $schema->txn_do(sub {
- $ars->create({ name => 'in_outer_transaction' });
- $schema->txn_do(sub {
- $ars->create({ name => 'in_inner_transaction' });
- });
- ok($ars->search({ name => 'in_inner_transaction' })->first,
- 'commit from inner transaction visible in outer transaction');
- throws_ok {
- $schema->txn_do(sub {
- $ars->create({ name => 'in_inner_transaction_rolling_back' });
- die 'rolling back inner transaction';
- });
- } qr/rolling back inner transaction/, 'inner transaction rollback executed';
- $ars->create({ name => 'in_outer_transaction2' });
- });
-
- ok($ars->search({ name => 'in_outer_transaction' })->first,
- 'commit from outer transaction');
- ok($ars->search({ name => 'in_outer_transaction2' })->first,
- 'second commit from outer transaction');
- ok($ars->search({ name => 'in_inner_transaction' })->first,
- 'commit from inner transaction');
- is $ars->search({ name => 'in_inner_transaction_rolling_back' })->first,
- undef,
- 'rollback from inner transaction';
-}
+use DBIx::Class::_Util qw( sigwarn_silencer modver_gt_or_eq modver_gt_or_eq_and_lt );
# check that we work somewhat OK with braindead SQLite transaction handling
#
# As per https://metacpan.org/source/ADAMK/DBD-SQLite-1.37/lib/DBD/SQLite.pm#L921
# SQLite does *not* try to synchronize
+#
+# However DBD::SQLite 1.38_02 seems to fix this, with an accompanying test:
+# https://metacpan.org/source/ADAMK/DBD-SQLite-1.38_02/t/54_literal_txn.t
+
+require DBD::SQLite;
+my $lit_txn_todo = modver_gt_or_eq('DBD::SQLite', '1.38_02')
+ ? undef
+ : "DBD::SQLite before 1.38_02 is retarded wrt detecting literal BEGIN/COMMIT statements"
+;
for my $prefix_comment (qw/Begin_only Commit_only Begin_and_Commit/) {
note "Testing with comment prefixes on $prefix_comment";
# FIXME warning won't help us for the time being
# perhaps when (if ever) DBD::SQLite gets fixed,
# we can do something extra here
- local $SIG{__WARN__} = sub { warn @_ if $_[0] !~ /Internal transaction state .+? does not seem to match/ }
- unless $ENV{TEST_VERBOSE};
+ local $SIG{__WARN__} = sigwarn_silencer( qr/Internal transaction state .+? does not seem to match/ )
+ if ( $lit_txn_todo && !$ENV{TEST_VERBOSE} );
my ($c_begin, $c_commit) = map { $prefix_comment =~ $_ ? 1 : 0 } (qr/Begin/, qr/Commit/);
);
ok ($schema->storage->connected, 'Still connected');
{
- local $TODO = 'SQLite is retarded wrt detecting BEGIN' if $c_begin;
+ local $TODO = $lit_txn_todo if $c_begin;
ok (! $schema->storage->_dbh->{AutoCommit}, "DBD aware of txn begin with comments on $prefix_comment");
}
);
ok ($schema->storage->connected, 'Still connected');
{
- local $TODO = 'SQLite is retarded wrt detecting COMMIT' if $c_commit;
+ local $TODO = $lit_txn_todo if $c_commit and ! $c_begin;
ok ($schema->storage->_dbh->{AutoCommit}, "DBD aware txn ended with comments on $prefix_comment");
}
{
# this never worked in the 1st place
- local $TODO = 'SQLite is retarded wrt detecting COMMIT' if ! $c_begin and $c_commit;
+ local $TODO = $lit_txn_todo if ! $c_begin and $c_commit;
# odd argument passing, because such nested crefs leak on 5.8
lives_ok {
}
}
+# test blank begin/svp/commit/begin cycle
+warnings_are {
+ my $schema = DBICTest->init_schema( no_populate => 1 );
+ my $rs = $schema->resultset('Artist');
+ is ($rs->count, 0, 'Start with empty table');
+
+ for my $do_commit (1, 0) {
+ $schema->txn_begin;
+ $schema->svp_begin;
