| 1 |
On Thu, 06 Feb 2014 09:45:10 Fabian Groffen wrote: |
| 2 |
> All, |
| 3 |
> |
| 4 |
> I started last FOSDEM to give the bootstrap process some desparately |
| 5 |
> needed love again. |
| 6 |
> |
| 7 |
> Sofar, I've got bootstraps succeed on OpenIndiana x86 and x64. (yay) |
| 8 |
> |
| 9 |
> Currently, bootstrap is broken on: |
| 10 |
> - Darwin: it hangs on bash, which DEMANDS readline, despite disabling |
| 11 |
> it, I think the best course of action here is to actually compile |
| 12 |
> readline before bash, but I don't recall why we didn't do that |
| 13 |
> (probably because it used to break). |
| 14 |
> - Solaris 10: the old gcc compiler can't compile |
| 15 |
> baselayout-prefix/libeinfo. I'm about to patch that piece of code |
| 16 |
> tonight so it will compile. |
| 17 |
> - Fedora 17: Python 3 (whatelse) gives problems compiling modules. The |
| 18 |
> oddest one _decimal, which breaks on some weird funcs, but maybe those |
| 19 |
> are arithmetic ones from the compiler that we may or may not have due |
| 20 |
> to the bootstrap setup. Solution might be to block Python 3 during |
| 21 |
> bootstrap stages if it demands a newer compiler or something. |
| 22 |
> |
| 23 |
> I intend on running the bootstraps again on OpenIndiana once I fixed the |
| 24 |
> above issues. Any help is welcome of course. Especially now the |
| 25 |
> readline case for Darwin just needs people to try and see if that |
| 26 |
> results in problems for them. |
| 27 |
> |
| 28 |
> Over the last few days I've fixed quite some issues caused by implicit |
| 29 |
> dependencies. One reason I think Darwin currently breaks is because it |
| 30 |
> doesn't come with readline (but libedit) while most other systems /do/ |
| 31 |
> have readline installed. Similar reason I had a problem with bison on |
| 32 |
> Fedora 17 because its configure demanded perl, and Fedora doesn't |
| 33 |
> install that by default, while most other systems have it around... |
| 34 |
> |
| 35 |
Where is the current bootstrap script? I tried the one at |
| 36 |
dev.gentoo.org/~heroxbd/boostrap-rap.sh |
| 37 |
on a SLES11SP1 ppc64 install and it failed while emerging gcc, strangely |
| 38 |
it build gcc all right during the initial bootstrap. |
| 39 |
The failure is something I have seen numerous times while building gcc |
| 40 |
(not in a prefix) on that hardware (both aix and linux). Namely stage 2 |
| 41 |
of gcc fails over because it fails to find on the library stage 1 gcc |
| 42 |
was built against: |
| 43 |
|
| 44 |
make[3]: Leaving directory `/hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 45 |
devel/gcc-4.7.3/work/build/gcc' |
| 46 |
mkdir -p -- powerpc64-pc-linux-gnu/libgcc |
| 47 |
Checking multilib configuration for libgcc... |
| 48 |
Configuring stage 1 in powerpc64-pc-linux-gnu/libgcc |
| 49 |
configure: creating cache ./config.cache |
| 50 |
checking build system type... powerpc64-pc-linux-gnu |
| 51 |
checking host system type... powerpc64-pc-linux-gnu |
| 52 |
checking for --enable-version-specific-runtime-libs... no |
| 53 |
checking for a BSD-compatible install... /usr/bin/install -c |
| 54 |
checking for gawk... gawk |
| 55 |
checking for powerpc64-pc-linux-gnu-ar... |
| 56 |
/hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/bin/ar |
| 57 |
checking for powerpc64-pc-linux-gnu-lipo... lipo |
| 58 |
checking for powerpc64-pc-linux-gnu-nm... |
| 59 |
/hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 60 |
devel/gcc-4.7.3/work/build/./gcc/nm |
| 61 |
checking for powerpc64-pc-linux-gnu-ranlib... |
| 62 |
/hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/bin/ranlib |
| 63 |
checking for powerpc64-pc-linux-gnu-strip... |
| 64 |
/hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/bin/strip |
| 65 |
checking whether ln -s works... yes |
| 66 |
checking for powerpc64-pc-linux-gnu-gcc... |
| 67 |
/hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 68 |
devel/gcc-4.7.3/work/build/./gcc/xgcc - |
| 69 |
B/hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 70 |
devel/gcc-4.7.3/work/build/./gcc/ -B/hpc/scratch/frb15/gentoo/usr/powerpc64- |
| 71 |
pc-linux-gnu/bin/ -B/hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/lib/ |
| 72 |
-isystem /hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/include -isystem |
| 73 |
/hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/sys-include |
| 74 |
checking for suffix of object files... configure: error: in |
| 75 |
`/hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 76 |
devel/gcc-4.7.3/work/build/powerpc64-pc-linux-gnu/libgcc': |
| 77 |
configure: error: cannot compute suffix of object files: cannot compile |
| 78 |
See `config.log' for more details. |
| 79 |
|
| 80 |
|
| 81 |
And when you look at config.log in build/powerpc64-pc-linux-gnu/libgcc |
| 82 |
you find this: |
| 83 |
configure:3562: checking for suffix of object files |
| 84 |
configure:3584: /hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 85 |
devel/gcc-4.7.3/work/build/./gcc/xgcc -B/hpc/scratch/frb15/gentoo/var/tmp/p |
| 86 |
ortage/sys-devel/gcc-4.7.3/work/build/./gcc/ - |
| 87 |
B/hpc/scratch/frb15/gentoo/usr/powerpc64-pc-linux-gnu/bin/ - |
| 88 |
B/hpc/scratch/frb15/gentoo/usr/ |
| 89 |
powerpc64-pc-linux-gnu/lib/ -isystem /hpc/scratch/frb15/gentoo/usr/powerpc64- |
| 90 |
pc-linux-gnu/include -isystem /hpc/scratch/frb15/gentoo/usr/ |
| 91 |
powerpc64-pc-linux-gnu/sys-include -c -g -O2 -pipe conftest.c >&5 |
| 92 |
/hpc/scratch/frb15/gentoo/var/tmp/portage/sys- |
| 93 |
devel/gcc-4.7.3/work/build/./gcc/cc1: error while loading shared libraries: |
| 94 |
libmpc.so.3: ca |
| 95 |
nnot open shared object file: No such file or directory |
| 96 |
|
| 97 |
|
| 98 |
I have seen this so many time while building gcc on that hardware that I have |
| 99 |
taken to build it against static gmp/mpfr/mpc/{ppl,isl}/cloog. |
| 100 |
|
| 101 |
Francois |
Archives