META: gg.top should depend on gg

[whitequark]

Subject says it. I don't think there is a reason to require gg.top but not gg, and adding a dependency means one #require less.

[dbuenzli]

You've got things the wrong way around, see this line. Just #require "gg" (sorry for the unbound values in the released version):

# #require "gg";;
/Users/dbuenzli/.opam/4.01.0/lib/ocaml/unix.cma: loaded
/Users/dbuenzli/.opam/4.01.0/lib/ocaml/bigarray.cma: loaded
/Users/dbuenzli/.opam/4.01.0/lib/gg: added to search path
/Users/dbuenzli/.opam/4.01.0/lib/gg/gg.cma: loaded
/Users/dbuenzli/.opam/4.01.0/lib/gg/gg_top.cma: loaded
Unbound value Raster.pp_buffer.
Unbound value Raster.pp_sample_format.

[whitequark]

Oops. Sorry for the noise.

[dbuenzli]

@whitequark Actually I think that what you assumed maybe the right thing to do. The problem with my current toplevel scheme, is that I don't give the choice to people. If they #require "gg", the damn thing will open Gg and that may be annoying for some reason (e.g. name clashes).

So I think it would be better if #require "gg" wouldn't do anything and require "gg.top" would depend on gg and do the open Gg and install the printers. So that you are actually given the choice.

What do you think ?

[whitequark]

@dbuenzli I think the scheme you propose is much more clear for the users. I certainly like it more; a package shouldn't behave differently when loaded in toplevel and when not.

[dbuenzli]

@whitequark Yes. I'm going to gradually change that in my distributed packages.