Thanks very much for the super quick turnaround. The leak is now gone :-)

Michael

On Thu, Sep 27, 2012 at 10:37 AM, K. John Wu <[email protected]> wrote:
> Hi, Michael,
>
> Looks like this particular problem might be caused by an incorrect
> handling of a special case in the function
> ibis::selectClause::addAgregado.  The variable generated from
> selectParser.yy:252 (selectParser.cc:252 actually refers to
> selectParse.yy:252) was passed to the function addAgregado, however,
> the expression count(*) (generated to replace avg(something) with
> sum(something) / count(*)) must have appeared in the select clause
> already and therefore is no longer needed.  The fix is to free the
> pointer to the variable object in the function addAgregado.
>
> Please give SVN Revision 582 a try when you get the chance and let us
> know if it continues to have problems.
>
> Thanks.
>
> John
>
>
> On 9/27/12 9:41 AM, Michael Beauregard wrote:
>> Here is the stack trace of a memory leak reported by xcode leak profiler:
>>
>>    0 libsystem_c.dylib malloc
>>    1 libstdc++.6.dylib operator new(unsigned long)
>>    2 libstdc++.6.dylib operator new[](unsigned long)
>>    3 Diagnostics ibis::util::strnewdup(char const*) 
>> ./FastBit/src/util.cpp:1383
>>    4 Diagnostics ibis::math::variable::variable(char const*)
>> ./FastBit/src/qExpr.h:789
>>    5 Diagnostics ibis::selectParser::parse() ./FastBit/selectParser.cc:252
>>    6 Diagnostics ibis::selectClause::parse(char const*)
>> ./FastBit/src/selectClause.cpp:107
>>    7 Diagnostics ibis::selectClause::selectClause(char const*)
>> ./FastBit/src/selectClause.cpp:17
>>    8 Diagnostics ibis::table::select(std::vector<ibis::part const*,
>> std::allocator<ibis::part const*> > const&, char const*, char const*)
>> ./FastBit/src/filter.cpp:2659
>>    9 Diagnostics ibis::mensa::select(char const*, char const*) const
>> ./FastBit/src/mensa.cpp:566
>>
>> The line number selectParser.cc:252 doesn't seem correct for me which
>> doesn't make much sense, but the rest seem to line up fine.
>>
>> Digging into it a little, I fear that the bug is actually in the
>> parser generator, but I'm not terribly familiar with how all of that
>> is put together. Let me know if I can help at all.
>>
>> Thanks,
>>
>> Michael
>> _______________________________________________
>> FastBit-users mailing list
>> [email protected]
>> https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users
>>
_______________________________________________
FastBit-users mailing list
[email protected]
https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users

Reply via email to