[ardour-dev] 2.0.2 fixup release

Fernando Lopez-Lezcano nando at ccrma.Stanford.EDU
Wed May 9 19:01:48 PDT 2007


On Wed, 2007-05-09 at 18:50 -0700, Fernando Lopez-Lezcano wrote:
> On Wed, 2007-05-09 at 18:42 -0700, Fernando Lopez-Lezcano wrote:
> > On Wed, 2007-05-09 at 16:59 -0400, Paul Davis wrote:
> > > There was an error in the source tarball for 2.0.1. This error did NOT
> > > affect the OS X DMG. 2.0.2 has been released as a correction for the
> > > error - it does not change any listed fixes for 2.0.1 in any way.
> > 
> > Has there been a change in how cpus are detected during the build
> > process? With the same spec file I used with 2.0.0 for building Planet
> > CCRMA packages I get this (on 32 bit fc6):
> > 
> > g++ -c -fPIC libs/ardour/sse_functions_64bit.s -o
> > libs/ardour/sse_functions_64bit.os
> > libs/ardour/sse_functions_64bit.s: Assembler messages:
> > libs/ardour/sse_functions_64bit.s:37: Error: bad register name `%rbp'
> > libs/ardour/sse_functions_64bit.s:38: Error: bad register name `%rsp'
> > libs/ardour/sse_functions_64bit.s:41: Error: bad register name `%rbx'
> > libs/ardour/sse_functions_64bit.s:42: Error: bad register name `%rdi'
> > libs/ardour/sse_functions_64bit.s:43: Error: bad register name `%rsi'
> > libs/ardour/sse_functions_64bit.s:46: Error: bad register name `%rdx'
> > libs/ardour/sse_functions_64bit.s:51: Error: bad register name `%rdi'
> > 
> > (I did not get to try with 2.0.1... :-)
> 
> Never mind, DIST_TARGET is (I think) what I needed. 
> Still, it did work on 2.0.0...

One more question, I see that:

      USE_SSE_EVERYWHERE
      BUILD_SSE_OPTIMIZATIONS

are gone. What does this mean? Are all sse/3dnow optimizations dynamic
by default? (meaning the proper routine is chosen at runtime based on
the cpu ardour is running on?)

-- Fernando





More information about the Ardour-Dev mailing list