Re: Bug#341174: boost FTBFS on hppa
On Mon, Dec 19, 2005 at 01:29:05PM +0800, Randolph Chung wrote:
> >> Is anyone looking into this problem with building boost on hppa? As there
> >> are quite a few packages which build-depend on boost (including parts of
> >> KDE, and aptitude), this is likely to cause hppa to hold up the c2a
> >> transition unless some progress is made here.
> >
> > i'm seeing.. i tried some builds on paer.d.o, to see how new
> > compilers/linkers behavaed with respect to this. maybe latest binutils
> > 2.16.1cvs20051214-1 fixes this... i'm not doing more, sorry.
>
> This is a bug in gcc, but while it is being debugged some more, the
> toolchain actually suggested a workaround for you:
>
> /usr/bin/ld: [...] cannot reach
> 00000aae__ZSt10_ConstructIN5boost7archive6detail19basic_iarchive_impl7aobjectES4_EvPT_RKT0_+0,
> recompile with -ffunction-sections
>
> Try adding that flag (-ffunction-sections) to the gcc compile flags for
> hppa.
i just finished to re-check the build with this option. unfortunately it still
fails. the log is at http://people.debian.org/~cavok/boost_1.33.0-5_hppa.build.
> The problem is that a branch instruction on hppa has limited range
> (17-bits), so if the branch target is too large (this happens often with
> large problems, especially c++ code with lots of templates), the
> toolchain needs to add special stubs for far branches. gcc has logic to
> figure this out, but sometimes its calculations do not match what
> binutils calculates.
thank you for the first explanation i read about this error message.
cheers
domenico
-----[ Domenico Andreoli, aka cavok
--[ http://people.debian.org/~cavok/gpgkey.asc
---[ 3A0F 2F80 F79C 678A 8936 4FEE 0677 9033 A20E BC50
Reply to: