Code review comment for lp:~andreas-pokorny/mir/bump-input-platform-abi-for-0.22

> I think this adds to an already confusing situation.
>
> We have input and graphics ABIs that relate to the corresponding interfaces
> used by mirserver.
>
> These are implemented in the assorted platform modules which use them to name
> the stanzas in their symbols maps.
>
> The .soname numbering (which this MP is addressing) shouldn't be tied to to
> the ABI numbers, nor should the different platform modules be tied to the same
> .soname numbering.
>
> There is a problem with the X11 platform that needs to support both input and
> graphics ABIs needed by the server but, theoretically at least, could support
> just one. But fiddling with the .soname doesn't help.

Sorry for conflating probing and other topic into this discussion.

This MP just avoids incrementing the kms and android counter when not necessary. I guess the names are misleading, instead of
set(MIR_SERVER_INPUT_PLATFORM_ABI 5)
set(MIR_SERVER_GRAPHICS_PLATFORM_ABI 9)
we could have
set(MIR_SERVER_INPUT_PLATFORM_SUFFIX 5)
set(MIR_SERVER_GRAPHICS_PLATFORM_SUFFIX 9)
set(MIR_SERVER_COMBINED_PLATFORM_SUFFIX 10)

or _COUNTER ..