I am sorry if this isn't the right newsgroup for this, but I'm not sure
where else to turn at this point. Someone went to great effort to help
me obtain the source for AKCL (the base dist and the 1530 release).
Now I'm trying to get it built to use for a class I'm taking (at the
rate I'm going the class will be over before I get this built :-)
Anyway, I'm using cc, not gcc -- and this is the tail end of a
"make -f Smakefile":
(cd cmpnew; make all)
cc -DVOL= -I. -I/usr2/bill/lisp/akcl/o -O -c -I../h cmptop.c
"./cmptop.h", line 27: illegal character: 043 (octal)
"./cmptop.h", line 27: cannot recover from earlier errors: goodbye!
*** Error code 1
Stop.
*** Error code 1
Stop.
*** Error code 1
Stop.
%
-------
akcl/cmpnew/cmptop.h looks like this:
(when reading it in to vi, vi complains of an "incomplete last line."
Looking at it with :set list in vi doesn't show anything funny.
static L1();
static L4();
static L21();
#define VC1 object V16 ,V15 ,V14 ,V13 ,V11 ,V7 ,V5;
static object LI2();
#define VMB2 register object *base=vs_top; object V42 ,V40;
#define VMS2 register object *sup=vs_top+10;vs_top=sup;
#define VMV2 vs_check;
#define VMR2(VMT2) {object CMPret2=VMT2;vs_top=base;return(CMPret2);}
static object LI3();
#define VMB3 register object *base=vs_top;
#define VMS3 register object *sup=vs_top+1;vs_top=sup;
#define VMV3 vs_check;
#define VMR3(VMT3) {object CMPret3=VMT3;vs_top=base;return(CMPret3);}
#define VC4
static object LI5();
#define VMB5 register object *base=vs_top;
#define VMS5 register object *sup=vs_top+1;vs_top=sup;
#define VMV5 vs_check;
#define VMR5(VMT5) {object CMPret5=VMT5;vs_top=base;return(CMPret5);}
static object LI6();
#define VMB6 register object *base=vs_top; object V77;
#define VMS6 register object *sup=vs_top+1;vs_top=sup;
#define VMV6 vs_check;
#define VMR6(VMT6) {object CMPret6=VMT6;vs_top=base;return(CMPret6);}
static object LI7();
------
Has anyone else out there run across anything like this, and what
should I do to correct it? I think it is unlikely that I got a
corrupt distribution, because it was a tarfile that unpacked correctly.
The person who sent it compiled it on Esix and did not run into this
problem.
Any suggestions on this would be most appreciated (I really hate the
thought of using the VMS machine at school ...)
Thanks in advance!
Bill
--
home: ...!{uunet,bloom-beacon,esegue}!world!unixland!bill
····@unixland.uucp Public Access Unix - Esix SYSVR3
508-655-3848(12/24) 508-651-8723(12/24/96-HST) 508-651-8733(12/24/96-PEP-V32)
other: ······@world.std.com
[problems installing AKCL under SVR3; compilation aborts with
some bizarre error message]
From memory (I installed AKCL a couple of years ago): these kinds of
strange errors can happen when the relative dates on the distribution
are wrong. If you unpacked AKCL before unpacking the KCL distribution,
and you did not allow/instruct tar (depending on your version of tar)
to set the file dates correctly, most change files will be older than
their corresponding source files and will not get applied.
In article <················@volterra.ai.mit.edu> ···@ai.mit.edu (Thomas M. Breuel) writes:
>
> [problems installing AKCL under SVR3; compilation aborts with
> some bizarre error message]
>
>From memory (I installed AKCL a couple of years ago): these kinds of
>strange errors can happen when the relative dates on the distribution
>are wrong. If you unpacked AKCL before unpacking the KCL distribution,
>and you did not allow/instruct tar (depending on your version of tar)
>to set the file dates correctly, most change files will be older than
>their corresponding source files and will not get applied.
I just went thru and 'touch'ed all of the files in the akcl directory,
so they would definitely be newer than the files in the kcl direcotry,
and the compile aborted the same way. Strange, eh!
--
home: ...!{uunet,bloom-beacon,esegue}!world!unixland!bill
····@unixland.uucp Public Access Unix - Esix SYSVR3
508-655-3848(12/24) 508-651-8723(12/24/96-HST) 508-651-8733(12/24/96-PEP-V32)
other: ······@world.std.com
From: Darryl Okahata
Subject: Re: Compiling AKCL on SYSVR3 Unix (Esix) System (LONG)
Date:
Message-ID: <7490001@hpnmdla.HP.COM>
In comp.lang.lisp, ····@unixland.uucp (Bill Heiser) writes:
> I am sorry if this isn't the right newsgroup for this, but I'm not sure
> where else to turn at this point. Someone went to great effort to help
> me obtain the source for AKCL (the base dist and the 1530 release).
> Now I'm trying to get it built to use for a class I'm taking (at the
> rate I'm going the class will be over before I get this built :-)
>
> Anyway, I'm using cc, not gcc -- and this is the tail end of a
> "make -f Smakefile":
>
> (cd cmpnew; make all)
> cc -DVOL= -I. -I/usr2/bill/lisp/akcl/o -O -c -I../h cmptop.c
> "./cmptop.h", line 27: illegal character: 043 (octal)
> "./cmptop.h", line 27: cannot recover from earlier errors: goodbye!
> *** Error code 1
This looks like a compiler (C-preprocessor?) bug that I ran into
(at least, I think it was on a 386 Unix). I'm not sure what you can do.
Under Esix rev D, I compiled akcl 1-505 using gcc, and I had no
problems (I just followed the directions), except for one where the the
"ld" command line exceeded the kernel max line length limitation. As I
was compiling from within GNU Emacs, I just cut out the linker line into
a file, edited out unneeded stuff like C preprocessor definitions
(making the line fall just below the max line length), and did a "sh file".
After restarting the make, the akcl build finished without any
problems.
However, the build did take a few hours.
-- Darryl Okahata
UUCP: {hplabs!, hpcea!, hpfcla!} hpnmd!darrylo
Internet: ·············@hp-sde.sde.hp.com