Csound5.12.1 on 64-bit amd64cpu setup

lI lux-integ at btconnect.com
Wed Jun 23 18:08:36 PDT 2010

On Wednesday 23 June 2010 22:12:11 Ken Moffat wrote:
> On 23 June 2010 18:22, lI <lux-integ at btconnect.com> wrote:
> > Now to short-circuit things,  Could you let me know if you have built
> > Csound5.12.1 (or equivalent) on a pure 64bit clfs setup and if so could
> > you let me have your recipe?
>  This one pique'd my interest, because I remembered building an earlier
> version in the past (turned out to be last August - always good to keep
> notes :-) although I dropped it because I couldn't find any benefit from
> it in gnash, and I've since dropped gnash.
>  I've got to rebuild my current system anyway (pure64, LFS-svn as of last
> week *except* sticking with gcc-4.4.3) because my buildscripts were
> severely trashed by gnome-2.30 and I need to see if I've got the correct
> new build order (FWIW, I had to rebuild atk, pango, gtk2 because I had
> compiled them without introspection).  So, adding a couple more packages
> to this system (libsndfile-1.0.21 and Csound-5.12.1) is no problem.
>  All I can say is "builds and installs for me" :)  Libsndfile is a straight
>  CMMI (this time, it installed a static lib, so I got rid of that), and
>  then I built Csound
> with the following instructions based, I think, on what debian was doing
>  last year - at that time there was a suggestion it might need useWord64=1
>  but not, of course, useLib64=1 on pure64.  I didn't need that, merely
> scons prefix=/usr useJack=0 useFLTK=0 dynamicCsoundLibrary=1 \
>  BuildRelease=1 useALSA=1 LINKFLAGS='-Wl,-as-needed'
> python install.py --prefix=/usr

 I followed your suggestion and now everything compiles smoothly.  This 
includes the 'scons; options shipped with Csound5  save  buildCsoundVST=1 and  

I made a few variations
--I used the stuff from CVS  instead of Csound5.12.1
--I added  gcc4opt=k8 (for X86_64) optimizations
--I have stuff  in /opt so I edited customy.py  and  replaced the reference to 
/usr/linb for fltk

My problem seems to have been the custom.py file I was using.  I had created 
one from scratch  and there were clearly duplicated  PATHS therein (For 
example I had one pointing to /usr/local/include/FL  (for  fltk and this  also 
had a math.h  as well as the one in /usr/include.)

once again thanks for all the help


More information about the blfs-support mailing list