[PD-dev] [pure-data:bugs] #1185 configure.ac script changes for fftw3

Ticket 1185 1185 at bugs.pure-data.p.re.sf.net
Mon Jan 19 03:00:44 CET 2015




---

** [bugs:#1185] configure.ac script changes for fftw3**

**Status:** open
**Group:** v0.46
**Created:** Mon Jan 19, 2015 02:00 AM UTC by Anonymous
**Last Updated:** Mon Jan 19, 2015 02:00 AM UTC
**Owner:** nobody

This patch updates the configure.ac file to check the library for 32-bit floating point fftw3.

The function fftwf_execute occurs in pd and in fftw3 but not in fftw2.  In the 64-bit version of fftw, the function occurs as "fftw_execute" instead.

So it's safe to check against, and precision specific.  The autoconfigure test will write a short program declaring the function header:
char fftwf_execute();

and testing that the program links correctly with the fftw library


--- pd-0.45-4.orig/configure.ac	2013-11-22 10:43:24.000000000 -0600
+++ pd-0.45-4/configure.ac	2015-01-18 19:43:06.378612708 -0600
@@ -212,13 +212,13 @@
 AM_CONDITIONAL(PORTMIDI, test x$portmidi = xyes)
 if test x$portmidi = xyes; then echo Using included portmidi; fi
 
-dnl fftw v2
+dnl fftw v3
 AC_ARG_ENABLE([fftw],
 	[AS_HELP_STRING([--enable-fftw],
 		[use FFTW package])],
     [fftw=$enableval])
 if test x$fftw = xyes; then
-    AC_CHECK_LIB(fftw, fftw_one, [LIBS="$LIBS -lfftw"],
+    AC_CHECK_LIB(fftw3f, fftwf_execute, [LIBS="$LIBS -lfftw3f"],
         [AC_MSG_NOTICE([fftw package not found - using built-in FFT]); fftw=no])
 fi
 AM_CONDITIONAL(FFTW, test x$fftw = xyes)



---

Sent from sourceforge.net because pd-dev at lists.iem.at is subscribed to http://sourceforge.net/p/pure-data/bugs/

To unsubscribe from further messages, a project admin can change settings at http://sourceforge.net/p/pure-data/admin/bugs/options.  Or, if this is a mailing list, you can unsubscribe from the mailing list.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-dev/attachments/20150119/623226e6/attachment.html>


More information about the Pd-dev mailing list