1 |
Hey Micheal, |
2 |
|
3 |
Michael Haubenwallner <haubi@g.o> writes: |
4 |
|
5 |
> Well, it is not "official" part of Gentoo Prefix, but it is a |
6 |
> de-facto requirement to bootstrap on AIX and HP-UX (maybe others). |
7 |
> It _may_ work to bootstrap without prefix-launcher there, |
8 |
> but it is a lot harder than on Linux and Solaris fex. |
9 |
> The reason it is on sf is that the "prefix"-idea was born before |
10 |
> "Gentoo Prefix" as the implementation. |
11 |
|
12 |
I see. |
13 |
|
14 |
I have struggled through the gcc phase. I did very dirty hacks including |
15 |
replacing failing steps with linking against hand-pointed libraries. |
16 |
|
17 |
So now I have gcc-4.2.4 at hand. |
18 |
|
19 |
Next comes the coreutils-6.9. |
20 |
|
21 |
I did ./configure --prefix=$HOME/tools then make |
22 |
|
23 |
,---- |
24 |
| gcc -std=gnu99 -I. -g -O2 -MT cycle-check.o -MD -MP -MF |
25 |
| .deps/cycle-check.Tpo -c -o cycle-check.o cycle-check.c |
26 |
| mv -f .deps/cycle-check.Tpo .deps/cycle-check.Po |
27 |
| gcc -std=gnu99 -I. -g -O2 -MT dirchownmod.o -MD -MP -MF |
28 |
| .deps/dirchownmod.Tpo -c -o dirchownmod.o dirchownmod.c |
29 |
| mv -f .deps/dirchownmod.Tpo .deps/dirchownmod.Po |
30 |
| gcc -std=gnu99 -I. -g -O2 -MT dirfd.o -MD -MP -MF .deps/dirfd.Tpo |
31 |
| -c -o dirfd.o dirfd.c |
32 |
| mv -f .deps/dirfd.Tpo .deps/dirfd.Pogcc -std=gnu99 -I. -g -O2 -MT |
33 |
| dirname.o -MD -MP -MF .deps/dirname.Tpo -c -o dirname.o dirname.c |
34 |
| mv -f .deps/dirname.Tpo .deps/dirname.Po |
35 |
| gcc -std=gnu99 -I. -g -O2 -MT dup-safer.o -MD -MP -MF |
36 |
| .deps/dup-safer.Tpo -c -o dup-safer.o dup-safer.c |
37 |
| dup-safer.c:1: error: expected '=', ',', ';', 'asm' or '__attribute__' |
38 |
| before '.' token |
39 |
| make[2]: *** [dup-safer.o] Error 1 |
40 |
| make[2]: Leaving directory `/home/k/benda/coreutils-6.9/lib' |
41 |
| make[1]: *** [all] Error 2 |
42 |
| make[1]: Leaving directory `/home/k/benda/coreutils-6.9/lib' |
43 |
| make: *** [all-recursive] Error 1 |
44 |
`---- |
45 |
|
46 |
the dup-safer.c file is very strange: |
47 |
|
48 |
,---- |
49 |
| $ cat lib/dup-safer.c |
50 |
| ccTvzRll.i.o : /tmp/ccTvzRll.i |
51 |
`---- |
52 |
|
53 |
I never saw this kind of c program. |
54 |
|
55 |
Given this box already have a very old version coreutils (in the age of |
56 |
gcc-3.3) in /opt/freeware/bin (IBM shipped GNU tools?), I decided just |
57 |
go ahead, ignoring the coreutils-6.9 failure. |
58 |
|
59 |
so I make config coreutils_SUPPLY = EXTERNAL and proceeded. |
60 |
|
61 |
Finally, prefix-launcher is done. (Hooray!) |
62 |
|
63 |
then I executed eprefix-bootstrap. |
64 |
|
65 |
But problems keeps emerging. (wel, I swear I did not $ emerge problems) |
66 |
|
67 |
One very annoying problem is, the portage system keeps using /bin/sh |
68 |
(which is a very very very ancient version of ksh of AIX) as the |
69 |
building SHELL. A fair envoke of "/bin/sh <some long command with |
70 |
long arguments can result in "Arg list too long" error, and booooom, |
71 |
emerge exits. |
72 |
|
73 |
I tried many times (set -xv in ebuild.sh, python -pdb with ebuild....) |
74 |
and could not find a way to tell portage to use my favorite bash |
75 |
compiled by prefix-launcher. (Dahhh) |
76 |
|
77 |
So each time there is an error, I have to spot which step goes wrong and |
78 |
open another shell (with modern bash of course) to execute the failed |
79 |
step. It works, just using a good SHELL. |
80 |
|
81 |
Hours of hours doing by hand. Now I am exhausted. I have to find a way |
82 |
to tell portage to use bash, and get rid of the f***ingly old sh. I am a |
83 |
normal user, I can't replace /bin/sh with the symlink to bash. (Dahhhh!) |
84 |
|
85 |
Any hints or ideas? |
86 |
|
87 |
Yours, |
88 |
-- |
89 |
XU Benda |
90 |
Research Center for Neutrino Science |
91 |
Tohoku University |
92 |
JAPAN |
93 |
|
94 |
http://www.awa.tohoku.ac.jp/~benda |