| 1 |
begin quote |
| 2 |
On Fri, 13 Feb 2004 13:03:32 +0000 |
| 3 |
James Harlow <james@××××××××××××××.nu> wrote: |
| 4 |
|
| 5 |
|
| 6 |
> > Here's the point, we can easly remove the dependency on esd, and if |
| 7 |
> > the code works, okay. However, then you have an untracked |
| 8 |
> > dependency when esd -is- installed but set as USE="-esd" , as it |
| 9 |
> > will be needed, but not listed as such. |
| 10 |
|
| 11 |
|
| 12 |
|
| 13 |
> Sure. I'm implicitly assuming that the --without-esd ./configure |
| 14 |
> switch does the right thing, which obviously needs to be tested. But |
| 15 |
> it's a pretty simple test. |
| 16 |
> |
| 17 |
> That's what you were referring to, right? |
| 18 |
|
| 19 |
No, if that isn't used its a simple thing to add a USE flag. The issue |
| 20 |
is with things that go... |
| 21 |
"do I have alsa? no... not building alsa interface" |
| 22 |
"do I have esound? no .. not building esound interface" |
| 23 |
|
| 24 |
|
| 25 |
without a choice to -disable (or enable) such things, when you can state |
| 26 |
USE="-*" and then still get everything included, no dependencies tracked |
| 27 |
and a hosed system when you finally try to live without alsa+esound. |
| 28 |
|
| 29 |
too many builds use auto* functions like that, and no. "dynamic |
| 30 |
dependencies" isn't a solution here either. Metal hacking configure.* |
| 31 |
is however one. |
| 32 |
|
| 33 |
|
| 34 |
//Spider |
| 35 |
|
| 36 |
-- |
| 37 |
begin .signature |
| 38 |
Tortured users / Laughing in pain |
| 39 |
See Microsoft KB Article Q265230 for more information. |
| 40 |
end |
Archives