MOO-cows Mailing List Archive
[Prev][Next][Index][Thread]
Re: PowerMac MOO
-
Date: Tue, 21 Nov 1995 16:35:39 PST
-
From: mpauker@peganet.com (Matt Pauker)
-
Content-Type: text/plain; charset="us-ascii"
At 3:54 PM 11/21/95, Theodore F. Vaida ][ wrote:
>>Anybody know of any development underway for a MOO server on the PowerMac?
>>
>>Nick
>
>I have been on a hiatus due to an overload of work but I am finally getting
>back to my MOO projects - the biggest is the Mac port of the server.
>rcn@netcom.com was the original MacMOO port author - he took the complete
>moo-1.7.6 source tree and created a MPW makefile, but it had some problems
>- namely a totally non mac interface and non mac friendly activity
>(faceless background app the refused to quit or release its memory,
>required the old and somewhat disfunctional ThreadsLib and was
>SLLLLOOOOOWWWW). I am currently trying to create a port that is capable of
>being compiled under MetroWerks Codewarrior 7, the only problem I to hack
>is the lack of a sockets interface - but the CWGUSI package ought to do,
>and the lack of a multi-process sytem - but the new threads package from
>Apple is nice and should be directly compaible with Copeland when it
>arrives. As a plus CodeWarrior can create FAT binaries - so I will let you
>know how thigns go. I am not sure how to import this stuff into the source
>tree and Pavel hasn't thought of anything either - a whole set of files and
>a totally different compile is required (CW Project files and a set of
>MacHeader files). Any ideas?
There are a number of problems involved in porting a Unix server to the Mac
- The main problem is obviously supporting MacTCP. You may be able to find
some packages out there to help with that, though (I don't know about the
CWGUSI package you mentioned). Also, consider supporting Open Transport -
It's much faster (PowerMac native) and is slowly replacing MacTCP.
Multi-processing shouldn't be a problem - The MOO server is only one
process anyways - Just make sure you support the Mac's cooperative
multitasking using WaitNextEvent().
As far as Copland goes, don't expect anything you write now to be
compatible later - The kernel is being completely rewritten, so calls will
most likely be changed. Also, Copland will use Open Transport rather than
MacTCP...
Matt Pauker
Home |
Subject Index |
Thread Index