gwl-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: How to install GWL?


From: zimoun
Subject: Re: How to install GWL?
Date: Thu, 23 Jan 2020 15:18:50 +0100

Hi Ricardo,

On Thu, 23 Jan 2020 at 14:02, Ricardo Wurmus <address@hidden> wrote:

> I believe that by spawning “guix repl” and using it as an inferior we
> can remove the confusion, and we would make the separation of what is
> library code and what is run-time code clearer.

I think we agree on a clear separation between what is library code
and what is run-time code.
I trust you that "guix repl" should fit the needs.

My concern is:

 - 'guix describe' or 'guix --version' returns the current usable Guix
 - 'guix workflow' depends on the 2 versions of Guix when it was installed:
      + version of all the building machinery
      + and this version can be different of the version of Scheme
library as inputs
 - 'guix pull' modifies the result of 'guix describe' or 'guix --version'
 - then 'guix workflow' refers to 3 versions of Guix:
      + the 2 above
      + the current version from which the packages are fetched

This happens all the times when using any package and I am not
confused. Because the separation between install-time and run-time is
clear.

Note that the 2 first version are linked so it is not an issue and the
version of Guix when installing GWL is enough for reproducibility
because this very version implies the other one.

However, it still can be confusing because 2 versions of Guix are
required for reproducibility. The version to build GWL itself and the
version at run-time.
In 'guix workflow', the first term 'guix' refers to one version of
Guix and the second term 'workflow' refers to another one.


What you are proposing probably address well my concern. :-)


> Currently, that’s not what we’re doing.  Instead we use Guix as a
> library (whatever version of the “guix” package is available in the
> version of Guix used), and *also* use that version to install software.
> That’s clearly not desirable, hence my email.

I agree.
And I agree too that "inferior" seems the right approach.


Cheers,
simon



reply via email to

[Prev in Thread] Current Thread [Next in Thread]