johnm@trsvax.UUCP (02/24/88)
If this question goes unanswered for a second time then I'll restart the PDP-11 discussion! Does anybody know if the C compiler that comes with the Atari ST port of Minix will compile programs larger than 64K? If so then it seems like the IBM and ST versions of Minix are going to diverge radically. If not then there doesn't seem to be much point in my buying Minix. Please say something...
wigle@canisius.UUCP (Ken Wigle) (03/01/88)
Not much help, but I did not want a re-opening of the PDP-11
discussion!
Good luck in finding an answer - one that is helpful
--
USMAIL: Ken Wigle / Canisius College / 2001 Main St. / Buffalo,NY 14208
UUCP: ...!{decvax|watmath|allegra|rocksvax}!sunybcs!canisius!wigle
" ...!ames!canisius!wigle
BITNET: CNSKMW@CANISIUSjohnm@trsvax.UUCP (03/04/88)
>>If this question goes unanswered for a second time then I'll restart the >>PDP-11 discussion! > >Get a life! I hope someone will answer your question, but your manner >of putting it is childish. It seems apparent that the smiley was invented for people too simple to understand when one is implied. Obviously, I am not going to (nor could I) restart the PDP discussion, it was merely a way to draw attention to the fact that it was a question that needed answering. Fortunately two people did so (Andy Tanenbaum says he answered it the first time but we never saw it at this site. Sorry Andy...) >>If not then there doesn't seem to be much point in my buying Minix. > >Aw ... gee ... ain't that *awful*? Apparently the rest of my question went right by you. The remainder of it was a concern that if the two versions of Minix differed on this point (the size of the executables that they could create) that there could be a drastic split in the programs available for the two systems. The ST version of Minix with its large executables can support a lot of programs that are a real porting nightmare otherwise. If this is the kind of response we can expect then I would hope that the ST owners don't lose any sleep over not having their programs run on the IBM Minix.
paradis@encore.UUCP (Jim Paradis) (03/09/88)
In article <207600004@trsvax> johnm@trsvax.UUCP writes: >Apparently the rest of my question went right by you. The remainder of it >was a concern that if the two versions of Minix differed on this point (the >size of the executables that they could create) that there could be a drastic >split in the programs available for the two systems. The ST version of Minix >with its large executables can support a lot of programs that are a real >porting nightmare otherwise. If this is the kind of response we can expect >then I would hope that the ST owners don't lose any sleep over not having >their programs run on the IBM Minix. I've thought about this myself, having more than once experienced frustration at downloading some UNIX goodie off the net only to have it be far too big to squeeze into 64K I+D. Two things are required to break this barrier: an OS that supports an arbitrary number of text and data segments per process, and a compiler that can produce large-model code. The former is not all that difficult to do, but the latter is more of a problem. Anyone out there with ACK who might be able to tackle the latter problem? (Andy has told me that you need all of the ACK tools to produce the back-end; the source provided with the 1.2 compiler is NOT sufficient!) +----------------+ Jim Paradis linus--+ +--+-------------+ | Encore Computer Corp. necntc--| | | E N C O R E | | 257 Cedar Hill St. ihnp4--+-encore!paradis | +-------------+--+ Marlboro MA 01752 decvax--| +----------------+ (617) 460-0500 talcott--+ Well, what's the pleasure in THAT??!!
