Code review comment for lp://qastaging/~alan-griffiths/mir/add-mir_surface_spec_attach

> My gut tells me we should maybe be resolving bug 1420334 at the same time. And
> this API appears to not.

There are various feature requests in this area that may require additional APIs. These are the ones I'm aware of ATM:

1. Repositioning with respect to parent. (relates to, but doesn't solve lp:1603086)
2. "Update MirAttachmentEdge API for GTK requirements" (essentially support for GDK APIs https://bugzilla.gnome.org/show_bug.cgi?id=756579#c311)
3. An API for movement of a surface relative to its current location lp:1420334
4. The ongoing discussion around possible inconsistencies of the window management spec and the toolkit API. (c.f. the "tooltips - everything is simple until you understand the problem" thread.)

Resolving any or all of these may introduce or update some toolkit APIs in this area.

My gut says that the API proposed here helps with an immediate problem without getting in the way of further enhancements. But maybe this is a good time for sketching out a final solution?