Re: Is inclusion in GIR files transitive?

[Colin Walters <walters verbum org>]

On Mon, Jan 19, 2015, at 03:53 PM, Philip Withnall wrote:
> I’ve attached a really quick patch which fixes this for me, but it
> probably misses some code paths, and probably doesn’t match with Colin’s
> original intention for dependencies[1]. I don’t really know.

I think it's better to have this be non-transitive, and offer an API to compute the full
transitive set.  If it was transitive at compilation time, it would require a rebuild of
all consumers if a library dropped an unnecessary dependency, for example.
It wouldn't allow accurately distinguishing what libraries a given library *really*
depended on directly.