robertl@bucsf.bu.edu (Robert La Ferla) (02/02/90)
In article <90031.143256UH2@PSUVM.BITNET> UH2@psuvm.psu.edu (Lee Sailer) writes: > How much does it matter whether there is ever a 680x0? Does Mach really > take any special features of the 68000 line? Won't the main problem just be > a new compiler? If only life was that simple... If NeXT were to go to the Motorola 88k series any code with variable arguments is likely to break since the 88000 puts these in registers. How do you take an address of a register? You can't. I don't know if you could make a compiler smart enough to address this issue but in any case, I bet it wouldn't be easy. It would be wonderful to see a NeXT running with an 88000. __ / \ / __/_ /___/ __ /_ __ __ / INTERNET: robertl%bucsf@bu-it.bu.edu / \ '_' /_/ |_- / ' / BITNET: mete0pc@buacca.bu.edu
rick@hanauma.stanford.edu (Richard Ottolini) (02/02/90)
In article <ROBERTL.90Feb2013012@bucsf.bu.edu> robertl@bucsf.bu.edu (Robert La Ferla) writes: >If only life was that simple... If NeXT were to go to the Motorola 88k series >any code with variable arguments is likely to break since the 88000 puts these >in registers. How do you take an address of a register? Read varargs.h of a MIPS-chip based system and you see how. It is complicated, but works. Address of subroutine arguments don't mean anything in C because it is pass by value. The value can be an address such as an array.
chris@island.uu.net (Chris King) (02/10/90)
In article <ROBERTL.90Feb2013012@bucsf.bu.edu> robertl@bucsf.bu.edu (Robert La Ferla) writes: >In article <90031.143256UH2@PSUVM.BITNET> UH2@psuvm.psu.edu (Lee Sailer) writes: ->> How much does it matter whether there is ever a 680x0? Does Mach really >>> __ > / \ / __/_ > /___/ __ /_ __ __ / INTERNET: robertl%bucsf@bu-it.bu.edu >/ \ '_' /_/ |_- / ' / BITNET: mete0pc@buacca.bu.edu Newsgroups: comp.sys.next Subject: Re: next the next NeXT Summary: Expires: References: <1990Jan27.195239.6953@eng.umd.edu> <1990Jan29.051444.26588@nueces.cactus.org> <194@atncpc.UUCP> <90031.143256UH2@PSUVM.BITNET> <ROBERTL.90Feb2013012@bucsf.bu.edu> Sender: Reply-To: chris@island.uu.net (Chris King) Followup-To: Distribution: Organization: Island Graphics, Marin County, California Keywords: In article <ROBERTL.90Feb2013012@bucsf.bu.edu> robertl@bucsf.bu.edu (Robert La Ferla) writes: >In article <90031.143256UH2@PSUVM.BITNET> UH2@psuvm.psu.edu (Lee Sailer) writes: >> How much does it matter whether there is ever a 680x0? Does Mach really >> take any special features of the 68000 line? Won't the main problem just be >> a new compiler? > >If only life was that simple... If NeXT were to go to the Motorola 88k series >any code with variable arguments is likely to break since the 88000 puts these >in registers. How do you take an address of a register? You can't. I don't >know if you could make a compiler smart enough to address this issue but in >any case, I bet it wouldn't be easy. It would be wonderful to see a NeXT >running with an 88000. > __ > / \ / __/_ > /___/ __ /_ __ __ / INTERNET: robertl%bucsf@bu-it.bu.edu >/ \ '_' /_/ |_- / ' / BITNET: mete0pc@buacca.bu.edu Newsgroups: comp.sys.next Subject: Re: next the next NeXT Summary: Expires: References: <1990Jan27.195239.6953@eng.umd.edu> <1990Jan29.051444.26588@nueces.cactus.org> <194@atncpc.UUCP> <90031.143256UH2@PSUVM.BITNET> <ROBERTL.90Feb2013012@bucsf.bu.edu> Sender: Reply-To: chris@island.uu.net (Chris King) Followup-To: Distribution: Organization: Island Graphics, Marin County, California Keywords: In article <ROBERTL.90Feb2013012@bucsf.bu.edu> robertl@bucsf.bu.edu (Robert La Ferla) writes: >In article <90031.143256UH2@PSUVM.BITNET> UH2@psuvm.psu.edu (Lee Sailer) writes: ->> How much does it matter whether there is ever a 680x0? Does Mach really ->>ake any special features of the 68000 line? Won't the main problem just be ->>a new compiler? -> ->If only life was that simple... If NeXT were to go to the Motorola 88k series ->any code with variable argments is likely to break since the 88000 puts these ->in registers. How do you take an address of a register? You can't. I don't ->know if you could make a compiler smart enough to address this issue but in ->any case, I bet it wouldn't be easy. It would be wonderful to see a NeXT ->running with an 88000. Sun Sparcs use arguments in registers. Using the stdarg or vararg macro set, source code compatibiltiy between Sun 3s and 4s is not to much of a problem. I think the same applys here. Chris King