Re: '-light' vs. '-minimal' packages

[Mathieu Lirzin <mthl@openmailbox.org>]

Alex Kost <alezost@gmail.com> writes:

> Ludovic Courtès (2015-09-02 16:09 +0300) wrote:
>
>> Ricardo Wurmus <rekado@elephly.net> skribis:
>>
>>> An alternative is to inherit from the “r” package in “statistics.scm”.
>>>
>>>     (define-module (my own packages)
>>>       #:use-module (guix packages)
>>>       #:use-module (gnu packages statistics)
>>>       #:use-module (srfi srfi-1))
>>>
>>>     (define-public my-r
>>>       (package (inherit r)
>>>         (name "my-r")
>>>         ;; modify only the inputs here
>>>         (native-inputs (alist-delete "texlive" (package-native-inputs r)))
>>>         (inputs (alist-delete "icedtea6" (package-inputs r)))))
>>
>> Speaking of which: should we get rid of icedtea6:jdk in the default R
>> package (closure size: 1 GiB), and maybe of TeX Live (4 GiB)?  Or should
>> we provide, say, ‘r-light’ with the definition above?
>
> We have ‘bash-light’ and ‘wpa-supplicant-light’; ‘bioperl-minimal’ and
> ‘cups-minimal’.  I think we should stick to a single name for such
> light/minimal packages.  (I prefer "…-minimal")
                            ^^^
+1

--
Mathieu Lirzin