palowoda@megatest.UUCP (Bob Palowoda) (08/31/89)
Well I got some answers to some of my Xenix questions. Sort of. ------------------------------------------------------------------- From: purdue!haven.umd.edu!ames!rogerk@decwrl.UUCP (Roger Knopf 5502) Organization: The Santa Cruz Operation, Inc. [some stuff deleted] [a quote from Mr, Knopf] Not all the utilities were recompiled as 386 because most were small model and would actually run faster as a small model 8086 binary with 16 bit ints and pointers. Any utility which would benefit from being compiled 386 has been. Roger Knopf The Santa Cruz Operation, Inc. ------------------------------------------------------------------ | | V ------------------------------------------------------------------- From: ucscc.UCSC.EDU!seanf@sun.UUCP (Sean Fagan) Organization: The Santa Cruz Operation, Inc. Status: OR In article <7396@megatest.UUCP> you write: > I would be interested to know what 386 Utilities on Xenix are > still compilied with the 286 compiler. Can someone do a > file * on the /bin and /usr/bin directories and pipe me the > output. Well, if you have a Xenix 2.3 system, all of the binaries on the N floppies (I think) are '386. The others are '286. -------------------------------------------------------------------- | V -------------------------------------------------------------------- From: jim@bahamut.fsc.com (James O'Connor) Subject: Re: SCO MicroSoft C Compiler comments [Thank you Mr. O'Connor for the list] Too damn many!!! I know it would not take much for SCO to re-compile everything, but they seem reluctant to. They do say, however, that SCO Unix is completely compiled in 386 mode. Here's the list from this machine: --------------------------------------------------------------------- I won't publish the list but to some it up there where: 59 --- 80386 Executables (Some of these where not even from SCO) 5 --- 80286 Executables 179 -- 8086 Executables I agree with Mr Knopf to a point. I'm still interested why these following programs are 286 (or 8086)executables and *NOT* 386. THIS IS ON A 2.3 OS. SCO? /bin/as: 80286 separate executable, Middle model /bin/asm: 80286 separate executable, Middle model /bin/asx: 8086 separate executable /bin/cc: 8086 separate executable /bin/cc286: 8086 separate executable /bin/nm: 8086 separate executable /usr/bin/lex: 8086 separate executable And about the lib's ---------------------------------------------------------------------------- From: ucscc.UCSC.EDU!seanf@sun.UUCP (Sean Fagan) Organization: The Santa Cruz Operation, Inc. Status: OR > Also is the lib files for Xenix386 created with the optimizer > turned on? In SCO UNIX? Yes. All of the libraries have at least some optimization turned on, but not all of them have full optimization. This is true in both SCO Xenix and SCO Unix. Sean. ----------------------------------------------------------------------------- | | V ----------------------------------------------------------------------------- From: Dr. T. Andrews <uunet!ki4pv!tanner@moldev.UUCP> [some stuff deleted] [a quote from Mr Andrews] The libraries are NOT built with ANY opptimization, which shows both little regard for the final program and for their optimizer. ----------------------------------------------------------------------------- Ok so what give's are the lib's built with optimization or what. How can you tell anyways? One thing I do know is Dr. Andrews indicated that when he turned on loop optimaztion programs would crash. I ran into this too. ---Bob -- Bob Palowoda *Home of Fiver BBS* login: bbs Work: {sun,decwrl,pyramid}!megatest!palowoda Home: {sun}ys2!fiver!palowoda (A XBBS System) 2-lines BBS: (415)623-8809 2400/1200 (415)623-8806 1200/2400/9600/19200
fr@icdi10.UUCP (Fred Rump from home) (09/03/89)
In article <7477@megatest.UUCP> palowoda@megatest.UUCP (Bob Palowoda) writes: > > > Well I got some answers to some of my Xenix questions. Sort of. [binary compile stuff deleted] > And about the lib's > >---------------------------------------------------------------------------- >From: ucscc.UCSC.EDU!seanf@sun.UUCP (Sean Fagan) >Organization: The Santa Cruz Operation, Inc. >Status: OR > > >> Also is the lib files for Xenix386 created with the optimizer >> turned on? In SCO UNIX? > >Yes. All of the libraries have at least some optimization turned on, but >not all of them have full optimization. This is true in both SCO Xenix and >SCO Unix. >Sean. >---------------------------------------------------------------------------- >From: Dr. T. Andrews <uunet!ki4pv!tanner@moldev.UUCP> >The libraries are NOT built with ANY opptimization, which shows both >little regard for the final program and for their optimizer. >----------------------------------------------------------------------------- > > Ok so what give's are the lib's built with optimization or what. > How can you tell anyways? > > One thing I do know is Dr. Andrews indicated that when he turned > on loop optimization programs would crash. I ran into this too. Programs with the loop optimizer on definitely do crash or behave erratically. Dr T has been known to go into mental loops trying to convince SCO of the cost of these problems on productivity and customer satisfaction. The latest set of "proofs" accompanied by the usual set of demonstration programs has been acknowledged (finally) by SCO as being beyond their control as MS is the real culprit behind the compiler scenes. Being that the problem can easily be demonstrated one wonders how Mr Sagan (SCO) managed to optimize 'any' library routines per his statement. Perhaps if we were let in on the secret we could use the same trick. There are other internal problems with the MS C compiler. Prototypes and other ANSI or even K&R standards seem to be in a state of confusion. One would think that mega-bucks Microsoft would have the resources and personnel to produce a compiler that is worthy of their size and power in the C development community. But perhaps the rumors are correct: that everyone is working on OS/2 and PM. Come hell or high water they're going to have those babies work on a 386 in native mode yet. So who cares about plain old C or UNIX? Why should they get so far ahead of basic bread and butter DOS/OS/2/3. Fred Rump -- This is my house. My castle will get started right after I finish with news. 26 Warren St. uucp: ...{bpa dsinc uunet}!cdin-1!icdi10!fr Beverly, NJ 08010 domain: fred@cdin-1.uu.net or icdi10!fr@cdin-1.uu.net 609-386-6846 "Freude... Alle Menschen werden Brueder..." - Schiller
allbery@NCoast.ORG (Brandon S. Allbery) (09/04/89)
It doesn't hurt for /bin/cc to be compiled for the 8086; it's nothing more than a dispatcher for the actual compiler (/lib/p[123] and possibly /lib/cpp), /bin/as, and /bin/ld. However, all of *those* should be compiled for the 80386 -- and *please*, compile cpp, lex, and yacc with larger table sizes! ++Brandon -- Brandon S. Allbery, moderator of comp.sources.misc allbery@NCoast.ORG uunet!hal.cwru.edu!ncoast!allbery ncoast!allbery@hal.cwru.edu "Why do trans-atlantic transfers take so long?" "Electrons don't swim very fast." -john@minster.york.ac.uk and whh@PacBell.COM
palowoda@megatest.UUCP (Bob Palowoda) (09/06/89)
From article <411@icdi10.UUCP>, by fr@icdi10.UUCP (Fred Rump from home): > In article <7477@megatest.UUCP> palowoda@megatest.UUCP (Bob Palowoda) writes: > > But perhaps the rumors are correct: that everyone is working on OS/2 and PM. > Come hell or high water they're going to have those babies work on a 386 in > native mode yet. So who cares about plain old C or UNIX? Why should they get > so far ahead of basic bread and butter DOS/OS/2/3. > Fred Rump What do you think thier compiling OS2 in? Why they don't have a 32bit version out? ---Bob > -- Bob Palowoda *Home of Fiver BBS* login: bbs Work: {sun,decwrl,pyramid}!megatest!palowoda Home: {sun}ys2!fiver!palowoda (A XBBS System) 2-lines BBS: (415)623-8809 2400/1200 (415)623-8806 1200/2400/9600/19200
davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr) (09/06/89)
I have to agree that thyere are problems in Xenix C. I had a program which gave me the noxious "infinite spill". I called SCO and they explained that I had to simplify the line which caused the problem (as if that was a secret fix). I asked for a hint on simplifying "return 1" but not no useful info. I did, at their request, send a copy of the program about three months ago, but I never got a reply. Due to another SCO bug (fsck) I have lost the problem number on this, so they can disavow all knowlege when I call. They did send me a new fsck copy, although the person on the line didn't know about the bbs so I got it surface mail. I am looking forward to updating to SCO UNIX, just to get the pcc compiler. There are some programs which just won't go thru the standard compiler due to using bugs in pcc as features. -- bill davidsen (davidsen@crdos1.crd.GE.COM -or- uunet!crdgw1!crdos1!davidsen) "The world is filled with fools. They blindly follow their so-called 'reason' in the face of the church and common sense. Any fool can see that the world is flat!" - anon