Re: [Dragora-members] Relation of Qi and Qire

[Kevin "The Nuclear" Bloom]

Hi Michael,

I partly agree with this line of thought: why have 2 interfaces for the
same command? There really is only 1 reason for such a thing:
easy-of-use. Instead of folks trying to remember if they're to use Qi or
Qire, they can just use Qire since that's how most people will be
working with packages (if they choose to use remote repos, of course).

Yes the interface is being implemented twice but not the actual
code. For example, Qire may have a `-f` option but all Qire really does
with that information is pass it into Qi. Sort of like a puppet master
controller his puppet. This requires very very little code, really only
adding in a few lines to make the options valid and then a more generic
Qi execution line(s).

I'm fine with either approach since implementation is fairly simple and
would only be an extra source of ease for the end user.

Kev

Michael Siegel writes:

> Hi everyone,
>
> thinking about Dragora's package management again, I realized something
> about the relation of Qi and Qire that should probably be changed.
>
> At the moment, the idea is that a user would normally not have to touch
> Qi and would instead always use Qire because Qire is not only a remote
> extension for Qi, but it is also a complete Qi wrapper, if I understand
> it correctly.
>
> What this means is: There are two complete user interfaces to Qi, or
> rather: The Qi user interface is more or less being implemented twice.
>
> I want to suggest a simpler approach: Let Qire just be what it claims to
> be – the Qi remote _extension_, i.e., a piece of code that Qi will make
> use of when necessary. Then, only implement a user interface for Qi,
> including “remote commands”. But whenever those remote commands are
> being executed, Qi will simply hand anything it can't do by itself over
> to Qire, i.e., Qi itself would still not include the code implementing
> those remote commands.
>
> What do you think?
>
>
> Best
> Michael