Katherine Cox-Buday <cox.katherine.e@gmail.com> skribis:

> That is a mistake; there's no reason it can't be listed as a regular input.
>
> During my tests to confirm this, I realized that sbcl-cl-rdkafka
> wasn't actually producing any useful output. This led me down a
> rabbit-hole (albeit a productive one). Below is a new patch set.
>
> 11:23 kate says: guix refresh -l sbcl-lparallel
> Building the following 4 packages would ensure 8 dependent packages
> are rebuilt: ecl-cl-rdkafka@1.0.2 cl-rdkafka@1.0.2
> cl-random-forest@0.1-0.85fbdd4 next@1.5.0
>
> All dependent packages continue to build.
>
> On Fri, Jun 5, 2020 at 9:01 AM Guillaume Le Vaillant <glv@posteo.net> wrote:
>>
>>
>> Hi,
>>
>> Katherine Cox-Buday <cox.katherine.e@gmail.com> skribis:
>>
>> > (propagated-inputs
>> >  `(("librdkafka" ,librdkafka)))
>>
>> Why is the librdkafka library a propagated input instead of a regular
>> input?

Thanks. Patches pushed as 388a89750ec60a320dd7b6a0b0f174cbed4f50c6 and
64174aff927f4f9da18ea452247c838e6bab3306. Except I didn't push the
definition for ecl-cl-rdkafka because it didn't build (because of the
phase moving a file before the cleanup phase IIRC).