[gmx-users] Compilation problems for 3.2.1

[Erik Lindahl]

Hi Ross,
>
> This can be fixed by adding the 'static' attribute to bOne:
>
> 395   static bool bOne = TRUE;
>
> Is this a reasonable fix please?

Oops. We must have missed that one when fixing all the others - yes, 
that's the way it should be.

> ---
> PROBLEM 2:
> make_edi.c:467: initializer element is not computable at load time
> make_edi.c:467: (near initialization for `pa[0].option')
> make_edi.c:468: initializer element is not computable at load time
> make_edi.c:468: (near initialization for `pa[0]')
> make_edi.c:469: initializer element is not computable at load time
> make_edi.c:469: (near initialization for `pa[1].option')
> make_edi.c:470: initializer element is not computable at load time
> make_edi.c:470: (near initialization for `pa[1]')
> ...
> make_edi.c:491: initializer element is not computable at load time
> make_edi.c:491: (near initialization for `pa[12].option')
> make_edi.c:492: initializer element is not computable at load time
> make_edi.c:492: (near initialization for `pa[12]')
>
> This problem is (for someone with my knowledge of C at least) a bit 
> more
> complicated! Can anyone suggest a fix for this please?
>

Ah, make_edi is a new program contributed by Oliver Lange and Bert de 
Groot. The options seem to be static, but they use indices in arrays 
instead of specifying some strings explicitly.

Try replacing them with the values above, i.e. for line 467:

{ "-mon", FALSE, etSTR, {NULL},


etc., and let me know if that works.

Cheers,

Erik