[Users] cctk_ScheduleFunctions.h assuming "int" for a "void" scheduled routine?

[Bernard Kelly]

Thanks, Roland!

Moving out of STARTUP solved it completely.

Bernard

On 27 June 2018 at 16:48, Roland Haas <rhaas at illinois.edu> wrote:
> Hello Bernard,
>
> STARTUP and SHUTDOWN are special (no grid hierarchy yet or none
> anymore), also you can abort Cactus in STARTUP routines if you
> returned non-zero (the docs says it is ignore right now though). See:
>
> http://einsteintoolkit.org/usersguide/UsersGuidech14.html#x19-205000D4
>
> Yours,
> Roland
>
>> Hi.
>>
>> I recently took a simple thorn --- one routine scheduled at  POSTEVOL
>> --- and enlarged it so that now there's a second routine scheduled at
>> STARTUP.
>>
>> Both routines are written in C, taking in CCTK_ARGUMENTS, and
>> returning void. The new routine is called Interpolate_setup.c
>>
>> During compilation, I get the following warning:
>>
>> ------------------------------
>> /nobackupp2/bjkelly1/codes/Cactus/configs/sim/build/interp_metric_sph_grid/Interpolate_setup.c(47):
>> warning #147: declaration is incompatible with "int
>> sph_grid_metric_initialize(void)" (declared at line 7 of
>> "/nobackupp2/bjkelly1/codes/Cactus/configs/sim/bindings/include/interp_metric_sph_grid/cctk_ScheduleFunctions.h")
>>   void sph_grid_metric_initialize(CCTK_ARGUMENTS){
>>          ^
>> ------------------------------
>>
>> Can anyone tell me why the build process is assuming that the new
>> routine is taking a void input and returning int? I can edit the
>> thorn's auto-generated "cctk_ScheduleFunctions.h" by hand to correct
>> this behavior, but it seems like it shouldn't be necessary.
>>
>> Bernard
>> _______________________________________________
>> Users mailing list
>> Users at einsteintoolkit.org
>> http://lists.einsteintoolkit.org/mailman/listinfo/users
>
>
>
> --
> My email is as private as my paper mail. I therefore support encrypting
> and signing email messages. Get my PGP key from http://pgp.mit.edu .