Linux X demo
in [ASM]
|
Prev: masm linking from console
Next: NASM HelloWorld - DOS
From: Rod Pemberton on 21 Jul 2007 04:14 "Wolfgang Kern" <nowhere(a)never.at> wrote in message news:f7q1sm$ek5$4(a)newsreader2.utanet.at... > > I once used a similar (more Zilog Z-80 based) syntax on my HEX_WORK85 > disassembler/Hex-editor for several different CPU/MC/MP families. > But I also have some trouble to read through Herberts code when it > isn't accompanied by hexadecimal machine-code. > Sorry to tell you, but I tried your HexTutor about six months ago. It blew up... Rod Pemberton
From: Wolfgang Kern on 21 Jul 2007 05:04 Rod Pemberton mentioned: [] > Sorry to tell you, but I tried your HexTutor about six months ago. > It blew up... How ? Wich trigger did you pull ? It should work on win98 and XP. Ok, the demo still got a few known syntax and sizecast bugs, but 'blew up'? __ wolfgang
From: Rod Pemberton on 21 Jul 2007 06:27 "santosh" <santosh.k83(a)gmail.com> wrote in message news:f7ql53$c8c$1(a)aioe.org... > > case 0x05: {la[--j] = la[j] / la[j+1]; break;} > > > case 0x05: > { > la[j-1] = la[j] / la[j+1]; > j--; > break; > } > 1) K&R C and the early DMR C compilers aren't LALR(1). 2) Most "modern" (old today...) C compilers are based on Steve Johnson's 1975 (or so) Portable C Compiler. PCC is YACC based and LALR(1). 3) PCC replaced DMR in most versions of UNIX. 4) Typedef's were added to C in 1976, and aren't LALR(1)... 5) Although not required by the standards, most people intuitively understand C to be LALR(1) (which is due to PCC's strong influence on later compilers). 6) The BNF grammars in the various C standards are YACC-able BNF, i.e., LALR(1). 7) This history is largely lost... I.e., assuming LALR(1) and applying precedence, the above is: case 0x05: { --j; la[j] = la[j] / la[j+1]; break; } It appears that DJGPP (GCC) works properly while OpenWatcom v1.3 fails. I'd guess that most K&R style compilers work. Rod Pemberton
From: Rod Pemberton on 21 Jul 2007 06:50 "Wolfgang Kern" <nowhere(a)never.at> wrote in message news:f7sioq$bgv$1(a)newsreader2.utanet.at... > > Rod Pemberton mentioned: > [] > > Sorry to tell you, but I tried your HexTutor about six months ago. > > It blew up... > > How ? Wich trigger did you pull ? > It should work on win98 and XP. > Ok, the demo still got a few known syntax and sizecast bugs, > but 'blew up'? > :) All I remember is that it halted or hung right after startup... I figured it might've been due to various SE "enhancements" from Axcel216... ;) I'm using Win98SE full using a special non-destructive install over Win98FE upgrade using Win3.1 auth. plus Axcel216 stuff. :) I've also got my partitions reversed. I.e., Windows isn't on C, but D. Rod Pemberton
From: T.M. Sommers on 21 Jul 2007 07:10
Rod Pemberton wrote: > "santosh" <santosh.k83(a)gmail.com> wrote in message > news:f7ql53$c8c$1(a)aioe.org... > >>>case 0x05: {la[--j] = la[j] / la[j+1]; break;} >> >>case 0x05: >>{ >> la[j-1] = la[j] / la[j+1]; >> j--; >> break; >>} > > I.e., assuming LALR(1) and applying precedence, the above is: > case 0x05: > { > --j; > la[j] = la[j] / la[j+1]; > break; > } > > It appears that DJGPP (GCC) works properly while OpenWatcom v1.3 fails. I'd > guess that most K&R style compilers work. The origial line invokes undefined behavior; anything the compiler does is "proper". -- Thomas M. Sommers -- tms(a)nj.net -- AB2SB |