[Patch, fortran] PR28174 - actual args with substring refs and PR28167 - character array constructors

[Paul Thomas]

Steve,

Thanks for looking at it - I was just about to ping the list again.

>Paul, the patch is ok with the possible exception of the above.
>If you removed "static" to make the function visible to a function
>in another file, then I believe our convention is start the 
>function with the "gfc_" prefix.  Hence, resolve_character_array_constructor
>becomes gfc_resolve_character_array_constructor.
>
>  
>
You are absolutely right; I will make it so.

I am increasingly convinced that a number of the character related 
patches that we have fixed in trans-xxxx.c over the last year were 
really front end problems of this kind.  In fact, following in the 
tradition, I had an initial fix for this problem that resided in 
trans-array.c.  Then the light bulb came on - why do we not have a 
char_len for these objects? After that the fix was straight forward.

Anyway, back to installing SUSE10.1 on a faintly recalcitrant 64bit laptop.

Thanks again - I will commit tomorrow.

Paul