This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch] intrinsic.texi: 13 down


Hello,

Daniel Franke schrieb:
>> Shouldn't one use z'F' rather than 'F'z?
> citing the latest GFortran Manual, section 7.8:
> "BOZ literal constants can also be specified by adding a suffix to the string. 
> For example, Z'ABC' and 'ABC'Z are the same constant."

I thought the idea is to use - if possible - the Fortran standard way.
According my Fortran 2003 copy:

R411 boz-literal-constant
     is binary-constant or octal-constant or hex-constant

R412 binary-constant
     is   B ’ digit [ digit ] ... ’
     or   B " digit [ digit ] ... "
C408 (R412) digit shall have one of the values 0 or 1.

R413 octal-constant
     is O ’ digit [ digit ] ... ’
     or O " digit [ digit ] ... "
C409 (R413) digit shall have one of the values 0 through 7.

R414 hex-constant
     is Z ’ hex-digit [ hex-digit ] ... ’
     or Z " hex-digit [ hex-digit ] ... "
R415 hex-digit is digit or A or B or C or D or E or F


Therefore I would like to see z'F' rather than 'F'z, even
though gfortran supports both.


> The other points you raise are valid, but maybe beyond the scope of those 
> showcase snippets?!

Well, my idea is that those snippets give new Fortran programmers
some good example. Putting the initialization in the same line
I would not recommend because of the SAVE side effect. I didn't
say you should input all those versions.

I actually wanted to say that I prefer something like
  real :: a,b
  a = z'F'
  b = z'3'
to the current
  INTEGER :: a, b
  DATA a / 'F'Z /, b / '3'Z /
and to the previous
  INTEGER :: a = 'F'X, b = '3'X

Tobias


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]