Is it possible that at least thigns like fixing the endian-ness of network code be submitted to the svn as a patch?
Last time I tried it was rejected.. was apparently "unnecessary" at the time..
This mac port relies on 1 person alone who is only here a few days of the year and is fallign hopelessly behind.
Actually, I'm getting up to date, faster than there are changes added; don't make assumptions because they sound good.
And the refusal to work with the svn means that the people who finally do use the mac port are forced into a userbase of 4 or 5 people, because they cant play with linux and windows people, precisely because the necessary changes to support mac havent been committed to the svn and have been left out, OR because the linux/windows users are running a later version of spring than is available to the mac users.
Or rather that there's just some major incompatibilities with how the network code is written. It is that architecturally dependant that there's that big a network sync problem. Even g5s can't sync with g4s without moving to a slower compatibility mode. It is possible to get the network sync fine with intel macs and windows/linux; but ppc macs are what I have and my primary target; intel should "just work" bar a few minor changes to take into account apple's network stack.
RE g4 build: Spring is that heavy on the memory bus that it runs really slow on g4s, even if the cpu's running comfortably, thus focused on making the g5 build run faster, at g4 compatibility's expense. At the time there was little interest in a g4 build but now the interest's there, I'll focus on getting up to date first, then give you a build, and we'll go from there.
the build that's online does infact use 64 bit numbers, and set a requirement for g5 for loading, that's probably why it bounces once or twice then quits.
Why would you support Mac Intel chips when they can just run windows natively? I thought this thread was about a PPC version of spring!!! I don't think low bus bandwidth of the PPC chip would hurt performance.
Why? because apple does have some useful tools to make it easy to port, and it helps catch and explain the loads of endian issues that pop up (if it works on an intel mac, but not ppc, then it's most likely an endian issue at that point. Besides, isn't windows expensive?
It's the low bus bandwidth of the g4; the g5's fine; here it's my graphics card and low integer optimisations that slow spring down; bi directional 1ghz+ buses are useful; g4s even the 1.5ghz ones are still sub 500mhz, so yes, with how spring runs, this is a problem.
Just have some patience people, it's not far from current up to date. There were some rather big changes that I need to step through, so it's taken time, but it's working :)