Re: libunwind detection problem

[Daniel Kegel <dank at kegel dot com>]

Josh Parsons wrote:
> I'm working with a crosstool-0.28-rc37 / gcc-3.3.5 toolchain for armv5b-
> softfloat and discovered that when the toolchain is built on a build
> system with libunwind, the target system is assumed to have libunwind as
> well.
>
> The culprit seems to be the libunwind detection code in libstdc++-v3's
> configure script, and I have been able to work around the problem by
> modifying armv5b-softfloat.dat to add --disable-libunwind-exceptions to
> GCC_EXTRA_CONFIG.
>
> This seems as though it shouldn't be necessary, though - gcc shouldn't
> be assuming that the target system has libunwind just because the build
> system does.
>
> Has anyone else had trouble with this? Should it be regarded as a
> crosstool bug, a gcc bug, or neither?

Hrm.  Probably a gcc bug, but I'm biased :-)
(crosstool should at least show people how to work around it.)

Can you check to see if this bug is present in gcc-4.0.1, and if so,
report it to gcc's bugzilla?

Thanks!
- Dan


------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com