Bug#342267: gcc-4.0: boost 1.33.x FTBFS with "cannot handle R_PARISC_PCREL17F..."

[Domenico Andreoli <cavok@tiscali.it>]

On Tue, Dec 06, 2005 at 05:41:14PM +0100, Matthias Klose wrote:
> did you check the suggestion from the error message? please notice
> that the error message comes from the linker (binutils).

-ffunction-sections? yes, i did some time ago but nothing changed.
i already submitted a bug report to binutils, #334497.

i'd be happy to try with gcc 4.1, but i suppose it will not enter
unstable before gcc 4.0 transition is finished.

i'm now building boost 1.33.1 for experimental. then i will try to build
it on paer.d.o, but i already know it is going to fail, recently i tried
from upstream cvs. i will then try again with -ffunction-sections to
be sure it really does not solve the problem.

i also tried both gcc 3.4 and 4.0 with different levels of optimization,
but again nothing changed.

it looks like something deep in gcc vs. ld relationship. boost
1.32.0-6.1 built with binutils 2.16.1-2 and gcc 4.0.1-2. but
boost 1.32.0+1.33.0-cvs20050727-1 cvs snapshot already broke with
binutils 2.16.1-2 and gcc 4.0.1-3. it again built with binutils
2.16.1cvs20050902-1 and gcc 3.4.4-9 but suddenly broke again with gcc
3.4.4-10 and binutils 2.16.1cvs20050902-1

whatever bug/feature makes gcc fail the build, it entered in debian
with gcc 4.0.1-3, gcc 3.4.4-10 or with some broken binutils. if i could
install packages on paer.d.o, i surely would try different combinations
of linker/compiler.

cheers
domenico

-----[ Domenico Andreoli, aka cavok
 --[ http://people.debian.org/~cavok/gpgkey.asc
   ---[ 3A0F 2F80 F79C 678A 8936  4FEE 0677 9033 A20E BC50