Errors, bugs, questions - page 2426

[Alexey Viktorov]

Alexandr Gavrilin:

What the hell is wrong with the double type in the latest versions of metaeditor? In both mt4 and mt5 it writes some nonsense instead of -0.01. Where did these numbers come from, the variables are reset to zero in my case.

I have to change the memory. It was the same in older versions.

[Igor Zakharov]

the compiler behaves differently in the same cases.

Downloaded 2 files from kodobase:https://www.mql5.com/en/code/15440

were made, obviously before iTime() appeared. One of them compiles, the other one swears:

Functions are identical:

calls too.

The reason for the scolding is obvious, understandable and easily fixed by auto-replacement, but the different behaviour in similar situations is alarming.

PS:

The compiled indicator works correctly.

[Alexandr Gavrilin]

Alexey Viktorov:

You have to replace the memory. It was exactly the same in the old versions.

What memory? What's memory got to do with it?

[pivomoe]

I ran the following code :

 Temp.ask=0; Temp.bid=0; Temp.time=0;
 Print("Старт Ask=",Temp.ask," Bid=",Temp.bid," Time=",Temp.time);
 Print("Ответ SymbolInfoTick=", SymbolInfoTick("",Temp) );
 Print("Ask=",Temp.ask," Bid=",Temp.bid," Time=",Temp.time);

And got a strange result:

Старт Ask=0.0 Bid=0.0 Time=1970.01.01 00:00:00
Ответ SymbolInfoTick=true
Ask=69.75 Bid=69.33 Time=2019.04.08 21:21:12

I.e. SymbolInfoTick with the first parameter equal to "" for some reason returns the last tick by the chart symbol instead of false and a tick from zeros. Although it should not be so according to definition in documentation.

bool  SymbolInfoTick( 
   string    symbol,     // символ 
   MqlTick&  tick        // ссылка на структуру 
   );

[Andrei Trukhanovich]

Igor Zakharov:

property strict?

[Vitaly Muzichenko]

Alexandr Gavrilin:

What the hell is going on with the double type in the latest versions of metaeditor? In both mt4 and mt5 it writes some nonsense instead of -0.01. Where do these numbers come from? I set the variables to zero.

This is the build from 2007.

It has always been 0.01, but now I don't understand it:

[Igor Zakharov]

Andrei Trukhanovich:

property strict?

in both are the same:

I haven't checked up to the symbol (as pointed out above - the code isn't mine), but on a cursory glance they are 95% identical. And the difference is not in the places where the error pops up. But, for some reason one compiles, the other doesn't. I have already fixed it for myself, but I am informing the developers about the strangeness.

[Marat Sultanov]

This bug seems to have appeared in recent builds of MT5.

Bug in CryptEncode() and CryptDecode() functions. Problem with the CRYPT_AES256 conversion method

You may reproduce the bug in the latest (2009) build with this script:

//+------------------------------------------------------------------+
//| Script program start function                                    |
//+------------------------------------------------------------------+
void OnStart()
{
   string Txt = "The quick brown fox jumps over the lazy dog. A0 A1 A2 A3 A4 A5 A6 A7 A8 A9"; 

//---Key
   uchar Key[32];
   ArrayInitialize(Key, 3);

//---encode
   uchar SourceArr[], EncodedArr[];
   
   StringToCharArray(Txt, SourceArr); 
   
   int EncodeRes = CryptEncode(CRYPT_AES256, SourceArr, Key, EncodedArr);
   
//---decode
   uchar DecodedArr[];
   
   int DecodeRes = CryptDecode(CRYPT_AES256, EncodedArr, Key, DecodedArr);
   
   string DecodedTxt = CharArrayToString(DecodedArr);

//---res
   Print("Txt: ",Txt);
   Print("EncodeRes: ",EncodeRes);
   
   Print("DecodedTxt: ",DecodedTxt);
   Print("DecodeRes: ",DecodeRes);
   
//---
   return;
}

[Artyom Trishkin]

Marat Sultanov:

This bug seems to have appeared in recent builds of MT5.

The bug is in CryptEncode() and CryptDecode() functions. Problem with the CRYPT_AES256 conversion method

You may reproduce the bug in the last (2009) build with this script:

Forum on trading, automated trading systems & strategy testing

CryptEncode in MT5 2009 started giving a different result

Ilyas, 2019.04.08 10:28

Thanks for the post.

AES encryption error has been fixed.

The fix will most likely only appear when upgrading to a new build.

[Slava]

Vitaly Muzichenko:

Here in the build since 2007

It was always 0.01 and now it's not clear:

Show output code