Simple Array Initialisation

PGK

unread,

Dec 27, 2009, 7:19:54 PM12/27/09

to gg95

The short test program below,

program main
integer, dimension(3,4) :: A
integer, dimension(2) :: B = shape(A)
end program main

gives the following error on g95 v0.92 on Ubuntu 9.10 Netbook Remix:

integer, dimension(2) :: B = shape(A)
1
Error: Can't convert INTEGER(0) to INTEGER(4) at (1)

I have other systems with g95 v0.92 installed which don't fail, and
gfortran also gives no error. Any suggestions?

Graham

Jimmy

unread,

Dec 28, 2009, 5:12:22 AM12/28/09

to gg95

Also fails on my Win7 32bit PC. Puzzling why you have found that it
doesn't fail on other systems with g95. Silverfrost compiler also
fails with error ***Constant expression required. I think that the
instruction is not legal F95 because the initialization is a function
result, but I am not certain of this. If it is legal then there is a
bug in G95, if not there is a bug in gfortran. Writing it as

program main
integer, dimension(3,4) :: A
integer, dimension(2) :: B
B = shape(A)
end program main

works okay.

Jimmy.

PGK

unread,

Dec 28, 2009, 11:34:36 AM12/28/09

to gg95

Looking at the standard, it seems that functions appearing within
initialisation expressions must be elemental and intrinsic. "shape"
then is, I think, non-elemental, so that's one result...

I checked again with my Desktop 64-bit Ubuntu 9.04. It does pass, and
it's also g95 v0.92. Here's the entire output of "g95 -v" on that
machine:

Using built-in specs.
Target:
Configured with: ../configure
Thread model: posix
gcc version 4.0.3 (g95 0.92!) Nov 24 2008

And the result of "gcc -v"
...
gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4)

...and here's the same from the failing Ubuntu Netbook Remix. "g95 -
v":

Using built-in specs.
Target:
Configured with: ../configure --enable-languages=c
Thread model: posix
gcc version 4.0.3 (g95 0.92!) Jun 24 2009

...and the result of "gcc -v"
...
gcc version 4.4.1 (Ubuntu 4.4.1-4ubuntu8)

I see that the g95 v0.92 dates are different. It may be worthwhile for
me to try and obtain the older (Nov 2008) version for the Netbook.
(Could this be a version 0.93.)
Graham

jfh

unread,

Dec 30, 2009, 4:29:14 PM12/30/09

to gg95

On Dec 29, 5:34 am, PGK <graham.k...@gmail.com> wrote:
> Looking at the standard, it seems that functions appearing within
> initialisation expressions must be elemental and intrinsic. "shape"
> then is, I think, non-elemental, so that's one result...

That's not the problem. Shape(A) is an inquiry intrinsic function.
That, with its argument A as given, is OK in this context by the F95
standard 7.1.6.1. I get the error with g95 on a machine on which g95 -
v says

Using built-in specs.
Target:
Configured with: ../configure
Thread model: posix

gcc version 4.0.3 (g95 0.92!) Jun 24 2009

and gcc -v says

Using built-in specs.
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --enable-languages=c,c+
+,fortran,objc,obj-c++,treelang --prefix=/usr --enable-shared --with-
system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-
threads=posix --enable-nls --with-gxx-include-dir=/usr/include/c++/4.2
--program-suffix=-4.2 --enable-clocale=gnu --enable-libstdcxx-debug --
enable-objc-gc --enable-mpfr --enable-checking=release --build=x86_64-
linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.2.4 (Ubuntu 4.2.4-1ubuntu4)