ray@asci.UUCP (Ray Tripamer) (03/05/87)
After reading the problems with zapped directories in Altos Xenix (Yes, it happened to us, too), I was wondering if anyone has notices that the C compiler for Xenix 3.2f doesn't work all the time. What happens is that the compiler will make it to one of the passes (i've seen it happen on /lib/p2 and /lib/p3), when it seems to get stuck in an infinite loop. It has happened more than once, and it doesn't matter which memory model I use. It prevented me from compiling news, ctex, and rogue. I have not been able to track down the cause. Any help would be appreciated. -- Ray Tripamer E-MAIL: ray@jimi.uucp (until we get into mod.maps)
ries@trwrb.UUCP (03/06/87)
In article <121@asci.UUCP> ray@asci.UUCP (Ray Tripamer) writes: [...] > if anyone has notices that the C compiler for Xenix 3.2f > doesn't work all the time. [...] > when it seems to get stuck in an infinite loop. It has > happened more than once, and it doesn't matter which > memory model I use. [...] > have not been able to track down the cause. Any help > would be appreciated. >Ray Tripamer BTW: You didn't mention what hardware version of Altos you are running (ie, 8086, 80286, 68020, etc.). FYI: I have no solution [don't buy an Altos 8-(], but I have only seen this happen once on my Altos 586 running XENIX 3.0b when trying to compile ICON. I don't remember the module off hand, but the compiler just goes off into never- never land. Which is surprising, since the Altos compiler usually gags on what would be considered *normal* C syntax (for BSD 4.2 or ATT), or runs out of space (either compiler or preprocessor). Marc Ries (trwrb!ries)
jlt@lll-crg.UUCP (03/07/87)
In article <121@asci.UUCP> ray@asci.UUCP (Ray Tripamer) writes: > > I was wondering > if anyone has notices that the C compiler for Xenix 3.2f > doesn't work all the time. > > What happens is that the compiler will make it to one of > the passes (i've seen it happen on /lib/p2 and /lib/p3), > when it seems to get stuck in an infinite loop. It has > happened more than once, and it doesn't matter which > memory model I use. > > It prevented me from compiling news, ctex, and rogue. I > have not been able to track down the cause. Any help > would be appreciated. >-- >Ray Tripamer >E-MAIL: ray@jimi.uucp (until we get into mod.maps) I'm running 3.3a Xenix, and have noticed the same problem. My compiler always seems to get stuck in /lib/p1. The only fix I've found is to to simplify the file in question. E.G. in rogue, I broke apart save.c into save.c and restore.c. Then it compiled, and even runs! I had a different problem when compiling MicroEMACS 3.8b. In a couple of files I got an error message something like "XXX.c (678) Compiler internal error: infinite spill in grammar.c" (the name of the file and the line number have been changed to protect the guilty). The solution was to simplify the statement at line number 677. This only arose when compiling with -Ml. John L. Thomas jlt@crg or ...seismo!lll-lcc!iei-ca!jlt
brian@asci.UUCP (brian) (03/10/87)
Summary: Expires: Sender: Followup-To: Distribution: Keywords: In article <1646@trwrb.UUCP> ries@trwrb.UUCP (Marc Ries) writes: >In article <121@asci.UUCP> ray@asci.UUCP (Ray Tripamer) writes: >[...] >> if anyone has notices that the C compiler for Xenix 3.2f >> doesn't work all the time. >[...] >> when it seems to get stuck in an infinite loop. It has >> happened more than once, and it doesn't matter which >> memory model I use. >[...] >> have not been able to track down the cause. Any help >> would be appreciated. >>Ray Tripamer > > BTW: You didn't mention what hardware version of Altos you are > running (ie, 8086, 80286, 68020, etc.). > > FYI: I have no solution [don't buy an Altos 8-(], but I have > only seen this happen once on my Altos 586 running XENIX > 3.0b when trying to compile ICON. They Altos CPUs we work on are 80186, and 80286 based. On the 186, the code compiles fine under 3.1 Development system. The 3.2F is Development and runtime on 886 and 686. It is 3.2F development on 2086 with 3.3 Runtime. Altos knows about these and other problems and has recently signed off version 3.4 runtime for 1086 and higher. 3.3 runtime for 686 and 886. 3.4 development for all 286 based units. 3.3 development does not correct problems mentioned. Contact your Dealer or Distributor for availablity and shipping dates. Altos is continuing to support but appears to have phased out all production on 8086 and 80186 based products. Watch for news on their 386 based box. Should be dynamite. I know some Altos gurus should be reading this and would appreciate more direct input for from said people on these and other problems, and any corrections to the above Brian Douglass Applied Systems Consultants, Inc. (ASCI)