[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Libunwind-devel] libunwind, autotools, and ARM?
From: |
Steffen Sledz |
Subject: |
Re: [Libunwind-devel] libunwind, autotools, and ARM? |
Date: |
Thu, 05 Aug 2010 08:15:59 +0200 |
User-agent: |
Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.9.2.7) Gecko/20100713 Lightning/1.0b2 Thunderbird/3.1.1 |
Am 05.08.2010 00:20, schrieb Arun Sharma:
>>>> AC_CHECK_LIB([unwind], [_Unwind_Backtrace], [], [ AC_MSG_ERROR([*** Could
>>>> not find libunwind ***]) ])
>>>
>>> _Unwind_Backtrace could be provided by other libraries as well. I'd
>>> test for something specific to libunwind (eg: unw_step).
>>
>> This is not a good idea. unw_step is not a symbol from the
>> library but only a define. So the check definitely fails.
>
> The problem is: if you configure libunwind with
> --enable_cxx_exceptions=no (which is the default on x86, arm,
> mips), you'll not have these symbols defined.
>
> True - unw_step gets rewritten as _UL${plat}_step (see
> tests/check-namespace.sh) - but why should it be a problem for
> writing an autoconf macro as long as you're including the right
> header file with #define unw_step ...?
That's not the way autoconf (especially configure) works. To check if a symbol
is available in a library it temporarily creates simple test source code
calling this symbol and tries to link with the library.
Steffen
- [Libunwind-devel] libunwind, autotools, and ARM?, Steffen Sledz, 2010/08/02
- Re: [Libunwind-devel] libunwind, autotools, and ARM?, Arun Sharma, 2010/08/02
- Re: [Libunwind-devel] libunwind, autotools, and ARM?, Steffen Sledz, 2010/08/03
- Re: [Libunwind-devel] libunwind, autotools, and ARM?, Arun Sharma, 2010/08/04
- Re: [Libunwind-devel] libunwind, autotools, and ARM?,
Steffen Sledz <=
- Re: [Libunwind-devel] libunwind, autotools, and ARM?, Arun Sharma, 2010/08/05
- Re: [Libunwind-devel] libunwind, autotools, and ARM?, Sven Neumann, 2010/08/05