arla on NeXTSTEP3.3/Intel

[Helmut Heller]

[ this was bounced by majordomo as it was rather large.  I have cut out the large attachments.
Helmut: I think people would appreciate it if you made them available by ftp/http and just included pointers to them in your mail.  I don't think all the readers of the mailinglist are that interested in these long files.   Thanks.  /assar ]
You wrote:
> > Well, as a matter of fact, I *do* have a working X11 (this is
> > available as shareware), but my X11 does not have
> > IceConnectionNumber() and other symbols, so the linking fails. So this
> > is a more local thing, nothing to include in a general config file.
> 
> But we never call IceConnectionNumber so I don't see why it should
> matter.  So, does configure thinks you have X11, right?  What does
> fail then when building amon?

Ah, I *did* find the sources (arla 0.21), so I can play with it.

Well, IceConnectionNumber is checked for in configure:

next 49 cpu> find . -exec grep IceConnectionNumber  ; -print
echo $ac_n "checking for IceConnectionNumber in -lICE""... $ac_c" 1>&6
echo "configure:2201: checking for IceConnectionNumber in -lICE" >&5
ac_lib_var=`echo ICE'_'IceConnectionNumber | sed 'y%./+-%__p_%'`
char IceConnectionNumber();
IceConnectionNumber()
./configure


I will try a 
make clean 
and then a
rm config.cache
and 
./configure
without any arguments:

[ ... config.log cut out ... ]
[ ... config.cache cut out ... ]
[ ... config.status cut out ... ]
[ ... config.stdout cut out ... ]
[ ... several make logs cut out ... ] 
now we have two choices: either use cc -bsd here, or change config.h to 
/* #undefine HAVE_SYS_UTSNAME_H  */
which I do. Perhaps the check in configure for HAVE_SYS_UTSNAME_H should use the -posix flag or make otherwise sure that only POSIX header files are considered. Or it should have hard wired that NeXTSTEP does  not  have HAVE_SYS_UTSNAME_H in POSIX mode.

[ ... and more make logs ... ] 
mode_t is only available in bsd mode:

next 80 cpu> cd /usr/include
next 81 cpu> find . -exec grep mode_t  ; -print
typedef unsigned short  mode_t;
./bsd/sys/types.h
mode_t  st_mode;        /* File type and access */
extern int chmod(const char *path, mode_t mode);
extern int mkdir(const char *path, mode_t mode);
extern int mkfifo(const char *path, mode_t mode);
extern mode_t umask(mode_t cmask);
extern mode_t umask();
./bsd/sys/stat.h
extern creat(const char *path, mode_t mode);
./bsd/sys/fcntl.h
typedef int /* default */ _G_mode_t;
./g++/_G_config.h



same for nlink_t:

next 82 cpu> find . -exec grep nlink_t  ; -print
typedef short   nlink_t;
./bsd/sys/types.h
nlink_t st_nlink;       /* File link count */
./bsd/sys/stat.h
typedef int /* default */ _G_nlink_t;
./g++/_G_config.h

 Probably here the real porting effort starts. Well, well....

I never did any kernel stuff before...if you want to, I can give you a guest account on my machine :-) And there are other kernel modules, e.g., the PPP stuff at 
http://www.thoughtport.com:8080/PPP/Latest_version.html
ftp://ftp.peak.org/pub/next/apps/internet/ppp/dev/




---
Servus, Helmut  (DH0MAD)     ______________NeXT-mail welcome_________________
FAX: +49-89-280-9460        " Knowledge must be gathered and cannot be given "
heller at lrz.de                                        ZEN, one of BLAKES7
Phone: +49-89-289-28823      ------------------------------------------------
Dr. Helmut Heller                                    
Leibniz-Rechenzentrum (LRZ)                         
High Performance Computing Group
Barer Str. 21, Zi S2515, 80333 Munich, GERMANY