+ $schema->svp_rollback;
+
+ $schema->svp_begin;
+ $schema->svp_rollback;
+
+ $schema->svp_release;
+
+ $schema->svp_begin;
+
+ $schema->txn_rollback;
+
+ $schema->txn_begin;
+ $schema->svp_begin;
+ $schema->svp_rollback;
+
+ $schema->svp_begin;
+ $schema->svp_rollback;
+
+ $schema->svp_release;
+
+ $schema->svp_begin;
+
+ $do_commit ? $schema->txn_commit : $schema->txn_rollback;
+
+ is_deeply $schema->storage->savepoints, [], 'Savepoint names cleared away'
+ }
+
+ $schema->txn_do(sub {
+ ok (1, 'all seems fine');
+ });
+} [], 'No warnings emitted';
my $schema = DBICTest->init_schema();
# make sure the side-effects of RT#67581 do not result in data loss
my $row;
warnings_exist { $row = $schema->resultset('Artist')->create ({ name => 'alpha rank', rank => 'abc' }) }
- [qr/Non-numeric value supplied for column 'rank' despite the numeric datatype/],
+ [qr/Non-integer value supplied for column 'rank' despite the integer datatype/],
'proper warning on string insertion into an numeric column'
;
$row->discard_changes;
is ($row->rank, 'abc', 'proper rank inserted into database');
# and make sure we do not lose actual bigints
+SKIP: {
+
+skip "Not testing bigint handling on known broken DBD::SQLite trial versions", 1
+ if modver_gt_or_eq_and_lt( 'DBD::SQLite', '1.45', '1.45_03' );
+
{
package DBICTest::BigIntArtist;
use base 'DBICTest::Schema::Artist';
$_[1]->do('ALTER TABLE artist ADD COLUMN bigint BIGINT');
});
+my $sqlite_broken_bigint = modver_gt_or_eq_and_lt( 'DBD::SQLite', '1.34', '1.37' );
+
+# 63 bit integer
+my $many_bits = (Math::BigInt->new(2) ** 62);
+
# test upper/lower boundaries for sqlite and some values inbetween
# range is -(2**63) .. 2**63 - 1
-SKIP: {
- skip 'This perl does not seem to have 64bit int support - DBI roundtrip of large int will fail with DBD::SQLite < 1.37', 1
- if ($Config{ivsize} < 8 and ! eval { DBD::SQLite->VERSION(1.37); 1 });
-
- for my $bi (qw/
- -9223372036854775808
- -9223372036854775807
- -8694837494948124658
- -6848440844435891639
- -5664812265578554454
- -5380388020020483213
- -2564279463598428141
- 2442753333597784273
- 4790993557925631491
- 6773854980030157393
- 7627910776496326154
- 8297530189347439311
- 9223372036854775806
- 9223372036854775807
- /) {
+#
+# Not testing -0 - it seems to overflow to ~0 on some combinations,
+# thus not triggering the >32 bit guards
+# interesting read: https://en.wikipedia.org/wiki/Signed_zero#Representations
+for my $bi ( qw(
+ -2
+ -1
+ 0
+ +0
+ 1
+ 2
+
+ -9223372036854775808
+ -9223372036854775807
+ -8694837494948124658
+ -6848440844435891639
+ -5664812265578554454
+ -5380388020020483213
+ -2564279463598428141
+ 2442753333597784273
+ 4790993557925631491
+ 6773854980030157393
+ 7627910776496326154
+ 8297530189347439311
+ 9223372036854775806
+ 9223372036854775807
+
+ 4294967295
+ 4294967296
+
+ -4294967296
+ -4294967295
+ -4294967294
+
+ -2147483649
+ -2147483648
+ -2147483647
+ -2147483646
+
+ 2147483646
+ 2147483647
+),
+ # these values cause exceptions even with all workarounds in place on these
+ # fucked DBD::SQLite versions *regardless* of ivsize >.<
+ $sqlite_broken_bigint
+ ? ()
+ : ( '2147483648', '2147483649' )
+) {
+ # unsigned 32 bit ints have a range of −2,147,483,648 to 2,147,483,647
+ # alternatively expressed as the hexadecimal numbers below
+ # the comparison math will come out right regardless of ivsize, since
+ # we are operating within 31 bits
