--- Comment #7 from kargl at gcc dot gnu dot org 2007-04-13 20:44 ---
(In reply to comment #6)
> Sorry, I cannot find another compiler that agrees with gfortran -
Then, I suggest you submit a bug report. The standard explicitly
says
"If a data-stmt-constant is a boz-literal-cons
--- Comment #6 from dir at lanl dot gov 2007-04-13 20:32 ---
Sorry, I cannot find another compiler that agrees with gfortran -
[EMAIL PROTECTED] ~/tests]$ ifort -o mask mask.f90
[EMAIL PROTECTED] ~/tests]$ mask
8000
FF00
qsc10:~/tests [6] > f90 -o mask mask.f90
qsc10:~/tests [7]
--- Comment #5 from kargl at gcc dot gnu dot org 2007-04-13 18:38 ---
(In reply to comment #4)
> With Hex and octal numbers - there is no overflow as long as there are enough
> bits - that is why g95 and Absoft do not complain.
>
g95 and absoft do not complain because these compilers e
--- Comment #4 from dir at lanl dot gov 2007-04-13 15:52 ---
With Hex and octal numbers - there is no overflow as long as there are enough
bits - that is why g95 and Absoft do not complain.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31563
--- Comment #3 from dir at lanl dot gov 2007-04-13 15:28 ---
I do not see why you say the numbers over flow - g95 gives exactly the correct
answer when I print them out
[dranta:~/junk] dir% g95 -o mask mask.f90
[dranta:~/junk] dir% mask
8000
FF00
[dranta:~/junk] dir% cat mask.f
--- Comment #2 from burnus at gcc dot gnu dot org 2007-04-13 15:08 ---
Well, gfortran is right:
x'8000' = 2147483648 > 2147483647 = huge(msk1)
and
o'377' = 4278190080 > 2147483647 = huge(msk4)
Thus the BOZ numbers are too big for the 4-byte variables. The other compilers
s
--- Comment #1 from fxcoudert at gcc dot gnu dot org 2007-04-13 15:02
---
The code you report is invalid, and gfortran is right to throw an error. If you
really want it to compile, please use -fno-range-check.
(see http://gcc.gnu.org/ml/fortran/2007-04/msg00123.html for details and a
p
