Errors, bugs, questions - page 534

[Sergey Gritsay]

gdtt:

will not

Exactly, every time you integrate the variable, it can store any value you want. so you need to initialize the variable explicitly

[Andrey Sharov]

I observe the following situation: on different machines on one and the same account with the same testing settings the Expert Advisor gives different results (and on each machine the results of repeated runs are the same)!
There are at least 4 different results:
on my machine in preparation for Championship 2011: ~700.00
automatic test : -3834.32
now : -3514.45
or : -4681.85

The last 2 results were obtained on the same terminal instance (located on a removable disk) on different machines. The EA is multi-currency.

???

[Rashid Umarov]

Ashes:

The last 2 results were obtained on the same instance of the terminal (located on a removable drive) on different machines. This Expert Advisor is multicurrency.

???

Are all testing conditions identical? Wasn't it the case that in one case you tested on EURUSD and in the other on GBPUSD?

It would be good if you wrote to servicedesk and provided an EA with all the details.

[Andrey Sharov]

When connected to one of the computers, the external drive goes into USB 1.1 mode. The MT5 is launched from it. When closing MetaTrader5 disappears from the list of applications, but remains in processes (now more than 9 hours since closing).

[Andrey Sharov]

I take a multi-currency Expert Advisor. Running in tester on championship account from October 3 to the current date. Several first runs have different results. After several runs (2-4) the result stabilizes.

[Andrey Sharov]

2 Rosh:
Everything is identical on the settings tab.

For some reason, when clicking on "reply" it doesn't quote (IE8).

[Ilyas]

sergeev:

there is such a lamer question.

When doing ArrayResize array of objects (pointers), CheckPointer does not return ==POINTER_INVALID and !=POINTER_DYNAMC.
That is, an array of pointers of the normal type.

is this a bug ?

it helps if all new elements=NULL, but new created pointers must return POINTER_INVALID at least
I think it's a bug. Pointers are initially blank... (although not explicitly set to NULL by ArrayResize)

You need to review the code.
If the array initially had size 100 and was filled with valid values, then it was reduced and again increased to 100, it's quite possible that the pointers remained valid.

[Alexey Da]

Ashes:
...

For some reason it's not quoted when you click on "reply" (IE8).

Tried it - it works.

Can you provide a link to the comment you tried to quote.

[Andrey Sharov]

2 alexvd: "Replying" to your message (different machine, same result). Previously (more than a month ago) I reported that "edit" doesn't work properly (I have to copy the original post by hand and paste it into the edit window, otherwise it gets lost). This hasn't been fixed either...

Previously replied to https://www.mql5.com/ru/forum/1111/page539#comment_104666

PS. In FireFox both reply and edit works...

[Stanislav Korotky]

x100intraday:

In general, if you dig into the rather simple code of the fractal indicator, it is symmetrical about the middle (the fractal itself) as for High:

as well as for Low. And, at first glance, the areas highlighted by red rectangles should not seem to occur...

Suspect >/>=.

Everything is displayed according to the code.