+ # P.S. 31 because one bit is lost for the sign
+ my $v_bits = ($bi > 0x7fff_ffff || $bi < -0x8000_0000) ? 64 : 32;
+
+ my $v_desc = sprintf '%s (%d bit signed int)', $bi, $v_bits;
+
+ my @w;
+ local $SIG{__WARN__} = sub {
+ if ($_[0] =~ /datatype mismatch/) {
+ push @w, @_;
+ }
+ elsif ($_[0] =~ /An integer value occupying more than 32 bits was supplied .+ can not bind properly so DBIC will treat it as a string instead/ ) {
+ # do nothing, this warning will pop up here and there depending on
+ # DBD/bitness combination
+ # we don't want to test for it explicitly, we are just interested
+ # in the results matching at the end
+ }
+ else {
+ warn @_;
+ }
+ };
+
+ # some combinations of SQLite 1.35 and older 5.8 faimly is wonky
+ # instead of a warning we get a full exception. Sod it
+ eval {
$row = $schema->resultset('BigIntArtist')->create({ bigint => $bi });
- is ($row->bigint, $bi, "value in object correct ($bi)");
+ } or do {
+ fail("Exception on inserting $v_desc") unless $sqlite_broken_bigint;
+ next;
+ };
+
+ # explicitly using eq, to make sure we did not nummify the argument
+ # which can be an issue on 32 bit ivsize
+ cmp_ok ($row->bigint, 'eq', $bi, "value in object correct ($v_desc)");
+
+ $row->discard_changes;
+
+ cmp_ok (
+ $row->bigint,
+
+ # the test will not pass an == if we are running under 32 bit ivsize
+ # use 'eq' on the numified (and possibly "scientificied") returned value
+ (DBIx::Class::_ENV_::IV_SIZE < 8 and $v_bits > 32) ? 'eq' : '==',
+ # in 1.37 DBD::SQLite switched to proper losless representation of bigints
+ # regardless of ivize
+ # before this use 'eq' (from above) on the numified (and possibly
+ # "scientificied") returned value
+ (DBIx::Class::_ENV_::IV_SIZE < 8 and ! modver_gt_or_eq('DBD::SQLite', '1.37')) ? $bi+0 : $bi,
+
+ "value in database correct ($v_desc)"
+ );
+
+# FIXME - temporary smoke-only escape
+SKIP: {
+ skip 'Potential for false negatives - investigation pending', 1
+ if DBICTest::RunMode->is_plain;
+
+ # check if math works
+ # start by adding/subtracting a 50 bit integer, and then divide by 2 for good measure
+ my ($sqlop, $expect) = $bi < 0
+ ? ( '(bigint + ? )', ($bi + $many_bits) )
+ : ( '(bigint - ? )', ($bi - $many_bits) )
+ ;
+
+ $expect = ($expect + ($expect % 2)) / 2;
+
+ # read https://en.wikipedia.org/wiki/Modulo_operation#Common_pitfalls
+ # and check the tables on the right side of the article for an
+ # enlightening journey on why a mere bigint % 2 won't work
+ $sqlop = "( $sqlop + ( ((bigint % 2)+2)%2 ) ) / 2";
+
+ for my $dtype (undef, \'int', \'bigint') {
+
+ # FIXME - the double-load should not be needed
+ # will fix in the future
+ $row->update({ bigint => $bi });
+ $row->discard_changes;
+ $row->update({ bigint => \[ $sqlop, [ $dtype => $many_bits ] ] });
$row->discard_changes;
- is ($row->bigint, $bi, "value in database correct ($bi)");
+
+ # can't use cmp_ok - will not engage the M::BI overload of $many_bits
+ ok (
+ $row->bigint
+
+ ==
+
+ (DBIx::Class::_ENV_::IV_SIZE < 8 and ! modver_gt_or_eq('DBD::SQLite', '1.37')) ? $expect->bstr + 0 : $expect
+ , "simple integer math with@{[ $dtype ? '' : 'out' ]} bindtype in database correct (base $v_desc)")
+ or diag sprintf '%s != %s', $row->bigint, $expect;
}
+# end of fixme
}
+ is_deeply (\@w, [], "No mismatch warnings on bigint operations ($v_desc)" );
+
+}}
+
done_testing;
# vim:sts=2 sw=2: