On Mon, Mar 10, 2008 at 09:02:16PM +0000, Jacob Meuser wrote:

> here's a new diff, with some improvements in the audio(4) interface,
> and an -arts FLAVOR.  yes, arts depends on esd, so you have to build
> a non-flavored esd, then arts, then -arts FLAVORed esd, and then
> install the FLAVORed esd with pkg_add -r -F installed esd-arts.  after
> all that, it has worked quite well in my testing.

that makes it possible to run esd as a client of artsd.

and, here's a patch to fix artsd as a client of esd.

in both cases, I suggest configuring artsd to use 2 buffers of 4096
bytes each.

-- 
[EMAIL PROTECTED]
SDF Public Access UNIX System - http://sdf.lonestar.org

Index: Makefile
===================================================================
RCS file: /home/cvs/OpenBSD/ports/x11/kde/arts3/Makefile,v
retrieving revision 1.63
diff -u -r1.63 Makefile
--- Makefile    24 Oct 2007 21:52:16 -0000      1.63
+++ Makefile    10 Mar 2008 21:02:40 -0000
@@ -5,7 +5,7 @@
 VERSION=               3.5.8
 DISTNAME=              arts-1.5.8
 MODKDE_VERSION?=       3.5.8
-PKGNAME=               ${DISTNAME}
+PKGNAME=               ${DISTNAME}p0
 
 SHARED_LIBS += mcop                 2.0      # .1.0
 SHARED_LIBS += artsflow_idl         2.0      # .1.0
Index: patches/patch-flow_audioioesd_cc
===================================================================
RCS file: patches/patch-flow_audioioesd_cc
diff -N patches/patch-flow_audioioesd_cc
--- /dev/null   1 Jan 1970 00:00:00 -0000
+++ patches/patch-flow_audioioesd_cc    10 Mar 2008 21:02:40 -0000
@@ -0,0 +1,16 @@
+$OpenBSD$
+--- flow/audioioesd.cc.orig    Mon Oct  8 02:47:09 2007
++++ flow/audioioesd.cc Sun Mar  9 22:52:42 2008
+@@ -204,10 +204,10 @@ int AudioIOESD::getParam(AudioParam p)
+               return write_fd;
+ 
+       case canRead:
+-              return ESD_BUF_SIZE;
++              return ESD_BUF_SIZE / 4;
+ 
+       case canWrite:
+-              return ESD_BUF_SIZE;
++              return ESD_BUF_SIZE / 4;
+ 
+               // ESD handles are actually socket descriptors, and I know not
+               // of any portable way to peek at the socket's send or receive

Reply via email to