| 1 |
This is what Thilo Bangert at Mon, May 13, 2002 at 09:30:59AM +0200 wrote: |
| 2 |
> its actually quite simply: don't use inetd! |
| 3 |
|
| 4 |
Fair enough. But who should decide which inetd replacement should be used? |
| 5 |
|
| 6 |
> i know, you meant this to be an example - but things are not meant to be |
| 7 |
> simple in gentoo linux. we want users to learn the tools they use - and |
| 8 |
> not force upon them some distribution specific way of doing things |
| 9 |
|
| 10 |
It wouldn't be forced on to the user. If it detects that the destination has |
| 11 |
been modified by someone else, then it wouldn't do anything. I'd also see it as |
| 12 |
more of an advisory tool, creating a list of instructions to complete the |
| 13 |
installation, such as "Add the following line to inet.conf", or "Add this to |
| 14 |
xinetd.conf", and have that advice be the most appropriate for the user's |
| 15 |
setup. |
| 16 |
|
| 17 |
This is more about decoupling the maintainer's assumptions from the user's |
| 18 |
choices. With the system in place, a package would have a single interface to |
| 19 |
the variants of the packages it depends on, so the user can install whatever |
| 20 |
they like, and there's less if/thens in the .ebuild. |
| 21 |
|
| 22 |
When a new variant is added, there doesn't need to be a single thing changed in |
| 23 |
the ebuild. Or, when a program's configuration syntax changes radically, it |
| 24 |
doesn't break things that depend on it. |
| 25 |
|
| 26 |
> thats one of the reasons gentoo is so great in the first place... |
| 27 |
|
| 28 |
Giving the user control is always good, in my book. |
| 29 |
|
| 30 |
James, |
| 31 |
-- |
| 32 |
James Vandenberg Email: james at vandenberg.dropbear.id.au |
| 33 |
GPG FP= 65AB 179A D884 EDC6 216D FE6A 6833 02BC 4425 4F70 |
| 34 |
Quidquid latine dictum sit, altum viditur. ICQ: 151135390 |
Archives