dave@kbvan.com (Dave Van Allen) (04/17/91)
My origional posting here stated that I was having trouble compiling elm on a Convergent Technologies MiniFrame. I did'nt really give very much detail. I'll be more specific this time. First, thanks to the following who emailed me with some initial advice: Syd Weinstein - syd@DSI.COM - (spent a hour on the phone, I thank him for his time. BTW, Syd I have Patch PL12 and that fixed the patch problem) Miles ONeal - meo@Dixie.com David Snyder - {SH}!trac.deci.com!das The specifics: CT Miniframe-1.5 meg memory-100meg HD ELM PL11 CTIX 3.2 SysV OS (suppose to be very vanilla) Has both cc and mcc (not Mcc) compiler -O works (I'm told that the compiler is limited to 8 character defines? variables?) History: I applied all patches, no problems, then ran Configure the only thing that looked out of the ordinary was some message about siginv() that said race conditions could occur? Everything else seemed OK, I took the defaults on anything I was not sure of; (most evrything was default) Then: I ran the make and here is the output of the MAKELOG: ==========================makelog output============================= cd src; /bin/make -b all /bin/chmod u+w ../hdrs/defs.h /bin/touch ../hdrs/defs.h /bin/chmod u+w ../hdrs/headers.h /bin/touch ../hdrs/headers.h mcc -O -I../hdrs -c addr_util.c ../hdrs/sysdefs.h: 46: DEFAULT_ redefined ../hdrs/sysdefs.h: 90: default_ redefined ../hdrs/sysdefs.h: 93: default_ redefined ../hdrs/sysdefs.h: 94: default_ redefined ../hdrs/sysdefs.h: 95: default_ redefined ../hdrs/sysdefs.h: 102: submitfl redefined ../hdrs/defs.h: 114: DONT_SAV redefined ../hdrs/defs.h: 120: default_ redefined ../hdrs/defs.h: 121: default_ redefined ../hdrs/defs.h: 122: default_ redefined ../hdrs/defs.h: 260: PMALLOC_ redefined "../hdrs/headers.h", line 53: syntax error "../hdrs/headers.h", line 156: redeclaration of mailfile "../hdrs/headers.h", line 163: syntax error "../hdrs/headers.h", line 167: redeclaration of alternat "../hdrs/headers.h", line 169: syntax error "addr_util.c", line 552: illegal function "addr_util.c", line 552: illegal function "addr_util.c", line 957: redeclaration of parse_ar *** Error code 1 Stop. *** Error code 1 Stop. ===========================end======================================= So, that's as far as I get. Geeez I hate xmail and _really_ want to run elm on this beast. BTW, I have two copies of the ELM source and have tryed both with the same results, that rules out a bad tar, I hope. Thanx in advance for any help Dave Van Allen +----------------------------------------------------------------------------+ | Lehigh Software Systems --- (Medical Industry Software and Hardware) | | Van Allen and Associates -- (anything for a buck enterprises) | | ---------------------------------------------------------------------------| | Dave Van Allen - dave@kbvan.com =internet ------------------------ | | uunet!kbvan!dave =uucp | This SPACE for SALE | | | CIS 73530,2463 =CompuServe ------------------------ | | "You'll never go forward unless you have something to look forward to" | | ~~~~~~ ~~~~~ ~~ ~~~~~~~ ~~~~~~ ~~~ ~~~~ ~~~~~~~~~ ~~ ~~~~ ~~~~~~~ ~~ | +----------------------------------------------------------------------------+
syd@DSI.COM (Syd Weinstein) (04/18/91)
dave@kbvan.com (Dave Van Allen) writes: >The specifics: CT Miniframe-1.5 meg memory-100meg HD > ELM PL11 > CTIX 3.2 SysV OS (suppose to be very vanilla) > Has both cc and mcc (not Mcc) compiler -O works >(I'm told that the compiler is limited to 8 character defines? variables?) Yup, from the messages: >../hdrs/sysdefs.h: 46: DEFAULT_ redefined >../hdrs/sysdefs.h: 90: default_ redefined Thats the problem. Its a killer for you. The older System III C compiles treated all names as only the first 8 were considered. Elm does not follow that rule and many names are the same for the first 8 and only differ there after. Long ago, Dave distributed a shortnames.h file that would get around this, but as system 3 got older and older, I have dropped that. Sorry, we only support compiles that have the modern length (min 32, normal 127 or 255) for variable names. Now, there are some programs in the c.s.unix or c.s.misc archives that will do a cross reference and translate the names via defines, but those depend on the pre-processor having the longer names and only the compiler the shorter ones. But as you can see from the messages yours has both being shorter. I vote for S. O. L. on this one. -- ===================================================================== Sydney S. Weinstein, CDP, CCP Elm Coordinator Datacomp Systems, Inc. Voice: (215) 947-9900 syd@DSI.COM or dsinc!syd FAX: (215) 938-0235
scs@lokkur.dexter.mi.us (Steve Simmons) (04/18/91)
