Hey guys, Dan Potter has posted the newest version of KOS on sourceforge! Here is the message from dan potter himself:

I’ve posted KOS 1.1.4 on SourceForge:

http://sf.net/projects/cadcdev/

You can find the release notes and change log here:

http://sourceforge.net/project/shownotes.php?group_id=23791&release_id=56054

If you’re doing any development using KOS, I highly recommend getting
this release unless you’re already using the CVS release, especially if
you’re doing anything with threads. Here’s the relevant part of the
CHANGELOG file:

– DC Store queue support [Andrew Kieschnick == ADK]
– DC New and improved matrix math routines [ADK]
– DC Patch to KGL to clip polygons with bad w value [ADK]
– DC Patch for serial I/O that does not reset the baud rate if it’s already
been set. This should eliminate a lot of problems with mismatched
serial baud in dcload [ADK]
– DC Patch to fix GL_QUADS constant [Greg Cooksey]
– DC ftan support, plus KGL usage [ADK]
– DC Fix for iso_ioctl
– DC New spinlock code works a lot better [ADK]
– DC Fixed a bug where thd_enabled wasn’t being set to 1 [ADK]
– DC Render done seems to work again, so it’s now re-enabled [ADK]
– DC Added ‘menu’ command in KOSH to exit to the DC menus [Dan Potter == DP]
– DC Added rtc_unix_secs() support for reading the current date/time [DP]
– DC Fixed some problems in VMU writing related to timing and a mystery
command (thanks to an oooold message from Nagra) and also did some
fixes to fs_vmu. Please see the notes in README, this code still
isn’t really stable. [DP]
– DC Added a new ‘hello’ example in the DC examples tree which shows
a basic project skeleton which you can start with. [DP]

Also, I’d like to specifically mention (as I do briefly in the README
file) that although I’ve gotten fs_vmu to be minimally working, I don’t
recommend serious usage of it. I got to seriously looking at it and
that’s some pretty gross code =). It’s not thread safe, parts of it
might not even work (or worse, mysteriously corrupt parts of your VMU,
or who knows what else). The previous versions had things like
re-allocating write buffers to be the same size and then promptly
writing over them and such… eek! =D Anyway, since VMU access is one
thing that’s prominently broken in KOS I’m probably going to work on
this in the next few versions. Probably I’ll seperate it into two parts;
one of them will work somewhat traditionally, loading or saving an
entire file at once, and a VFS-compatible part that uses ioctl’s and all
that.

Ya, a little long, but oh well =) -pSyCo