Errors, bugs, questions - page 2566

[[Deleted]]

Slava:

With this casting, there is no data loss. Either it is 0 or it is not 0.

Another case is when double -> any integer type (up to and including int32)

You're playing with warnings a bit too much in my opinion, there is a lack of consistency in your judgement

bool continuation()const {return this.last_level;}  // uint last_level
// expression not boolean	lrp_last_9.mq5	260	42

[A100]

Vict:

You've gone a bit overboard with the warnings in my opinion, there is a lack of consistency in judgement

Agreed

Here too,there is no data loss in this sense(either 0 or not 0)

void OnStart()
{
        int i = -1;
        while ( ++i ); //Warning: expression not boolean
}

but there is a warning. A uniform approach was expected

[Aliaksandr Hryshyn]

Result of an attempt to access a property of a class instance from a static function of the same class:

2019.09.18 20:07:41.043 Test_exec (EURUSD,M5)   Access violation at 0x000001E9CCD2963C read to 0x0000000000000014 in 'D:\Alpari MT5\MQL5\Scripts\Test_exec.ex5'
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)      crash -->  000001E9CCD2963C 837B1400          cmp        dword [rbx+0x14], 0x0
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)                 000001E9CCD29640 0F8E8F020000      jle        dword 0x1e9ccd298d5
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)   
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)                 000001E9CCD29646 90                nop        
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)   
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)   00: 0x000001E9CCD2963C
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)   01: 0x000000B56CA2F120
2019.09.18 20:07:41.080 Test_exec (EURUSD,M5)   02: 0x000001E9CCD2A04D

The compiler was silent before this.

[Roman]

Does not initialise an empty string with a terminal null.

void OnStart()
{      
   string str;
   StringInit(str, 100, 0);
   
   Print(StringLen(str));     
}

Result

2019.09.18 22:16:54.380 TestScript (EURUSD,H1)  0

Expected

2019.09.18 22:16:54.380 TestScript (EURUSD,H1)  100

Note in the help for the StringInit function

Примечание

Если  character=0 и размер new_len>0, то будет распределен буфер строки указанного размера и заполнен нулями. 
Размер строки будет равен нулю, так как весь буфер заполнен терминаторами строки.

In fact, there is no distribution.
I.e. allocation in this way, ten spaces

string str = "          ";

is not the same as.

string str;
StringInit(str, 10, 0);

The help text marked in red does not correspond to logical behavior.

[Igor Makanu]

Vict:

You're overdoing it with the warnings in my opinion, not enough consistency in judgement

Warnings don't work on boolean operations, now I accidentally noticed a typo in my code, so I reproduced it:

#define               getAsk(dummy)          SymbolInfoDouble(_Symbol, SYMBOL_ASK)
#define               getBid(dummy)          SymbolInfoDouble(_Symbol, SYMBOL_BID)
//+------------------------------------------------------------------+
void OnStart()
  {
      double sl = 0.0,tp=0.0;
      if(sl != ! OrderStopLoss() || tp != OrderTakeProfit()) Print("");  
  }
//+------------------------------------------------------------------+

double OrderStopLoss()
{
   return(getAsk());
}

double OrderTakeProfit()
{
   return(getBid());
}

[Stanislav Korotky]

Roman:

Does not initialise an empty string with a terminal null.

Result

Expected

Note in the help for the StringInit function

In fact there is no distribution.
I.e. allocation in this way, ten spaces

is not the same as.

And it doesn't seem to match the logic of the behaviour, highlighted in red in the help text.

You missed the StringBufferLen function - it returns the size of the allocated buffer. StringLen returns the string length, i.e. up to the terminal character, respectively, if it is at the very beginning - length 0. Buffer != string.

[A100]

Stanislav Korotky:

You have overlooked the StringBufferLen function - it returns the size of the allocated buffer. StringLen returns the length of the string, i.e. up to the terminal character, respectively, if it is at the very beginning - length 0. Buffer != string.

Otherwise it doesn't know itself.

[Roman]

Stanislav Korotky:

You have overlooked the StringBufferLen function - it returns the size of the allocated buffer.
StringLen returns the length of the string, i.e. up to the terminal character, respectively, if it is at the very beginning - length 0. Buffer != string.

This is clear, we're talking about initializing a string with terminal zeros.
Why not fill stringInit(str, 10, 0); with ten terminal zeroes ? And return the actual string length.
If you want to initialize an empty string, without filling it! for example with 100 characters,
then either 100 spaces as str = " many pokes at the keyboard ", or initialize rubbish StringInit(str, 100, 65)
I don't understand why I need to initialize with rubbish, there's already enough of it ))

[[Deleted]]

Roman:

It's understandable, we're talking about initializing the string with terminal zeros.
Why not fill in StringInit(str, 10, 0); with ten terminal zeros ? And return the actual string length.
If you want to initialize an empty string, without filling it! for example with 100 characters,
then either 100 spaces as str = " many pokes at the keyboard ", or initialize rubbish StringInit(str, 100, 65)
I don't understand why I need to initialize it with rubbish, there's already enough of it ))

What makes you think it's not filled in? It's just that the length in the µl string is not stored, it's recognised by the terminal zero.

StringInit(str, 100, ' ');

no?

And if you need zeros, there is a String class somewhere in the bowels of µl std.

[Roman]

Vict:

What makes you think it's not filled in? It's just that the length in the µl string is not stored, it's recognised by the terminal zero.

isn't it?

And if you need zeros, there is a String class somewhere in the bowels of the µl std.

Nah... It returns zero that way too. And third parameter ushort, requires integer character code.

If I initialize as str = " "; and then pass data from dll, everything is ok.
But if I initialize it as StringInit(str, 10, 0) then data won't arrive. That's because the code does not allocate memory for string with needed size.

If I had code for space, probably would have worked, but that I have not found such code in different tables.
We need to allocate memory for number of future characters which then will go to already initialized empty string.
But rubbish to initialize the string, not kommelpho.
StringInit(str, 100, 0); should be filled with terminal nulls and return the length 100.