| 1 |
Ciaran McCreesh wrote: |
| 2 |
> || ( |
| 3 |
> sdl? ( media-libs/libsdl ) |
| 4 |
> svga? ( media-libs/svgalib ) |
| 5 |
> opengl? ( virtual/opengl ) |
| 6 |
> ggi? ( media-libs/libggi ) |
| 7 |
> virtual/x |
| 8 |
> ) |
| 9 |
|
| 10 |
> As has been discussed in the past, the only correct way of handling |
| 11 |
> this from an ebuild perspective is lots of use && has_version calls |
| 12 |
> (relying upon use only is insufficient, since if two or more flags are |
| 13 |
> used, only one of the dependencies need be installed; relying upon |
| 14 |
> has_version only is insufficient, since a dependency may be installed |
| 15 |
> but its associated use flag could still be turned off). |
| 16 |
|
| 17 |
It's doable with just use. For the example above, this would be the |
| 18 |
solution: |
| 19 |
|
| 20 |
interface="x11" |
| 21 |
use ggi && interface="ggi" |
| 22 |
use opengl && interface="opengl" |
| 23 |
use svgalib && interface="svga" |
| 24 |
use sdl && interface="libsdl" |
| 25 |
|
| 26 |
After that, instead of using "use $flag", just use |
| 27 |
"[[ ${interface} == $flag ]]" |
| 28 |
|
| 29 |
A combination of use and has_version which does not use the concept |
| 30 |
shown above wouldn't work reliably anyway as a non-dependency might be |
| 31 |
installed and its flag turned on, but another interface should be |
| 32 |
selected since its flag has a higher priority. |
| 33 |
|
| 34 |
> So, is there a legitimate reason for this complication to exist? Or |
| 35 |
> should use? blocks being direct children of || ( ) be forbidden? |
| 36 |
|
| 37 |
I don't see how this is a complication, tbh. The alternative would be to |
| 38 |
have the same DEPEND but without the ||(). But then you would have to a) |
| 39 |
let the ebuild die if more then one flag is turned on or b) have it |
| 40 |
follow the same logic as shown above. |
| 41 |
|
| 42 |
I also don't see how it is a nuisance to explain. Explain how ||() works |
| 43 |
with a "normal" dependency, explain how use? () dependencies work and it |
| 44 |
will become clear to everybody how it will work. TBH, it looks to me |
| 45 |
like you were to explain how an if statement in a while loop works, even |
| 46 |
though it is well-known how both of them work if they aren't nested. |
| 47 |
|
| 48 |
-- |
| 49 |
Kind Regards, |
| 50 |
|
| 51 |
Simon Stelling |
| 52 |
Gentoo/AMD64 developer |
| 53 |
-- |
| 54 |
gentoo-dev@g.o mailing list |
Archives