Jan Wielkiewicz <tona_kosmicznego_smiecia@interia.pl> writes:

> * gnu/packages/jami.scm (libringclient):
> [propagated-inputs]: Remove qtbase.
> [inputs]: Add qtbase.
> ---
>  gnu/packages/jami.scm | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/gnu/packages/jami.scm b/gnu/packages/jami.scm
> index 041955f297..ea94efb2f9 100644
> --- a/gnu/packages/jami.scm
> +++ b/gnu/packages/jami.scm
> @@ -530,10 +530,11 @@ functionality.")
>                   (base32
>                    "1s1wi80drabfjhvxbiyl8k9c2jsafvmx5v84s3wldsmmdg4f6xdf"))))
>        (build-system cmake-build-system)
> -      (propagated-inputs
> -       `(("libring" ,libring)     ; For 'dring'.
> -         ("qtbase" ,qtbase)       ; Qt is included in several installed headers.
> +      (inputs
> +       `(("qtbase" ,qtbase)       ; Qt is included in several installed headers.
>           ("qttools" ,qttools)))
> +      (propagated-inputs
> +       `(("libring" ,libring)))     ; For 'dring'.
>        (arguments
>         `(#:tests? #f                      ; There is no testsuite.
>           #:configure-flags

What's the reason behind this change? I'd interpret the "headers"
comment as justifying why qtbase should be propagated.