Errors, bugs, questions - page 1270

 
A100:

Compilation error

I tried the C++ compiler - it understands that (where MQL has an error) it needs to call f() with Print("2" ), while MQL has an uncertainty.

A simpler variant also compiles with an error

Thank you for your feedback, we are working on it.
 
Can you tell me how to remove a site from the WebRequest's list of allowed sites?
 
avoitenko:
Can you tell me how to remove a site from the list of allowed sites for WebRequest?
"Service" -> "Settings" -> "Advisors".
 
avoitenko:
Can you tell me how to remove a site from WebRequest's list of allowed sites?
the "delete" button on the address line and "ok"
 
barabashkakvn:
"Tools" -> "Settings" -> "Advisors".
there is no context menu by right-clicking on the address bar. it took me a long time until i just pressed "delete".
 
sanyooooook:
the "delete" button on the address line and "ok"
Thank you. The 'delete' button has helped. But it's so not obvious!
 

There is a problem in the Expert Advisor optimizer. When all passes are completed by agents, the Stop button immediately changes to Start, i.e. optimization is already completed. But the results can still be processed in OnTesterPass or OnTesterDeinit.It turns out that we cannot stop the process. If you press Start, a new optimization is launched simultaneously with the old one and that can lead to some unpleasant consequences.

We want the button to change the value only after OnTesterDeinit is processed

 
meat:

There is a problem in the Expert Advisor optimizer. When all passes are completed by agents, the Stop button immediately changes to Start, i.e. optimization is already completed. But the results can still be processed in OnTesterPass or OnTesterDeinit.It turns out that we cannot stop the process. If you press Start, a new optimization is launched simultaneously with the old one and that can lead to some unpleasant consequences.

We want the button to change its value only after OnTesterDeinit is processed.

And what if, in case of a programming error or the like, OnTesterPass\OnTesterDeinit hangs/locks?

What prevents you from deleting the EA in the chart or closing the chart itself with the EA before restarting it?

The Stop button changes to Start, as the optimisation is indeed completed and the participation of agents is not required. Further parsing is up to the developer.

 
alexvd:

And if in case of programming error or other, OnTesterPass\OnTesterDeinit hangs/locks execution?

I don't see a problem. Optimisation then won't complete until the user presses the button. What's strange about it? And now when you press the button, another optimization is launched simultaneously with the halted one, exhausting system resources. The user may not even notice it until the system starts to slow down. Is that normal?

The Stop button changes to Start, as the optimization is indeed complete and no agents are needed. Further parsing is up to the developer.

If the optimization process was limited to agents' work, we would rather call it brute force and not optimization. But it's not only brute force that's important for us, it's an integral part of optimization, especially when we're talking about optimization of mathematical calculations.

As for manually deleting the chart before restarting - yes, of course it can be done, but it is a shamanistic option. And it is not always possible to understand which graph the optimization refers to, if there are many such graphs open, and we switched the current graph. It is much easier to use the button.

Распределенные вычисления в сети MQL5 Cloud Network
Распределенные вычисления в сети MQL5 Cloud Network
  • cloud.mql5.com
Заработать деньги, продавая мощности своего компьютера для сети распределенных вычислений MQL5 Cloud Network
 
meat:

//

If the optimization process was limited to the work of agents, it would rather be called a search and not optimization. But for us it's not just a search of parametersbut the subsequent processing of the obtained results that's an integral part of optimization. Especially when it comes to optimization of mathematical calculations.

//

Generally, there is no optimization in the tester. As a matter of fact, there is a reduction of parameter variant space (reduction of number of runs in comparison with complete enumeration) in accordance with set criteria.

Tester doesn't provide any means of optimization(identification of optimal parameters). Selection, analysis, grouping of results and subsequent selection of the best one (if it were, then one could call what the "optimizer" does as optimization) is something that the user takes over.

Reason: