[vsipl++] [patch] Fix quickstart doc typos, fix configure --disable-fft

[Jules Bergmann]

Mark Mitchell wrote:
> Jules Bergmann wrote:
> 
>> It also fixes configure.ac to recognize --disable-fft as a synonym for
>> --enable-fft= (with no backend).
> 
> Personally, I'd make --disable-fft an error.  Or, if that bothers you,
> I'd make it a synonym for no_fft.
> 

That doesn't bother me.  If we do that, we should rename the option 
'--with-fft' since the current '--enable-fft' implies that it could be 
disabled.

At a high-level, configure would behave like:

  - by default, the library provides FFT support, using the builtin FFT
    backend and perhaps SAL if we detect its existence

    i.e. on most platforms the following option is implied

	--with-fft=builtin

    and on mercury platforms the following option is implied

	--with-fft=sal,builtin

  - users that want to use a different backend can specify it manually:

    for example:

	--with-fft=ipp

    or

	--with-fft=ipp,fftw3

    Optionally, we can have infer which backends to use from the
    options indicating the presence of libraries
    such as IPP.

  - power users that want to disable fft can do so either by specifying
    no backends:

	--with-fft=

    (which causes FFT usage to fail at compile-time)

    or by specifying the empty backend:

	--with-fft=no_fft


How does this sound?

				-- Jules

-- 
Jules Bergmann
CodeSourcery
jules at codesourcery.com
(650) 331-3385 x705