|
|
sybperl-l Archive
Up Prev Next
From: Jamie Belanger <jamie_belanger at globalcrossing dot com>
Subject: compiler/make issues (fancy that)
Date: Oct 12 2000 8:36PM
Well, I'm having a devil of a time getting sybperl to compile.
# /link/make_temp/sybperl/sybperl-2.12> make
cc -c -I/link/sybase/include -Ae -D_LARGEFILE_SOURCE
-D_FILE_OFFSET_BITS
=64 -O +Onolimit -DVERSION=\"2.12\" -DXS_VERSION=\"2.12\" +z
-I/opt/perl5/
lib/5.6.0/PA-RISC1.1/CORE -DCTLIBVS=1000 -DSYBPLVER='"2.12"' -DDO_TIE
CTlib.c
cc: +Onolimit: No such file or directory
cc: +z: No such file or directory
*Initialization*:1: missing token-sequence in `#assert'
*** Error exit code 1
Stop.
*** Error exit code 1
Stop.
My problem is the C compiler. Initially, I only had the bundled C compiler
that comes standard with all HP-UX systems (this particular server is
running HP-UX 10.20). I downloaded a pre-compiled binary of the gcc compiler
and installed it on the system.
As far as I know, the gcc compiler (gcc version 2.95.2 19991024 (release))
should work to compile sybperl. I even created a symbolic link from cc to
the actual location of gcc. However, it still seems to be fumbling up. The
default location of the bundled cc compiler is /usr/ccs/bin/cc, and there is
a symbolic link from /usr/bin/cc to this location. The gcc compiler is in
/usr/gcc/bin, and it's also listed first in my PATH variable. I also created
a symbolic link from /usr/gcc/bin/cc to /usr/gcc/bin/gcc in an effort to
fake the thing out.
If I perform a "which cc" it comes up with the /usr/gcc/bin/cc, which should
be resolving to the gcc compiler. This may be more of a unix question than a
sybperl question, but any advice would be most appreciated. Other than
fiddling in my unix environment, is there any other way to change the
compiler that the make is looking at?
Jamie Belanger
Unix Administrator
Global Crossing
(319) 298-1489 or ext. 489
-----Original Message-----
From: Thomas Santamaria [mailto:Thomas.Santamaria@msdw.com]
Sent: Wednesday, October 11, 2000 2:19 PM
To: SybPerl Discussion List
Subject: "infected with 11"
When running a stored proc from a Perl script, we received the following
error;
Sybase error: 20047
Sybase error: DBPROCESS is dead or not enabled.
The Sybase errorlog reported the situation as;
00:00000:00018:2000/10/10 16:44:46.22 kernel current process (0x18cf0084)
infected with 11
00:00000:00018:2000/10/10 16:44:46.23 kernel Address 0x46d104
(_sparclock+0x4), siginfo (code, address) = (1, 0x0)
00:00000:00018:2000/10/10 16:44:46.23 kernel
************************************
00:00000:00018:2000/10/10 16:44:46.23 kernel SQL causing error :
some_stored_proc
00:00000:00018:2000/10/10 16:44:46.23 kernel
************************************
00:00000:00018:2000/10/10 16:44:46.23 server SQL Text: some_stored_proc
00:00000:00018:2000/10/10 16:44:46.23 kernel curdb = 9 pstat = 0x10000
lasterror = 0
00:00000:00018:2000/10/10 16:44:46.23 kernel preverror = 0 transtate = 0
00:00000:00018:2000/10/10 16:44:46.23 kernel curcmd = 195 program =
some_program
This is followed by a stack trace. We are inclined to believe that it may
have something to do with object id's being out of sink. The stored proc
referenced a table in another database that had earlier been restored from a
dump taken on a different server. I might also mention that the server
running the stored proc was running 11.9.2.2 whereas the dump was created on
a server running 11.0.3.3.
I know this sounds a bit odd, but I'm hoping if someone else has experienced
this that perhaps they can provide some insight?
|
