Hi Jeff,
I am restarting developement of my program for costing for shipstructures using the latest version of SF.
I did not use it for a longer time and found an unwanted reaction on hitting the windows cross on top of dioalogs: It causes a crash or it does not return to the calling window. All my dialogs  have buttons to exit and return to the calling window. The cross at top is not needed. Is it possible te deactivate it?
Regards Klaus

There is no way to deactivate the window's close button.  However, recent versions of AppGraphics include a routine setwindowclosecallback that will trigger a subroutine call when a window is closed, regardless of how it occurs.  If you set up a callback routine properly, it should be able to do the same thing gracefully4 that your exit buttons do.

When you specify a window size to initwindow, it actually is specifying the window's client area where drawing can take place, and it does not include the decorations.

I have a question regarding the screen resolution in pixels using APPGRAPHICS. When a window is created of width and height of 800,600 using initwindow(800, 600, title="Example Application"), the functions getwindowwidth ( ) and getwindowheight ( ) produce pixels of 806 and 629. In comparision, the functions getmaxx ( ) and getmaxy ( ) produce 799 and 599 as they should.
Why do these function calls produce different window dimension?

Moreover, when subsequently initwindow (getwindowwidth ( ), getwindowheight ( ), "EXAMPLE WINDOW", DEFAULT_POSITION , DEFAULT_POSITION , .FALSE., .TRUE.) is called, then functions getwindowwidth ( ) and getwindowheight ( ) produce pixels of 812 and 658. 
Why has the window dimensions gotten larger?

EXAMPLE CODE:
    INTEGER :: window, initwin

        !--- Display screen in pixel coordinates
        !--- Create a window 800 pixels wide by 600 pixels tall.
        ! (width,height)
        ! (0,0).....................................(800,0)
        !  .
        !  .
        !  .
        ! (0,600)...................................(800,600)

    !--- Initialized window x,y dimensions (800,600)
        window = initwindow(800, 600, title="Example Application")

        !-------------------------------------
        !INITWIN = initwindow (getwindowwidth ( ), getwindowheight ( ), &
        !"EXAMPLE WINDOW", DEFAULT_POSITION , DEFAULT_POSITION , &
        !.FALSE., .TRUE.)

        !--- Width in pixels of the current window
        !PRINT *, 'INITWIN-XTOTAL_PIX = ', getwindowwidth ( )    !--> 812
        !!--- Height in pixels of the current window
        !PRINT *, 'INITWIN-YTOTAL_PIX = ', getwindowheight ( )   !--> 658
        !-------------------------------------

        !--- Width in pixels of the current window
        PRINT *, 'XTOTAL_PIX = ', getwindowwidth ( )    !--> 806

        !--- Height in pixels of the current window
        PRINT *, 'YTOTAL_PIX = ', getwindowheight ( )   !--> 629

        !--- Largest horizontal position possible in the current window
        PRINT *, 'XMAX_PIX = ', getmaxx ( )    !--> 799
   
    !--- Largest vertical position possible in the current window
        PRINT *, 'YMAX_PIX = ', getmaxy ( )    !--> 599

btw, I running Windows 10 using a Microsoft Surface Pro PC. SimplyFortran is verison 3.19, build 3543 using GNU Fortan 10.2.0.

Frank

So calling the subroutine test twice is a bug.  We've fixed it, and it should be included in Simply Fortran 3.15, which we're finishing up now.

    i2=createtextbox(200,200,50,25)!文本框
    call settextboxcontents(i2, "2")!初始值
   
    i3=createscrollbar (155, 185, 40, 56, SCROLL_VERTICAL, test) !垂直滚动条
    call setscrollrange(i3,1,3) !设置垂直滚动条档位范围
    call setscrollposition(i3,1) !设置垂直滚动条当前档位
   
    i4=createtextbox (155, 200, 40, 25)!文本框
    call settextboxcontents(i4, "+")!初始值
   
    i5=createtextbox(250, 200,50,25)!文本框
    call settextboxcontents(i5, "=")!初始值
    i6=createtextbox(300, 200,100,25)!文本框
    call settextboxcontents(i6, "3")!初始值
   
    i = createbutton(450, 200, 50, 25, "Run", start)!按钮
    i = createbutton(600, 200, 60, 25, "Close", quit)!按钮
   
    call loop()
    call closewindow(myscreen) !关闭窗口
contains
    subroutine Quit()
        call stopidle()
    end subroutine
   
    subroutine start()
        integer::i,j,k,m=0
        character(80)::s1,s2,s3
       
        i=gettextboxcontents(i1,s1) !i返回字符串长度 s1返回文本字符串
        j=gettextboxcontents(i2,s2)
        read(s1,*)i !字符串转换为整数
        read(s2,*)j
       
        k=getscrollposition (i3)!返回滚动条当前位置

        select case(k)
        case(1)
            m=i+j
        case(2)
            m=i-j
        case(3)
            m=i*j
        end select
       
        write(s3,"(i0)")m !整数转换为字符串
        call settextboxcontents (i4, s)!设置文本框内容
        call settextboxcontents (i6, trim(adjustl(s3)))!设置文本框内容
    end subroutine
   
    subroutine test(j) !j返回滚动条位置
        integer::j

        write(*,*)j
        select case(j)
        case(1)
            s="+"
        case(2)
            s="-"
        case(3)
            s="*"
        end select
       
        call settextboxcontents (i4, s)!设置文本框内容
    End subroutine
end program main

In the subroutine test, why "write(*,*)j" run twice in the console when i change the scrollbar?

I'm not exactly sure I understand the problem.  After you call enablewindow(initW,.false.), what seems to be the problem?  You're creating a new window after that which should get all user input, but some is going to the background window?

I think it might be safer to reverse the two calls:

initW = getcurrentwindow()  ! keep the main window in memory for return to it
dlg_screen  = initwindow(660, 900, title = "Structural parts defined on cross sections", closeflag=.false.)
Call enablewindow(initW,.false.)

Disabling a window when there is no foreground window doesn't quite look right, though I might be wrong.  I'll try some experiments on my end too.

Whats wrong with it?
Klaus

The problem is related to setcurrentwindow.   When you're calling that subroutine, the description (and subroutine name, quite frankly) sound as if it should also bring that window to the foreground.  In reality, all the call does is set the internal indicator within the library to point at that window for any subsequent graphics calls.  The library should also raise the indicated window to the top when setcurrentwindow is called.

I'll make the change and release a new build with the fix within the week.

Let me work on that this weekend to see if I can both replicate the problem and come up with a solution.

How can I try to tackle the problem?

The dialog in the second window (initW) starts as follows:

initW = getcurrentwindow()  ! keep the previous window in memory for later return
Call enablewindow(initW,.false.)
! ______________________ open dlg-window__________________________________
dlg_screen = initwindow(250, 360, title = "Secundary Stiffening", closeflag = .false.)
CAll setdialoghotkeys(.TRUE.)
…..........
….........
And it ends with:
call loop()
   Call enablewindow(initW,.true.)
   Call closewindow(CURRENT_WINDOW)
   Call setcurrentwindow (initW)
!____________________________________________________________________________
Contains
.....
May be this info is too short to give a hint and I have to make a tiny testprogram first?

Regards, Klaus

Yes, that's what's necessary in Windows to construct multiline messages for use in message boxes.  The Conway's Game of Life example included with Simply Fortran does the same thing in the subroutine aboutgame in control.f90.

You could consider using the constant C_CARRIAGE_RETURN available in the iso_c_binding module as well to be more explicit rather than using CHAR(13).  I'm not sure there's any real advantage, though.