jsp@sp7040.UUCP (John Peters) (03/15/88)
In article <207600004@trsvax>, johnm@trsvax.UUCP writes: > > Apparently the rest of my question went right by you. The remainder of it > was a concern that if the two versions of Minix differed on this point (the > size of the executables that they could create) that there could be a drastic > split in the programs available for the two systems. The ST version of Minix > with its large executables can support a lot of programs that are a real > porting nightmare otherwise. If this is the kind of response we can expect > then I would hope that the ST owners don't lose any sleep over not having > their programs run on the IBM Minix. This is indeed going to be a real problem. However I can no answer to the problem. It really lies with the klude of 64k segments (yell at intel). As for progrems I write on ST-Minix, we will have to wait and see. No on to the real discussion. Who would be in charge of trying to keep the two versions the same. Here again I see no real answer. The two machines are going to diverse somewhat. I'm sure that somebody out there has the knowlege to write a compiler for PC-Minix that will generate large memory model code for the 80286 and port minix to a true '286 version. And as for the '386 port the problem should not exist. I can forsee there being as many ports of Minix as UNIX (every one has something different) however, they will (or should) all still have the same feel on the command line interface and that is about as good as it gets in this industry. -- Johnnie -- My opinions are my own and usually mis-spelled. I expect to get flamed but realize that I will try to write code that is portable across Minix machines. However I refuse to be limited because somebody else bought a machine that I can't stand. I'm of the opinion that segments are for worms.
ralphw@IUS3.IUS.CS.CMU.EDU (Ralph Hyre) (04/02/88)
In article <294@sp7040.UUCP> jsp@sp7040.UUCP (John Peters) writes: >In article <207600004@trsvax>, johnm@trsvax.UUCP writes: >>. this is the kind of response we can expect >> then I would hope that the ST owners don't lose any sleep over not having >> their programs run on the IBM Minix. > > This is indeed going to be a real problem. However I can no answer >to the problem. It really lies with the klude of 64k segments (yell at intel). Well, yell at more people than Intel, most processors I know of have some sort of segmentation warts on them. In the 68000 world, for example, parts of the Mac 'OS' limit code segments to 32K, presumably to provide for easy-to-relocate code. (I suppose you can use relative branches that way.) If I were to try to port Minix to a machine with 16 bit addresses (and bank-switched memory), it would also effectively have 64k segments. I think the Minix ST port (and other ports) should make the developer do something explicit to override the 64k segment limitation (like a special compiler switch), because it will be a win when you add swapping & paging anyway, since you will be able to shove code anywhere in the address space (modulo 64k). The problem of trying to port code between the Unix variants will also happen with Minix, unless someone (Andy?) takes a stand. It would be a shame. -- - Ralph W. Hyre, Jr. Internet: ralphw@ius2.cs.cmu.edu Phone:(412)268-{2847,3275} CMU-{BUGS,DARK} Amateur Packet Radio: N3FGW@W2XO, or c/o W3VC, CMU Radio Club, Pittsburgh, PA -- - Ralph W. Hyre, Jr. Internet: ralphw@ius2.cs.cmu.edu Phone:(412)268-{2847,3275} CMU-{BUGS,DARK} Amateur Packet Radio: N3FGW@W2XO, or c/o W3VC, CMU Radio Club, Pittsburgh, PA
wes@obie.UUCP (Barnacle Wes) (04/04/88)
In article <1280@PT.CS.CMU.EDU>, ralphw@IUS3.IUS.CS.CMU.EDU (Ralph Hyre) writes: [...much blathering about 64k segment sizes, paging, etc...] > The problem of trying to port code between the Unix variants will also happen > with Minix, unless someone (Andy?) takes a stand. It would be a shame. No, what would really be a shame would be forcing the stupid limitations of the Intel x86 architecture on those of us bright enough to avoid it like the plague. Arbitrarily making the Minix compiler create 64k segment limitations for M68000, NS32000, and even the 386 architecture would be more of a shame. I am sure it would also be "fixed" quickly by some intrepid hacker :-). For those of you still struggling with the 8088 or 286, the answer is simple: buy a processor that does not have these inherent limitations, or learn to live with them. Your argument sounds like a lot of the arguments I have heard in support of the 55 mph speed limit in the USA: "I am not safe driving my car faster than 55, so you should ALL have to drive 55." -- /\ - "Against Stupidity, - {backbones}! /\/\ . /\ - The Gods Themselves - utah-cs!utah-gr! / \/ \/\/ \ - Contend in Vain." - uplherc!sp7040! / U i n T e c h \ - Schiller - obie!wes