| 1 |
On 5/12/06, Jerry McBride <mcbrides9@×××××××.net> wrote: |
| 2 |
> On Thursday 11 May 2006 19:47, Richard Fish wrote: |
| 3 |
> > On 5/11/06, Jerry McBride <mcbrides9@×××××××.net> wrote: |
| 4 |
> > > I'm going one step further with gcc 4.1.0. After I emerged gcc and |
| 5 |
> > > glibc... I did an "emerge -e system" twice and am now following up with |
| 6 |
> > > two "emerge -e world" commands... |
| 7 |
> > |
| 8 |
> > Wow, you like to waste a lot of CPU cycles... |
| 9 |
> > |
| 10 |
> |
| 11 |
> Actually... nothing is wasted. I've read that this is the best way to rebuild |
| 12 |
> the tool chain, then the applications. Sources that rely on other sources are |
| 13 |
> guaranteed to be accurately built after the second pass of "world" |
| 14 |
|
| 15 |
I'm sorry, but what you read was simply wrong, written by somebody who |
| 16 |
probably didn't understand how compilers, linkers, dynamic libraries, |
| 17 |
and executables interact. |
| 18 |
|
| 19 |
I could see _some_ value in emerge -e system followed by emerge -e |
| 20 |
world. There can be some (very small) effects of system packages on |
| 21 |
each other. But you are building system again when you emerge -e |
| 22 |
world, and there is simply no reason at all to emerge -e world twice. |
| 23 |
|
| 24 |
-Richard |
| 25 |
|
| 26 |
-- |
| 27 |
gentoo-user@g.o mailing list |
Archives