On Fri, 10 Jan 2003 17:22:00 +0900, "Stephen J. Turnbull" said:
Valdis, Mike, I (perhaps incorrectly) associate you guys with some
AIX
knowledge. Comments?
I have an AIX 5.1 box here under my desk. Unfortunately, it's out to lunch
at the moment - Something Evil happened to 'bosboot'. Filesystems are there,
but installing the bootstrap mangles the LVM header on hd6 so it's not
recognized as a boot LVM.
I'll look into it as soon as I resurrect the box - if it looks like it will
be more than 2-3 days I'll do it on another test box I have access to (but
that one is *ancient* and builds will be painful).
Paul> #define config_opsysfile "s/aix4-2.h"
Paul> There is no aix5-1.h file in src/s.
There probably needs to be one. Unfortunately, somebody with access
to AIX 5 docs and a test system will have to write it.
There's not that much diff between AIX5.1 and 4.3.3 - in fact, an lsof
compiled for 4.3.3 worked just fine.
% cat > s/aix5.h
#define AIX5
#include <aix4-2.h>
^D
% cat > s/aix5-1.h
#define AIX5_1
#include <aix5.h>
^D
I *THINK* that should do it, but I'll have to look into the part about
'ALL_SOURCE'. This may be compiler rev dependent, and whether you use
the 'cc' or 'xlc' commands, and a few other things....
Paul> root@feynman /usr/local/xemacs-21.4.11$ Fatal error:
Paul> assertion failed, file specifier.c, line 2583, abort()
This is bad. That code hasn't changed in years (only comments
have
I'm getting a strange spider-sense tingling that this is an old IBM
compiler bug. Does it happen if you recompile without -O?
I'll check into it more as soon as I beat bosboot into submission (which
may involve sneakernetting in patches on floppy, blech).
/Valdis