| 1 |
On Thu, Jul 8, 2021 at 1:41 PM Peter Stuge <peter@×××××.se> wrote: |
| 2 |
> |
| 3 |
> Matt Turner wrote: |
| 4 |
> > If you can find a case where you wouldn't want to enable one of these |
| 5 |
> > USE flags, please let me know and I'll reconsider my position. |
| 6 |
> |
| 7 |
> My catalyst spec files all have use: -* foo bar x y z |
| 8 |
> specifically because the defaults are never what I want for any given |
| 9 |
> system. I build desktops, servers, containers, VM appliance images and |
| 10 |
> embedded system, and I know what I want in each one. Especially the |
| 11 |
> latter frequently have only very few USE flags set and I want zero |
| 12 |
> extra dependencies. |
| 13 |
|
| 14 |
I think you're making a great argument that you'd be completely |
| 15 |
unaffected by any of the suggestions in this thread. |
| 16 |
|
| 17 |
> I completely agree that the default USEs should rather be reduced, |
| 18 |
> not increased. Isn't this what profile inheritance is for? It would |
| 19 |
> be great if I didn't essentially have to create my own profile when I |
| 20 |
> want a very minimal system. |
| 21 |
> |
| 22 |
> |
| 23 |
> Matt Turner wrote: |
| 24 |
> > I'd claim most of these packages' bzip2/lzma/zstd USE flags should |
| 25 |
> > be removed in favor of statically enabling them |
| 26 |
> |
| 27 |
> That is the direct opposite of Gentoo's single most core value: choice |
| 28 |
|
| 29 |
Choice makes sense when there's a legitimate trade-off to be made. |
| 30 |
Choice isn't dogma. |
| 31 |
|
| 32 |
> Just don't do it. Kthx. |
| 33 |
|
| 34 |
This kind of thing is nothing but irritating. Please don't do this. |
Archives