unexpected gfortran compiling error [closed] - fortran

Question

Closed. This question is not reproducible or was caused by typos. It is not currently accepting answers.
This question was caused by a typo or a problem that can no longer be reproduced. While similar questions may be on-topic here, this one was resolved in a way less likely to help future readers.
Closed 4 years ago.
Improve this question
Recently, I started to lean fortran programming. I have seen the following code at youtube without any error it is compiled. But I have got some errors.
I appreciate any help
program
implicit none
real, parameter :: pi=4*atan(1.0)
integer, parameter :: n = 100
real :: dimension(1:n) :: x, y
real :: a=0.0, b = 2*pi
real :: increment
integer :: i
increment = (b-a)/(real(n)-1)
x(1)=0.0
do i =2,n
x(i) = x(i-1) + increment
end do
y = sin(x)
print *, x(1:5)
print *, y(1:5)
end program

Answer 1

real :: dimension(1:n) :: x, y is a syntax error. Replace the first :: with a comma. You may need to give a name on the program statement.

Answer 2

You also have mixed-mode arithmetic in line
increment = (b-a)/(real(n)-1)
It will probably compile, and it may not even affect the program, but you should never, never have mixed-mode arithmetic in any programming language as it can cause strange, hard to find bugs.
It should look like this:
increment = (b-a)/(real(n)-1.0)

Answer 3

Here are the results of a working example which addresses the concerns of #High Performance Mark.
host system = (redacted)
compiler version = GCC version 5.1.0
compiler options = -fPIC -mmacosx-version-min=10.9.4 -mtune=core2 -Og -Wall -Wextra -Wconversion -Wpedantic -fmax-errors=5
execution command = ./a.out
Compare mesh points
1.57017982 1.57080817 1.57143652
1.57016802 1.57079625 1.57142460
Compare function values at these mesh points
1622.04211 -84420.7344 -1562.01758
1591.57471 13245402.0 -1591.65527
There is a bad programming practice in the demo you found: moving through the mesh by adding steps (+ increment) instead of counting them (k * increment). The problem is widespread and appears with severe consequences (https://www.ima.umn.edu/~arnold/disasters/patriot.html).
For demonstration on your code, the size of the mesh was boosted to 10K points. Also the sample function changed from cos x to tan x and we examine points near the singularity at pi/2 = 1.57079633. While the novitiate may find the discrepancies in mesh values trivial, the difference in function values is significant.
(Mesh errors can be reduced by using increments which have exact binary representation like 2^(-13) = 1 / 8192.)
The code is shown here. The compilation command is gfortran -Wall -Wextra -Wconversion -Og -pedantic -fmax-errors=5 demo.f95. The run command is ./a.out.
program demo
use iso_fortran_env
implicit none
real, parameter :: pi = acos ( -1.0 )
integer, parameter :: n = 10001
real, dimension ( 1 : n ) :: x, y, z
real :: a = 0.0, b = 2 * pi
real :: increment
integer :: k, quarter, status
character ( len = * ), parameter :: c_options = compiler_options( )
character ( len = * ), parameter :: c_version = compiler_version( )
character ( len = 255 ) :: host = " ", cmd = " "
! queries
call hostnm ( host, status )
call get_command ( cmd )
! write identifiers
write ( *, '( /, "host system = ", g0 )' ) trim ( host )
write ( *, '( "compiler version = ", g0 )' ) c_version
write ( *, '( "compiler options = ", g0 )' ) trim ( c_options )
write ( *, '( "execution command = ", g0, / )' ) trim ( cmd )
increment = ( b - a ) / ( n - 1 )
quarter = n / 4
! mesh accumulates errors
x ( 1 ) = 0.0
do k = 2, n
x ( k ) = x ( k - 1 ) + increment
end do
y = tan ( x )
print *, 'Compare mesh points'
print *, x ( quarter : quarter + 2 )
! better mesh
x ( 1 ) = 0.0
do k = 2, n
x ( k ) = ( k - 1 ) * increment
end do
z = tan ( x )
print *, x ( quarter : quarter + 2 )
print *, 'Compare function values at these mesh points'
print *, y ( quarter : quarter + 2 )
print *, z ( quarter : quarter + 2 )
end program demo