QByteArray problem

[Zizione]

Hi guys. I am new in QT. I started developing a programm which working with bytes, and when I tried to use a QbyteArray, I've got some problem. There is a line in QByteArray class reference :
QByteArray ba;
ba.resize(5);
ba[0] = 0x3c;
ba[1] = 0xb8;
ba[2] = 0x64;
ba[3] = 0x18;
ba[4] = 0xca;
QByteArray must store 1 byte in each cell of it, but when i trying to do the same it gives me a warning:
QByteArray ba;
ba.resize(4);
ba[0] = 0xFF;
Result of the output ba[0] is symbol "?". It means that it stored not 8 bit but 6, symbol "?" is 63 in ascii table and 00111111 in binary code. Warning is C4309. How can i store all 8 bits in QByteArray ? (P.S. sorry for my awful English)

[aha_1980]

Hi @Zizione:

QByteArray ba;
ba.resize(4);
ba[0] = 0xFF;
Result of the output ba[0] is symbol "?".

How did you output it? Your example works perfectly fine for me, I see 0xFF in the debugger.

It means that it stored not 8 bit but 6, symbol "?" is 63 in ascii table and 00111111 in binary code.

I rather think, your output cannot display the character for 0xFF and prints a '?' instead.

Warning is C4309.

I don't know what the text for this warning is, but I guess it's because QByteArray contains chars and char is signed for most compilers, so it has a range of -128...127. 0xFF is outside this range, hence the warning. You can help yourself with a cast: ba[0] = char(0xFF); or with a literal: ba[0] = '\xFF'

How can i store all 8 bits in QByteArray ?

You already do, it's just a matter of output.

[Zizione]

@aha_1980
There is my full code , this is just test unit. I need exactly int value in it , and how i understand '\xFF' will store char in it ?

#include <QCoreApplication>
#include <QByteArray>
#include <QString>
#include <QTextStream>
int main(int argc, char *argv[])
{
    QTextStream cin(stdin); cin.setCodec("CP866");
    QTextStream cout(stdout); cout.setCodec("CP866");
    QCoreApplication a(argc, argv);
    QByteArray ba;
    ba.resize(4);
    ba[0] = 0xFF;
    ba[1] = 0xFF;
    ba[2] = 0xF;
    ba[3] = 0xF;
    char tdi[16];
    int b=0;
    for(int i=0;i<4;i++)
    {
        uint16_t mask = 0x8;
        for (int j = 0; j < 4; j++)
        {
            tdi[b] = (ba[i] & mask) ? '1' : '0';
            mask >>= 1;
            b++;
        }

    }
    for(int i=0;i<16;i++)
    {
        cout<<tdi[i]<<endl;
    }
    cout<<tdi<<endl;
    cout<<ba[0]<<endl;
    //cout<<sizeof(tdi)<<endl;
    return a.exec();
}

[VRonin]

welcome to the forum and you are jumping right into a controversial point of the API (see https://forum.qt.io/topic/89766/converting-qbytearray-to-unsigned-char/9). Basically QByteArray interface uses char while you are trying to use unsigned char and hence the warning.

You can either:

• ignore the warning (QByteArray is actually storing the correct value it is just casting it to signed)
• use manual casting: *reinterpret_cast<unsigned char*>(ba.data()[0])=0xFF;
• use memcpy: const unsigned char tempVal = 0xFF; std::memcpy(ba.data()+0,&tempVal ,1);

[Zizione]

@VRonin So it is ok with storing 0xFF value , if i need this value in my code from QByteArray it would be given ?

[VRonin]

Yes, the if you use [] that casts it automatiocally to signed but the value is still 0xff

[JonB]

@Zizione
In addition to the solutions offered by @VRonin....

I am not a C++-er, but I believe in your case your can use the U suffix to make your constant values unsigned int and thus convertible to unsigned char:

ba[0] = 0xFFU;

?

[J.Hilk]

@JonB
Wow, that totally slipped my mind, 👍

[aha_1980]

@JonB said in QByteArray problem:

@Zizione
In addition to the solutions offered by @VRonin....

I am not a C++-er, but I believe in your case your can use the U suffix to make your constant values unsigned int and thus convertible to unsigned char:

ba[0] = 0xFFU;

?

Yeah, but ba is signed char...

[J.Hilk]

@aha_1980
true, but the warning one gets with ba[0] = 0xFF is not about usigned and signed char but rather between int and char

0xFFU == unsigned char(0xFF) == uchar(0xFF) != 0xFF

the implicit conversation to ba[0] does not "drop" information

[aha_1980]

@J.Hilk have you tried?

    QByteArray ba;
    ba.resize(4);
    ba[0] = char(0xFF); // no warning
    ba[1] = 0xFFU; // warning: implicit conversion changes signedness: 'unsigned int' to 'char'
    ba[2] = 0xFF; // warning: implicit conversion changes signedness: 'int' to 'char'

Edit: screenshot

[J.Hilk]

@aha_1980 said in QByteArray problem:

@J.Hilk have you tried?

I always do, well 95%, before I post ;-)

to be honest, I'm surprised about the uint part.

[aha_1980]

@J.Hilk

then you can be surprised about the uchar() part too.

But the biggest surprise: your code will not even compile in gcc and clang because of line 65.

And one more hint: your code will crash as you access invalid index in ba ;)

[J.Hilk]

@aha_1980 said in QByteArray problem:

@J.Hilk
But the biggest surprise: your code will not even compile in gcc and clang because of line 65.

I saw you using linux in the screen shot and testet that there than, and noticed it too

And one more hint: your code will crash as you access invalid index in ba ;)

nope doesn't crash, ignore the project name its my test anything project x).
!

resize is apparently not a necessity?

Edit:
also ignore value 0, I commeted that line out, to run it with gcc, so theres a random value.
// ba[0] = unsigned char(0xFF);

[aha_1980]

@J.Hilk said in QByteArray problem:

resize is apparently not a necessity?

Ah, my fault: "If an assignment is made beyond the end of the byte array, the array is extended with resize() before the assignment takes place."

A big difference to QVector or QList here.