mbox series

[bug#54615,v2,0/2] Add riscv-pk and improve spike

Message ID 20220328185557.27310-1-arunisaac@systemreboot.net
Headers show
Series Add riscv-pk and improve spike | expand

Message

Arun Isaac March 28, 2022, 6:55 p.m. UTC 
Hi Efraim,

I didn't realize you had already packaged riscv-pk in the guix-bioinformatics
channel. I copied the idea of using #:out-of-source? from your version of the
package. Here is an improved patchset.

In your version of the package, I didn't understand why you had #:target and
native-inputs. Could you enlighten me?

Thanks,
Arun

Arun Isaac (2):
  gnu: spike: Substitute path to dtc instead of wrapping executables.
  gnu: Add riscv-pk.

 gnu/packages/virtualization.scm | 54 +++++++++++++++++++++++++++------
 1 file changed, 45 insertions(+), 9 deletions(-)

Comments

Efraim Flashner March 28, 2022, 8:10 p.m. UTC | #1 
On Tue, Mar 29, 2022 at 12:25:55AM +0530, Arun Isaac wrote:
> Hi Efraim,
> 
> I didn't realize you had already packaged riscv-pk in the guix-bioinformatics
> channel. I copied the idea of using #:out-of-source? from your version of the
> package. Here is an improved patchset.
> 
> In your version of the package, I didn't understand why you had #:target and
> native-inputs. Could you enlighten me?
> 
> Thanks,
> Arun

The native-inputs I got rid of because they weren't needed in the end, I
just hadn't committed the change yet. For target I was under the
impression that it had to be compiled targeting
riscv64-(unknown-)?linux-gnu so I had that there. The instructions
upstream involved using a configure-flag instead, which is why I started
with the native-inputs.

> Arun Isaac (2):
>   gnu: spike: Substitute path to dtc instead of wrapping executables.
>   gnu: Add riscv-pk.
> 
>  gnu/packages/virtualization.scm | 54 +++++++++++++++++++++++++++------
>  1 file changed, 45 insertions(+), 9 deletions(-)
> 
> -- 
> 2.34.0
>
Arun Isaac March 29, 2022, 10:39 a.m. UTC | #2 
> The native-inputs I got rid of because they weren't needed in the end, I
> just hadn't committed the change yet. For target I was under the
> impression that it had to be compiled targeting
> riscv64-(unknown-)?linux-gnu so I had that there. The instructions
> upstream involved using a configure-flag instead, which is why I started
> with the native-inputs.

Thanks, that clears it up! I've added #:target in v3 of my patchset, and
will send it shortly.