1 |
I've recently been working on how to effectively build a new |
2 |
system using gcc-3.0.2. Since the current glibc complains |
3 |
when built using gcc-3+, I decided to bootstrap using gcc-2.95. |
4 |
I followed the build-from-source docs, except for the following |
5 |
changes: |
6 |
|
7 |
1. Comment out >=gcc-2.96 in package.mask. |
8 |
2. Change binutils to be 2.11.92.7 in /etc/make.profile/packages |
9 |
and files/boot*. |
10 |
3. Comment out net-misc/netkit-telnetd in /etc/make.profile/packages |
11 |
(because netkit-telnetd won't build with gcc-3+). |
12 |
|
13 |
I bootstrapped normally and then emerged the latest gcc: |
14 |
scripts/bootstrap.sh files/boot[tab] && emerge sys-devel/gcc |
15 |
and then I edited /etc/make.conf to use "-mcpu=athlon -march=athlon" |
16 |
before doing |
17 |
export CONFIG_PROTECT="" && emerge system. |
18 |
|
19 |
I finished up the rest of the install, rebooted, and discovered |
20 |
that I couldn't log in. It seems that the PAM ebuild didn't build |
21 |
the PAM pwdb module. *Sigh* |
22 |
|
23 |
I tried again, only this time I left "-mcpu=i686 -march=i686" in |
24 |
/etc/make.conf. I haven't found any glaring problems with this |
25 |
version, so now I'm rebuilding using the latest iso and we'll |
26 |
see what happens! |
27 |
|
28 |
-g2boojum- |
29 |
-- |
30 |
___________________________________________________________________ |
31 |
| Grant Goodyear | The Secrets of Physics: | |
32 |
| Dept. of Chemistry - Clemson U |1. Add zero. | |
33 |
| Clemson, SC 29634 |2. Multiply by one. | |
34 |
|-------------------------------------|3. Expand in a Taylor series| |
35 |
|e-mail: goodyea@×××××××.edu |4. Integrate by parts. | |
36 |
|www:bernacchi.chem.uh.edu/~grant |5. Fourier transform. | |
37 |
| |6. Add auxiliary variables | |
38 |
|_____________________________________|____________________________| |