From: Henry W. <ha...@nc...> - 2001-03-12 04:55:24
|
Bill Fink wrote: > > > On 03/11/2001, Henry Worth wrote: > > > > Check your ulimits, I get a core dump frequently, but > > gdb (5.0) doesn't seem able to understand it. But running > > from within gdb does work. > > Do you mean the csh/tcsh limits. My coredumpsize is unlimited. Yep, also watch the -omit-frame-pointer compile option, even if you get a dump, you may not be able to do much with it and the aggressive levels of the default optimization will make a mess of source-level debugging. > > For the "#define HAVE_BYTESWAP_H", I was going to borrow code from > oms that checks this. Forgot to address that one, but the standard configure file checks for headers should work, HAVE_BYTESWAP_H pops up in lots of packages. >For the "#define AWACS_AUDIO", I was going > to check the output of "cat /dev/sndstat". And then there needs > to be a way not to include the libw32dll stuff in the Makefile. > At this point I just have a SOFT_AUDIO_SYNC define that includes the soft sync code, but it only gets used at runtime if tests of the realtime cap fails. Limiting of the sample rate to 44.1K is by a FORCE_44K_MAX define. Configure only defines these for ppc and I was planning on adding configure options for them. Testing for AWACS on the build machine will cause problems for distributing binaries built on the PMac's with other sound chips (iBooks and the latest G4's, at least). I think there is a PMAC define that could be used instead so that the other ppc archs don't get the code by default. I've handled the libw32dll and corba lib builds with automake conditionals. Corba was already there, just needed to have Makefile.am fixed -- if it needs to be disabled, it should be done in configure.in. |