| 1 |
On 11/15/2009 03:25 PM, Duncan wrote: |
| 2 |
> There's a couple of experimental projects whereby portage is modified |
| 3 |
> to be able to handle multiple ABI installations in parallel. As a |
| 4 |
> matter of practice, I don't know if they'll ever get merged, because |
| 5 |
> we've gone this long without it, and as I mentioned, the worst need |
| 6 |
> was along about 2006 or so, when a lot of folks had switched already |
| 7 |
> but Flash and etc weren't yet available for 64-bit, and even |
| 8 |
> mainstream FLOSS apps like Open Office hadn't been ported. Since |
| 9 |
> pretty much everything mainstream FLOSS has been ported now, and the |
| 10 |
> proprietaryware folks are coming around to 64-bit as well, there's |
| 11 |
> far less need for multilib in general than there used to be, and the |
| 12 |
> need/demand will be ever weaker with time. |
| 13 |
|
| 14 |
Multilib is still very useful. 64-bit Firefox is slow as molasses |
| 15 |
currently (it uses the legacy javascript engine), and every 32-bit app |
| 16 |
that uses Qt or Gtk looks extremely ugly due to missing 32-bit Qt/Gtk |
| 17 |
styles. Wine will never be 64-bit and it needs libraries not available |
| 18 |
in emul packages. |
| 19 |
|
| 20 |
Also, lets not forget that many people would prefer a 32-bit userland |
| 21 |
running on a 64-bit kernel. |
| 22 |
|
| 23 |
Bottom line, multilib will be needed (and useful) for many years to come. |
Archives