Re: [yoshimi-user] a little something for the weekend

  • From: Will J Godfrey <WillGodfrey@xxxxxxxxxxxxxxx>
  • To: yoshimi-user@xxxxxxxxxxxxxxxxxxxxx
  • Date: Sun, 18 Jul 2010 19:51:00 +0100

On Fri, 16 Jul 2010 19:08:44 +1000
cal <cal@xxxxxxxxxxxx> wrote:

I've tried to merge a couple of elements from the experimental stuff back
into the 0.058++ branch. The main feature is the use of boost smart pointers
and memory pool to take as many of the new/delete ops as possible out of the
realtime path. I think it has promise. Implicit in that ia a new dependency
on the boost development headers. On debian I believe the relevant package
is libboost-dev.

Jack session support is included, but the default setting in CMakeLists.txt
is 'No', with a dependency on jack>=0.115.6 (as it's been for ages). If you
want jack session support, you'll need to visit ccmake ('cd src; cmake ..'),
or set the appropriate option via command line feed to the cmake run.

There's a couple of other minor tweaks, so this one is by definition
experimental, and the 'rc1' tag means exactly what it implies.

<http://www.graggrag.com/yoshimi/yoshimi-0.059-rc1.tar.bz2>

cheers, Cal

Strange problem, most voice patches load without problems but the
following ones instantly cause yoshi to silently run away :(

Hyper Matrix
Hyper Pad

I suspect this is some obscure memory allocation problem, because these
are rather big patches and if you cycle through lots of other patches
first, then these one will load.

Needless to say, this doesn't occur with any other versions of yoshi.

... Otherwise, processor use looks good.

--
Will J Godfrey
http://www.musically.me.uk
Say you have a poem and I have a tune.
Exchange them and we can both have a poem, a tune, and a song.


Other related posts: