seahorse and gpg

[Richard PALO <richard%netbsd.org@localhost>]

pbulk balks at seahorse, which seems to compile fine normally.  It appears 
to come from a missing dependency on gpg (implicit via gpgme?).

I tried the following quick hack:
> richard@omnis:/home/richard/src/pkgsrc/security/seahorse$ git diff ../gpgme
> diff --git a/security/gpgme/buildlink3.mk b/security/gpgme/buildlink3.mk
> index bf9de4f..678952f 100644
> --- a/security/gpgme/buildlink3.mk
> +++ b/security/gpgme/buildlink3.mk
> @@ -17,9 +17,17 @@ pkgbase:=            gpgme
>  .include "../../mk/pkg-build-options.mk"
>  
>  .if !empty(PKG_BUILD_OPTIONS.gpgme:Mgpgsm)
> +.  if !empty(PKG_BUILD_OPTIONS.gpgme:Mgnupg21)
> +DEPENDS+=              gnupg21>=2.1:../../security/gnupg21
> +.  else
> +DEPENDS+=              gnupg2>=2.0<2.1:../../security/gnupg2
> +.  endif
>  GPGME_GPG=             ${PREFIX}/bin/gpg2
> +CONFIGURE_ARGS+=       ac_cv_path_GNUPG=${GPGME_GPG}
>  .else
> +DEPENDS+=              gnupg>=1.4.2:../../security/gnupg
>  GPGME_GPG=             ${PREFIX}/bin/gpg
> +CONFIGURE_ARGS+=       ac_cv_path_GNUPG=${GPGME_GPG}
>  .endif
>  
>  .endif # GPGME_BUILDLINK3_MK

which seems to now build okay, but is this reasonable? Too me it seems convoluted...
Could someone a bit more familiar with these waters take a look?

I believe there may be other "issues" around gpg use elsewhere where there 
may be a better solution. BTW, why are there so many flavours?

-- 
Richard PALO