dave@kbvan.com (Dave Van Allen) writes: >My origional posting here stated that I was having trouble compiling >elm on a Convergent Technologies MiniFrame . . . >The specifics: CT Miniframe-1.5 meg memory-100meg HD > ELM PL11 > CTIX 3.2 SysV OS (suppose to be very vanilla) ^^^^^^^^^^^^^^^^^ Your problem is the CTIX 3.2 C compiler. The preprocessor stage has an 8-character limit to macro names. Yes, that's insane but that's the way it is. Fortunately there is help. The MiniFrame is *extremely* similar to the UNIX-PC -- to the point that they could often interchange object files. Post a note to comp.sys.3b1, and ask if there's anyone who can uuencode the cpp from gcc and mail it to you. -- "FACT: less than 10% of the psychiatrists in the US are actually practicing cannibals." Rod Johnson
brad@bradley.bradley.edu (Bradley E. Smith) (04/18/91)
Onething I would try.....the CT-miniframe and UNIXPC are binary compatable.....how about trying to use gcc/gas with the CT-miniframe, or get someone to send you binaries (stay away from shared libraries) that were compiled on a UNIXPC. If you need binaries.....I don't have the current version compiled, but will do so. -- Bradley Smith Network & Technical Services @ Bradley University, Peoria, IL brad@bradley.edu --- 309-677-2337
wwm@pmsmam.uucp (Bill Meahan) (04/18/91)
Hmm... I've successfully compliled ELM PL11 on my AT&T 3B1/7300 which is really just about the same beast. Many others have as well. I Acutally had fewer problems getting it up on the 3b1 than on my HP-UX systems here at work. Are you SURE you've answered the questions in Configure properly? That ELM was unpacked with the right directory structure? THAT THE PROPER PERMISSIONS ARE SET ON ALL FILES AND DIRECTORIES (that's been a big killer for me) and that "you" own ALL the distribution files (not the system files, of course)???? If you can't get it going, send me mail at my home machine (wwm@wa8tzg.mi.org) and perhaps I can send you whatever files might be useful. Oh, and one thing more: if you own a MiniFrame (or MightyFrame or Motorola 6350 or AT&T Unix-PC/7300/3b1) you REALLY ought to subscribe to comp.sys.3b1 !! -- Bill Meahan |Product Design & Testing Section Production Test Engineer |Starter Motor Engineering wwm@pmsmam | +1 313 484 9320
das@trac2000.ueci.com (David Snyder) (04/19/91)
In article <1991Apr18.000403.17095@lokkur.dexter.mi.us>, scs@lokkur.dexter.mi.us (Steve Simmons) writes: > dave@kbvan.com (Dave Van Allen) writes: > > Fortunately there is help. The MiniFrame is *extremely* similar to the > UNIX-PC -- to the point that they could often interchange object files. > Post a note to comp.sys.3b1, and ask if there's anyone who can uuencode > the cpp from gcc and mail it to you. Better hope that the cpp from gcc hasn't been compiled with shared libraries. If it has been, you're (as Syd said) "SOL"! DAS -- David A. Snyder @ UE&C - Catalytic in Philadelphia, PA UUCP: ..!uunet!trac2000!das INTERNET: das@ueci.com
dick@cca.ucsf.edu (Dick Karpinski) (04/19/91)
I think the free gcc compiler is available for the 68010 and would be willing to handle the longer names of ELM. Dick
wieland@ea.ecn.purdue.edu (Jeffrey J Wieland) (04/22/91)
In article <1991Apr17.195859.26819@DSI.COM> syd@DSI.COM writes: >dave@kbvan.com (Dave Van Allen) writes: > >Thats the problem. Its a killer for you. The older System III C compiles >treated all names as only the first 8 were considered. Elm does >not follow that rule and many names are the same for the first 8 and >only differ there after. Long ago, Dave distributed a shortnames.h > >Now, there are some programs in the c.s.unix or c.s.misc archives >that will do a cross reference and translate the names via defines, >but those depend on the pre-processor having the longer names and only >the compiler the shorter ones. But as you can see from the messages >yours has both being shorter. > >I vote for S. O. L. on this one. The Miniframe is binary compatible with the AT&T 3B1 (7300, UNIX-PC, etc.), except that the Miniframe does not have shared libraries. If you can find someone who is willing to compile Elm for you on their 3B1 and ship you the binaries, you should be OK. You would have to work out how to handle running the Configure script. One note: you should replace the '-lcurses' with '-ltermlib' before compilation on the 3B1. I don't know if the Miniframe has the terminfo stuff -- I would suspect not. The termcap routines should work fine, since Elm doesn't really use curses-specific stuff anyway. One person suggested getting cpp from a 3B1 -- this won't work because cpp has the shared libraries linked. An alternate cpp might be the one from GNU-CC. It could also be compiled on a 3B1, and as long as the shared libraries aren't linked, it should work. -- Jeff Wieland wieland@ecn.purdue.edu