Jeff Zucker skrev:

> Henrik Tougaard wrote:
>> 
>> t/50dbm.t still fail after test 37:
>> 
>> --- Using DB_File () ---
>> DBD::DBM         0.01 using DB_File
>>   DBD::File      0.30
>>   DBI::SQL::Nano 0.01
>> DBI              1.42
>> OS               dec_osf (4.0g)
>> Perl             5.008003 (alpha-dec_osf)
>> 
>>  UPDATE DB_File_fruit SET dVal='apples' WHERE dKey=2
>> Segmentation fault (core dumped)
> 
> Henrik, could you try a quick test for me that may or may not fix
> this: I have two versions of fetch_row starting at line 523 of
> DBM.pm.  Could you comment out everything in the "# fetch with %each"
> section and uncomment everything in the "# fetch without %each"
> section and let me know if it makes any difference?

I tried rc3 using perl 5.006001 on alpha-dec_osf.

Still fails on 50dbm:
t/50dbm................dubious

        Test returned status -1 (wstat 139, 0x8b)
Executing /usr/bin/dbx "/usr/local/bin/perl" "core" (where)...
dbx version 5.1
Type 'help' for help.
Core file created by program "perl"

signal Segmentation fault at >*[__free_ovflpage, 0x3ff810cb040]         ldl
t2, 24(a1)
>  0 __free_ovflpage(0x4, 0x14039bcd0, 0x3ff810c6b34, 0x1403a1810,
0x14039bcd0) [0x3ff810cb040]
   1 __big_delete(0x3ff810c4314, 0x14058680a, 0x1, 0x4, 0x14039bcd0)
[0x3ff810c6b50]
   2 (unknown)() [0x3ff810c4310]

/usr/bin/dbx: fatal error: hit EOF on standard input file

DIED. FAILED tests 38-132
        Failed 95/132 tests, 28.03% okay

__________________
and with the change you suggested:
t/50dbm................dubious

        Test returned status -1 (wstat 139, 0x8b)
Executing /usr/bin/dbx "/usr/local/bin/perl" "core" (where)...
dbx version 5.1
Type 'help' for help.
Core file created by program "perl"

signal Segmentation fault at >*[__free_ovflpage, 0x3ff810cb040]         ldl
t2, 24(a1)
>  0 __free_ovflpage(0x4, 0x1403b3850, 0x3ff810c6b34, 0x140522010,
0x1403b3850) [0x3ff810cb040]
   1 __big_delete(0x3ff810c4314, 0x14053380a, 0x1, 0x4, 0x1403b3850)
[0x3ff810c6b50]
   2 (unknown)() [0x3ff810c4310]

/usr/bin/dbx: fatal error: hit EOF on standard input file

DIED. FAILED tests 38-132
        Failed 95/132 tests, 28.03% okay
___________________

And switching to perl 5.008003 on alpha-dec_osf, it still fails on 50dbm:
t/zz_50dbm_pp..........dubious

        Test returned status -1 (wstat 139, 0x8b)
        test program seems to have generated a core
DIED. FAILED tests 38-132
        Failed 95/132 tests, 28.03% okay


This being morning, and me being a tiny bit more alert, I tried tracing the
failure in the debugger, and found that it timed out in:
408:                return $v if uc $k eq uc $stmt_col_name;
DBI::SQL::Nano::Statement_::column_nums(/home/ht000/work/DBI-1.42/blib/lib/D
BI/SQL/Nano.pm:408):
408:                return $v if uc $k eq uc $stmt_col_name;
DBI::SQL::Nano::Statement_::UPDATE(/home/ht000/work/DBI-1.42/blib/lib/DBI/SQ
L/Nano.pm:364):
364:                for ($i = 0;  $i < @{$self->{column_names}};  $i++) {
DBI::SQL::Nano::Statement_::UPDATE(/home/ht000/work/DBI-1.42/blib/lib/DBI/SQ
L/Nano.pm:368):
368:                $affected++;
DBI::SQL::Nano::Statement_::UPDATE(/home/ht000/work/DBI-1.42/blib/lib/DBI/SQ
L/Nano.pm:369):
369:                if ($self->{fetched_from_key}){
DBI::SQL::Nano::Statement_::UPDATE(/home/ht000/work/DBI-1.42/blib/lib/DBI/SQ
L/Nano.pm:370):
370:                    $table->update_one_row($data,$array);
DBD::DBM::Table::update_one_row(/home/ht000/work/DBI-1.42/blib/lib/DBD/DBM.p
m:600):
600:        my($self,$data,$aryref) = @_;
DBD::DBM::Table::update_one_row(/home/ht000/work/DBI-1.42/blib/lib/DBD/DBM.p
m:601):
601:        my $key = shift @$aryref;
DBD::DBM::Table::update_one_row(/home/ht000/work/DBI-1.42/blib/lib/DBD/DBM.p
m:602):
602:        return undef unless defined $key;
DBD::DBM::Table::update_one_row(/home/ht000/work/DBI-1.42/blib/lib/DBD/DBM.p
m:603):
603:        if( ref $aryref->[0] eq 'ARRAY'){
DBD::DBM::Table::update_one_row(/home/ht000/work/DBI-1.42/blib/lib/DBD/DBM.p
m:606):
606:        $self->{hash}->{$key}=$aryref->[0];
... and everything stood still for a loooong time. That didn't look like the
coredumps I had earlier, so maybe the debugger interfered.

Trying with DBI->trace(9) shows that the core dump is in:
    >> execute     DISPATCH (DBI::st=HASH(0x14047f730) rc1/1 @1 g0 ima1041
pid#98253) at t/50dbm.t line 120 via t/50dbm.t line 64
    -> execute for DBD::DBM::st (DBI::st=HASH(0x14047f730)~0x14047f600)
(the entire trace is attached).


Is there somethingelse I can try?

--
Henrik

Attachment: trace.log
Description: Binary data

Reply via email to