[LORENE] new config-file reading routines

Reinhard Prix Reinhard.Prix at aei.mpg.de
Thu Dec 4 15:12:30 CET 2003


Hi,

this is just an announcement that there are now some new
config-file reading routines available in Lorene. These allow you to
write somewhat more flexible config-files with a "variable = value"
type syntax. Everything else is ignored as comments. The only
restriction for variable-assignments is that they have to occur at the
beginning (modulo whitespace) of a line. I think the full syntax is 
best explained  in an example:

--------------------------------------------------
// this is a comment
var1=10.9	// ok
    var2  = hi  # this is ok too, whitespace is ignored
 #  var3  = nix // this would be considered a comment
// you can also write:
var4 : 1       // ok
# or even just  
var5	99.9     // ok
--------------------------------------------------

Note that the above lines would constitute a valid config-file as they
are, defining the variables var1,var2, var4 and var5.

Note also that the comment-syntax is basically free, so you can use
whatever comment-marker you like (including none(!), but best use some
non-alphanumeric character!)

To read this, e.g. the file "settings.par", you can either use the
general-purpose function: 

read_variable ("settings.par", "var1", "%f", &some_float);

where the format-string "%f" corresponds to sscanf(),
or you can use a type-specific overloaded function, like this:

read_variable ("settings.par", "var1", some_float);

Note that the file "settings.par" is buffered in the first call and
will not be re-read from disk in subsequent calls if either the same
filename or NULL is passed as an argument.

read_variable (NULL, "var4", some_bool);
read_variable (NULL, "var5", some_double);

Return is 0 if OK, -1 on error.

These routines are not heavily tested yet, but they seem to work fine
as far as I've used them in sfstar.C.

Please let me know if there are any problems/suggestions.

Cheers,
Reinhard.

-- 
-----------------------------------------------------
MPI fuer Gravitationsphysik | 
Albert-Einstein Institut    | 
Am Muehlenberg              | Tel: +49(0)331 567 7128
14476 Golm b. Potsdam       | Fax: +49(0)331 567 7298
Germany                     |
-----------------------------------------------------


More information about the Lorene.list mailing list