Bugs item #2971516, was opened at 2010-03-16 16:22
Message generated for change (Comment added) made by mbucc
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2971516&group_id=56967
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Core
Group: MonetDB5 "stable"
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Mark Bucciarelli (mbucc)
Assigned to: Martin Kersten (mlkersten)
Summary: pci->fcn = 0 in malAtomProperty: segfault
Initial Comment:
Feb2010 CVS.
OpenBSD 4.6
gcc 3.3.5 (propolice)
Trying to start database, the mserver5 dumps core.
At line 92 of mal_atom.c, code tries to dereference null pointer:
case 'n':
if( idcmp("null",name)==0 && pci->argc==1){
-----> ptr atmnull = ((ptr (*)(void)) pci->fcn) ();
----------------------------------------------------------------------
>Comment By: Mark Bucciarelli (mbucc)
Date: 2010-04-15 14:37
Message:
short answer: yes this is stilll open; I'm not
sure if it needs a test.
The library loading code first calls a locate file
routine that uses open() with a specific
filename, which is more restrictive than a
dlopen() with the same file name.
the patch depends on if you consider this
an os-specific case or not. (There is already
#ifdef AIX logic in this code path, so it looks
like something similar has come up before.)
My current thought is to delete the open() code
path, and just use dlopen(). I can't think of any
value the open() call adds. I plan to build a patch
for this.
----------------------------------------------------------------------
Comment By: Sjoerd Mullender (sjoerd)
Date: 2010-04-15 10:20
Message:
Is this still open?
Do we need a test?
----------------------------------------------------------------------
Comment By: Martin Kersten (mlkersten)
Date: 2010-03-16 19:57
Message:
Indeed, the server should come to a graceful stop. More defensive code is
added.
----------------------------------------------------------------------
Comment By: Mark Bucciarelli (mbucc)
Date: 2010-03-16 18:44
Message:
Looks like a loader path issue.
But shouldn't the server fail earlier and harder?
Can the server ever function normally if the mdb (or blob, or calc, or
...)
library is not found?
2010-03-16 17:45:29 MSG test[3349]: # MonetDB server v5.18.2, based on
kernel v1.36.2
2010-03-16 17:45:29 MSG test[3349]: # Serving database 'test', using 1
thread
2010-03-16 17:45:29 MSG test[3349]: # Compiled for
i386-unknown-openbsd4.6/32bit with 32bit OIDs dynamically linke
d
2010-03-16 17:45:29 MSG test[3349]: # Found 191.590 MiB available
main-memory.
2010-03-16 17:45:29 MSG test[3349]: # Copyright (c) 1993-July 2008 CWI.
2010-03-16 17:45:29 MSG test[3349]: # Copyright (c) August 2008-2010
MonetDB B.V., all rights reserved
2010-03-16 17:45:29 MSG test[3349]: # Visit http://monetdb.cwi.nl/ for
further information
2010-03-16 17:45:29 MSG test[3349]: #WARNING:
LoaderException:loadLibrary:File not found:mdb
2010-03-16 17:45:29 MSG test[3349]: #WARNING:
LoaderException:loadLibrary:File not found:unix
etc.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2971516&group_id=56967
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
