| 1 |
At Mon, 21 Nov 2005 14:13:14 +0100 "Hemmann, Volker Armin" <volker.armin.hemmann@××××××××××××.de> wrote: |
| 2 |
|
| 3 |
> On Monday 21 November 2005 13:33, Steve B wrote: |
| 4 |
>> WTF.. I'm getting ready to rebuild my gentoo box. I have always did a stage |
| 5 |
>> 1 install. i was under the impression that if u used a stage 3 u couldn't |
| 6 |
>> muck with your CFLAGS or what not. |
| 7 |
|
| 8 |
We have a lot in common in this regard. |
| 9 |
|
| 10 |
I also always had done stage1 installs. |
| 11 |
I also misremembered that CFLAGS couldn't be changed if using stage3. |
| 12 |
I also posted this mis-remembrance to the present list. |
| 13 |
|
| 14 |
As was very politely explained to me: |
| 15 |
|
| 16 |
1. It was CHOST, not CFLAGS. |
| 17 |
|
| 18 |
2. You can (at least now, not sure about previously) actually change |
| 19 |
CHOST, but must run a utility afterward. |
| 20 |
|
| 21 |
3. If you fiddle with USE flags, /etc/package.* files, etc, you can |
| 22 |
get the effect of a customized stage1 install by recompiling |
| 23 |
everything twice. I suspect simply |
| 24 |
emerge -e world; emerge -e world |
| 25 |
|
| 26 |
3A. Only a few items really need to be compiled the first time. |
| 27 |
|
| 28 |
3B. There are scripts in the forums to do this "automatically". |
| 29 |
|
| 30 |
Thus, I conclude that, unless you are changing bootstrap.sh (which I |
| 31 |
never did), you really can get the effect of a stage1 starting with |
| 32 |
stage3. |
| 33 |
|
| 34 |
I would not use the term "faster" to describe the stage3 equivalent of |
| 35 |
stage1, especially if you are lazy and do the double emerge -e world. |
| 36 |
Instead, I would say that there are two advantages to the user |
| 37 |
|
| 38 |
1. You can use your machine for reading/sending mail, browsing the |
| 39 |
web, etc while all the compilation is proceeding. So the machine |
| 40 |
is usable sooner, even if the procedure is not faster. |
| 41 |
|
| 42 |
2. You avoid the chance to make the kind of error I did in changed |
| 43 |
USE flags too aggressively, *too early* in the process. |
| 44 |
|
| 45 |
Perhaps more important is that there is one big advantage to the |
| 46 |
developers and seasoned gentoo experts on this list. |
| 47 |
|
| 48 |
1. Fewer installation errors due to faulty understanding of a stage1 |
| 49 |
process results in fewer requests for help, requests that likely |
| 50 |
to the expert seem to all be essentially the same, and do not lead |
| 51 |
to improvement of the stage1 process. |
| 52 |
|
| 53 |
A converted supported of stage3 --> stage1-equivalent installs, |
| 54 |
allan |
| 55 |
-- |
| 56 |
gentoo-user@g.o mailing list |
Archives