[nas] patch: output gain scaling

[Paul Fox]

jon wrote:
 > On Mon, 24 Jul 2006, Paul Fox wrote:
 > >
 > > sure.  i thought of that, but figured that since we're already in
 > > trouble with "gain", i wouldn't worry about it.  but that was
 > > just laziness.  how about "outputscale"?
 > >
 > 
 >          Duh, you are correct.  :)  I thought this was a global option,
 >          and not per-section.  Applied and comitted.

thanks!

 >          I was trying to determine the cause of a coredump I am
 >          seeing.  I notice your patches do not affect it, so I will
 >          apply anyway and debug further when I get some time.
 > 
 >          Seems we are corrupting the stack somewhere in
 >          createServerComponents.  I can only avoid the core by
 >          compiling auvoxware.c with '-g'.  With the default -O3, it
 >          segv's (tries to execute NULL stack) right when CreateComponents
 >          returns.  Not sure when that started though.
 > 
 >          Anybody else see this? I am using gcc 3.4.1-4mdk.

i might be.  i was building last night on my aging server
(originally RH 7.2, w/ gcc 2.96) and moving the nasd binary to a
debian etch (testing) machine to run it (it's the machine
connected to the stereo, but it doesn't have a full devkit
installed), and was getting a core dump.  i suspected library
mismatch, or other voodoo, and switched the test builds to
another etch machine (gcc 4.0.4), and the core dump went away.

i can investigate further.

speaking of debian -- when upgrading, the current debian
installer for nas tries to preserve the current "releasedevice"
and "mixerinit" values, but everything else comes straight from
the release.  this means any other changes to the nas config get
overwritten on an upgrade.  seems like it might be time for the
nas package to start md5'ing the config in order to detect any user
changes.  is the debian maintainer on the list?

paul
=---------------------
 paul fox, pgf at foxharp.boston.ma.us (arlington, ma, where it's 65.8 degrees)