Errors, bugs, questions - page 1968

New comment
 

Why this silly and useless warning "no #import declaration" if you write it in the header file:

void f();
void f() {}

There is no limit to the number of declarations in decent places. Annoying, I want to have a brief description of available functions in the mqh header, I have to comment lines, which negatively affects readability. Someone may say, "put static methods in a class or structure (when using structures, another wonderful warning "struct has no members, size assigned to 1 byte" appears)". I'll answer: "I don't like µl classes, I want to use the first option" without this flood of goopy warnings. Why are you forcing me to abandon perfectly valid common practices?

 
pavlick_:

Why this silly and useless "no #import declaration" warning if you write it in the header file:

This is a special case. It makes sense in a more general solution to add an analogue

#pragma  warning (disable:xxxx)
and every experienced programmer could disable the annoying warning (the number can be found in the command compiler). The existing warning system is essentially useless... I don't even look through them because there are hundreds of the same-type ones that do not take into account the programmer's writing style and experience. And among these hundreds of them it is hard to find really important warnings worth paying attention to
  
class A {
public:
        A() { Print( A::a ); } //Результат: 0
        static const int a;
};
/*
...
*/
const int A::a = 1;
Don't believe me? I'll try to add the code tomorrow
 
A100:

This is a special case. It makes sense in a more general solution to add an analogue

and every experienced programmer could disable annoying warnings (the number can be found in the command compiler). The existing warning system is essentially useless... I don't even look through them because there are hundreds of the same kind of warnings that don't take into account the programmer's writing style and experience. And among those hundreds of them it's hard to find really important information worth paying attention to
Yes, I agree. But I think you'd better manage a whole set of warnings at once. A key mechanism (like -Wno_all) or #pragma warning (level:{0|1|2|...}). It's cumbersome to disable one at a time.
 

Dear developers! Please remind, were there any edits about the fact that if the indicator calculation is in a subwindow and the style of several of its buffers is DRAW_NONE, then they do not affect the scale of display in the subwindow? Or there were no such edits?

If we haven't made such changes, please, make them. Because now it turns out that the DRAW_NONE style affects the graphics in the subwindow, which should be at a completely different scale!

 
pavlick_:
Yes, I agree. Only I think it's better to manage a set of warnings at once. Key mechanism (like -Wno_all) or #pragma warning (level:{0|1|2|...}). It's cumbersome to disable one at a time.

I've probably been asking for years for an example for every warning and error that would clearly show why they're coming out.

In this light, tiering and explicit management of warrants looks like a fantasy.

 
Alexey Kozitsyn:

Dear developers! Please remind, were there any edits about the fact that if the indicator calculation is in a subwindow and the style of several of its buffers is DRAW_NONE, then they do not affect the scale of display in the subwindow? Or there were no such edits?

If you did not make such changes, please, make them. Otherwise now it turns out that the DRAW_NONE style affects the graphics in the subwindow, which should be at a completely different scale!

This has been fixed. I wrote a ticket and then checked...

Addendum: it turns out there were even 2 tickets. In MT4 it was fixed, but in MT5 it seems not.

 
Stanislav Korotky:

This has been fixed. I wrote a ticket and then checked.

That's what I thought was fixed, but now it's not. I just checked. Bild 1643.
 
Alexey Kozitsyn:
Here I thought it was fixed, but now it's not. Just checked. Bild 1643.

See above, I finished ;-)

 
Stanislav Korotky:

See above, I finished ;-)

Yeah I see... then I'll create another application.
1...196119621963196419651966196719681969197019711972197319741975...3190
New comment