[Discuss-gnuradio] Quick Change to fusb_darwin

[Michael Dickens]

I made a quick change to the fusb_darwin code, to allow for the 
possibility that the write() command might get called with more data 
than the maximum allowed per transfer.  While I've never seen the 
number of incoming Bytes get that high, I'd rather have the code handle 
it correctly on the off chance it does happen to someone.  I also 
updated "mld_threads" to the latest version, which allows for switching 
between darwin's pthreads and omni_threads, but provides a consistent 
interface no matter which mechanism is used.

In the near future, I will be updating both gr-audio-osx and 
fusb_darwin for cosmetic changes (mostly changing variable names from 
"CamelCase" to "camel_case").  I will also be updating some of the 
buffer classes to allow for optional use of a mutex to control access 
to class internals (i.e. being "thread safe" or not).  For example, 
gr-audio-osx's code must already be thread-safe because of the way 
OSX/Darwin provides access to audio data, and thus having the buffer 
also be thread-safe is redundant and unnecessary (and only adds 
latency).  fusb_darwin, OTOH, is not inherently thread-safe and thus 
the buffer needs to provide this functionality.

Has any OSX user tried these codes out?  Any success compiling and/or 
executing? - MLD

ps> Has anyone successfully gotten GNURadio (and thus background 
libraries and includes) to install and execute on an Intel-Mac?  It was 
good to see the note about ".align 4" versus ".align 16", so I know 
that folks are making progress.  I hope to have access (remotely, 
without USRP hardware) to an Intel-Mac soon and will start working my 
way through the process.  I will update my install guide (2) with 
advice once I get it working, but I certainly welcome advice from 
others who have already done this or are in the process of doing it.

(2) Currently at < http://www.nd.edu/~mdickens/GNURadio/ >, but
hopefully soon to be moved somewhere on < http://sdr.nd.edu/ >.