MetaTrader 4 Client Terminal build 604 - page 4

 
Ovo:

I mean those related to the BetaTrader.
... I see that the pun is spreading round the internet. First on Forex Factory, then on Forex Magnates, now here...
 
gchrmt4:
... I see that the pun is spreading round the internet. First on Forex Factory, then on Forex Magnates, now here...


Yes, apparently it is not my invention :)
 
Ovo:

Yes, apparently it is not my invention :)
Big <g> if you also came up with it independently!
 
gchrmt4:
Big <g> if you also came up with it independently!


So you may keep the g small then.
 

A bug in the new Custom Optimizer:

In backtest, OnTester() gives the correct result. But when optimizes always give 0. (attach image)

---------------

I think there is also a bug in the dll calls. In build 509 call urlmon.dll without problems, in build 600-604 seems to not work.

Regards.

In tester work well, in optimizer always 0

 
Ovo:
So you may keep the g small then.
Not my g. I wish I had thought of it, but I didn't. I've just seen it in a spreading number of places.
 
positiveday:

Hello, there is a bug still present in v 604 MetaEditor.

Keys for functions : "Function header" - "Block comment" - "Comment lines" - "Uncomment lines" are not for localized keyboard, please check localization scancodes.


This is instead a suggestion.
The dropdown list for List functions and methods in files, is useful for small prorams codes, but for large codes it's not very immediate to use.
Any time I open it starts from A letter, instead it should remember the last clicked item.

Apart from this could be better to create a classical navigator pane.
I dont' know why there is a navigator for files and there is not a navigator for function/methods, it seems that a programmer pass his time to open different files instead to jump from a function to another.

This is what I see in my dropdown list of functions (please note that list start from "A" and ends to "G") :



This is the same source from a classic editor, Geany, note the function navigator on right :



It's obvious that for a ten functions program there are not differencies, but not all developer on MT4 works only on Sunday or after their daytime work.


I completely agree, these points should be fixed as quick as possible for they are pesky and annoying.

 
Fer137:

A bug in the new Custom Optimizer: In backtest, OnTester() gives the correct result. But when optimizes always give 0. (attach image)

...

I think there is also a bug in the dll calls. In build 509 call urlmon.dll without problems, in build 600-604 seems to not work.


Ah, I see now it had already mentioned:

" VabRus: ... double OnTester() --> you can´t optimizer by custom, and this does not make much sense. ... For external calls, windows api, you can only work with unicode strings, ... "

 

StringToCharArray does not correctly convert Unicode string to ANSI string.

code:

#property strict

#import "kernel32.dll"
    void OutputDebugStringA(char& str[]);
    void OutputDebugStringW(string str);
#import

#import "UnicodeString.dll"
    int getAnsiStringLength(string unicodeString);
    bool convertUnicodeToAnsi(string unicodeString, char& ansiString[], int ansiLength);
#import

void OnStart()
{
    // Russian string
    string unicodeString = "Здравствуйте, меня зовут Ольга. Я живу во Владивостоке.";

    // Method 1: Convert Unicode string to ANSI string by StringToCharArray
    uchar ansiString1[];
    StringToCharArray(unicodeString, ansiString1);

    OutputDebugStringW("***** ansiString1:");
    OutputDebugStringA(ansiString1);

    // Method 2: Convert Unicode string to ANSI string by WideCharToMultiByte in DLL
    int ansiLength2 = getAnsiStringLength(unicodeString);
    uchar ansiString2[];
    ArrayResize(ansiString2, ansiLength2);
    convertUnicodeToAnsi(unicodeString, ansiString2, ansiLength2);

    OutputDebugStringW("***** ansiString2:");
    OutputDebugStringA(ansiString2);
}

result in DebugView:


Used DLL is attached.


From the above, ansiString2 is perfect converted by twice call of WideCharToMultiByte.

StringToCharArray may be buggy.

Files:
 

I do not yet have b604. Can anyone confirm whether the following things have been resolved within it?

"true/false gate .set file issues.

Behaviour - Load in a previously saved .set file and the true/false gates are messed up. Some have switched round, others have 1's or 0's displayed in their boxes.

The bug appears to occur when selecting .set files that were saved in the tester and using them to set up live bots. The opposite is also true, .set files saved from the live bot config on a chart and then used in the tester show the same issues."

and...

"It has been previously reported that parameter inputs based on enumerated types are displayed in the reverse order in the drop-down list in the parameter dialog."

.

Also, have the init bugs previously described been fixed?

Reason: