|
Повідомлення створено: 07. 03. 2017 [10:00]
|
|
amirkhan63
amir amir
Автор теми
Зареєстрован(а) с: 07.03.2017
Повідомлення: 2
|
Hi,
I was tested my project developed on openscada 0.8.17 with Valgrind and massif and reported memory leakage on tarchval.cpp (function: void TValBuf::TBuf<TpVal>::set( TpVal value, int64_t tm )) code below:
//>>> Memory consumption is made optimal for limited size
if((int)buf.grid->capacity() > size)
{
vector<TpVal> *ns = new vector<TpVal>;
*ns = *buf.grid;
delete buf.grid;
buf.grid = ns;
}
line '*ns = *buf.grid;' reported as memory leak and I wondering this part of code means what. Can you explain this part of code?
thank you
|
|
Повідомлення створено: 07. 03. 2017 [10:09]
|
|
roman
Roman Savochenko
Moderator Contributor Developer
Зареєстрован(а) с: 12.12.2007
Повідомлення: 3750
|
"amirkhan63" wrote:
line '*ns = *buf.grid;' reported as memory leak and I wondering this part of code means what. Can you explain this part of code?
thank you
There is no memory leakage and the code designed to drop spare (reserved by vector on push in twice size of the array on the push time) items into buf.grid by its recreation and copy the presented-real items:
Memory consumption is made optimal for limited size
The same line '*ns = *buf.grid;' only performs the items copy by the standard copy function of the STL Vector, which isn't OpenSCADA code at all!
And sure I have not any real leakages on my projects here!
P.S. If you about a reporter mean "valgrind" then it's wrong often for the STL containers.
Learn, learn and learn better than work, work and work.
|
