Is there a bug for the dislin library of macOS native？

mark jack

unread,

Apr 11, 2025, 9:06:15 AM4/11/25

to dislin-users

Dear developer,

When I test macOS native dislin library (https://www.dislin.de/macos.html, intel and arm64 has same result), it seems has issue (compiled by gfortran-14 and 13 has same result), my test code:

program dislin_3d_plot
use dislin
implicit none

integer, parameter :: N = 50
integer :: id_draw, id_but1, id_but2, id_lab2
double precision :: xrot1 = 50.0d0, xrot2 = 40.0d0
double precision :: xray(N), yray(N), zmat(N,N)
integer :: ip, ip1, ip2, i, j, ival
double precision :: step, x, y
double precision, parameter :: fpi = 3.14159265358979323846d0 / 180.0d0

! Initialize data arrays
step = 360.0d0 / dble(N-1)
do i = 1, N
x = dble(i-1)*step
xray(i) = x
do j = 1, N
y = dble(j-1)*step
yray(j) = y
zmat(i,j) = 2.0d0 * sin(x*fpi) * sin(y*fpi)
end do
end do

! Create GUI widgets
call swgtit('DISLIN 3D Plot')
call swgopt('track', 'scroll')
call swgopt('center', 'position')
CALL SWGOPT("TRACK","SCROLL")

call swgwth(70)
call wgini('hori', ip)
call wgbas(ip, 'vert', ip1)
call swgdrw(2100D0/2970D0)
call wglab(ip1, 'DISLIN 3D Plot:', id_lab2)
call wgdraw(ip1, id_draw)

call swgwth(20)
call wgbas(ip, 'form', ip2)
call swgwin(0, 0, 200, 100)
call wgpbut(ip2, 'Azimuthal',id_but1)
call swgwin(0, 120, 200, 100)
call wgpbut(ip2, 'Polar', id_but2)
call swgcbk(id_but1, myplot)
call swgcbk(id_but2, myplot)
! CALL REAWGT
call myplot(id_but1)
call wgfin()

contains
subroutine myplot(id)
integer, intent(in) :: id
integer :: lev

if (id == id_but1) then
xrot1 = xrot1 + 10
else if (id == id_but2) then
xrot2 = xrot2 + 10
end if

CALL GETLEV(lev)
write(*, "('Current level: ', I4)") lev
if (lev /= 0) then
write(*, *) "*******************Level ERROR*******************"
end if

call metafl('cons')
call setxid(id_draw, 'widget')
call scrmod('revers')
CALL PAGE(3000,3000)
CALL IMGFMT("RGB")

write(*,*) "**************call disini()************"
call disini()
call erase()
call hwfont()

call name('X-axis', 'x')
call name('Y-axis', 'y')
call name('Z-axis', 'z')
call labdig(-1, 'xyz')

call axspos(700, 2800)
call axslen(2100, 2100)

call view3d(xrot1, xrot2, 6.0D0, 'angle')
call height(40)
call graf3d(0.0D0, 360D0, 0D0, 90D0, 0D0, 360D0, 0D0, 90D0, -3D0, 3D0, -3D0, 1D0)
call surshd(xray, N, yray, N, zmat)
call disfin()
write(*,*) "**************call disfin()************"

end subroutine myplot

end program dislin_3d_plot

The program maybe crash when frequently click two push button, console message:

Current level: 0
**************call disini()************

<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
<< END OF DISLIN / VERSION 11.5.2 <<
<< Date : 11.04.2025 Time : 12:23:24 Pageformat: PAGE <<
<< Vectors : 1081 Warnings: 0 Fileformat: CONS <<
<< Metafile: <<
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

**************call disfin()************

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:

Could not print backtrace: executable file is not an executable

If my program more complex, the report `Current level: ` maybe > 0, I suspect if there's something wrong with the `disini` function.

Sincerely,

Mark

Helmut Michels

unread,

Apr 14, 2025, 5:05:06 AM4/14/25

to dislin-users

Dear Mark,

thank you for the example code. I will make some tests with it.

With best regards,

Helmut

Helmut Michels

unread,

Apr 15, 2025, 7:03:45 AM4/15/25

to dislin-users

Hi Mark,

I have now found a bug in the native macOS code of Dislin for Fortran, which has caused the

segmentation fault I think. Your example code is working fine now for me after the correction.

Updated Dislin distributions for macOS are available from https://www.dislin.de/macos.html