Bugs item #2017862, was opened at 2008-07-14 12:41 Message generated for change (Comment added) made by jflokstra You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2017862&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: PF/runtime Group: Pathfinder CVS Head Status: Open Resolution: None Priority: 5 Private: No Submitted By: Ying Zhang (yingying) Assigned to: Jan Flokstra (jflokstra) Summary: PF: shredding empty file causes MXQ to crash Initial Comment: Shred an empty file using pf:add-doc() in mclient crashes Mserver: In terminal 1: $ Mserver --dbinit="module(pathfinder);" # MonetDB Server v4.24.1 # based on GDK v1.24.1 # Copyright (c) 1993-2008, CWI. All rights reserved. # Compiled for x86_64-unknown-linux-gnu/64bit with 64bit OIDs; dynamically linked. # Visit http://monetdb.cwi.nl/ for further information. # PF/Tijah module v0.5.0 loaded. http://dbappl.cs.utwente.nl/pftijah # MonetDB/XQuery module v0.24.1 loaded (default back-end is 'milprint_summer') # XRPC administrative console at http://127.0.0.1:50001/admin MonetDB>Segmentation fault In terminal 2: $ mclient -lx xquery>pf:add-doc("/ufs/zhang/tmp/empty.xml", "empty.xml") more><> MAPI = [EMAIL PROTECTED]:50000 ACTION= read_line QUERY = pf:add-doc("/ufs/zhang/tmp/empty.xml", "empty.xml") ERROR = Connection terminated The crash happens in shredder.mx Empty file is not a valid XML document, however, I think it should not cause Mserver to segmentation fault. Shredding the same doc using MIL shred_doc() results (correctly) in an error: MonetDB>shred_doc("/ufs/zhang/tmp/empty.xml", "empty.xml"); !ERROR: Document is empty !ERROR: Start tag expected, '<' not found !ERROR: [shred_url]: 1 times inserted nil due to errors at tuples [EMAIL PROTECTED] !ERROR: [shred_url]: first error was: !ERROR: emit_tuple: node.level(18831648) >= XML_DEPTH_MAX(128) !ERROR: CMDshred_url: operation failed. Kind regards, Jennie ---------------------------------------------------------------------- >Comment By: Jan Flokstra (jflokstra) Date: 2008-10-14 17:12 Message: I found another problem here. In case of an empty document libxml2 calls end_document() without having called start_document(). I now detect this and bail out a.s.a.p when this happens. This could solve the problem but there may be more. I checked it in in the HEAD. ---------------------------------------------------------------------- Comment By: Sjoerd Mullender (sjoerd) Date: 2008-10-14 17:04 Message: Could this be a problem of non-initialized memory when the memory is used for the second time, i.e. it is used in the first run, freed at the end of the query, and the next query allocates the same memory again. Until today, allocated memory was not initialized, but it was assumed that it was. I have seen crashes because of this. Today I checked in a change to mem.c to initialize all memory allocated by PFalloc and PFrealloc (only the extension, obviously) to zero. If this is indeed the cause, the bug should be fixed as of today. Somebody needs to check that. ---------------------------------------------------------------------- Comment By: Jan Flokstra (jflokstra) Date: 2008-10-14 16:37 Message: I had a look at this a couple of weeks ago and it did not crash with me on my SuSe10.3 system. I now have the newest HEAD and it still does not crash. I think I will just have to guess what goes wrong and anticipate on this. I will try to solve it this week (in the HEAD) ---------------------------------------------------------------------- Comment By: Stefan Manegold (stmane) Date: 2008-10-14 16:27 Message: Jan, could you please have a look at this one? It seems to work fine in the stable release, but fails in the development trunk ... Thanks! Stefan ---------------------------------------------------------------------- Comment By: Lefteris Sidirourgos (lsidir) Date: 2008-08-12 16:07 Message: Logged In: YES user_id=1856546 Originator: NO Did some looking around for this bug. Although the original bug report was for milprint_summer, the same behavior is noticed with the algebra (since is a runtime error anyway). The stack trace from gdb looks like: 0x00002aaab9bb22f5 in emit_tuple (shredCtx=0x18a1508, pre=15842497851538791387, size=606348325, level=-606348325, prop=5185, kind=48 '0', nid=15842497851538791387) at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:574 574 shredCtx->dstBAT[PRE_SIZE].cast.intCAST[pre] = size; Missing separate debuginfos, use: debuginfo-install bzip2.x86_64 e2fsprogs.x86_64 glibc.x86_64 keyutils.x86_64 krb5.x86_64 libselinux.x86_64 libxml2.x86_64 openssl.x86_64 pcre.x86_64 zlib.x86_64 (gdb) bt #0 0x00002aaab9bb22f5 in emit_tuple (shredCtx=0x18a1508, pre=15842497851538791387, size=606348325, level=-606348325, prop=5185, kind=48 '0', nid=15842497851538791387) at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:574 #1 0x00002aaab9bb23c5 in emit_node (shredCtx=0x18a1508, node=0x18ba360) at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:626 #2 0x00002aaab9bb35be in shred_end_document (xmlCtx=0x18a1508) at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:822 #3 0x00000038aae48cc8 in xmlParseDocument () from /usr/lib64/libxml2.so.2 #4 0x00002aaab9bb4fdb in shredder_parse (shredCtx=0x18a1508, location=0x1006370 "/export/scratch0/lsidir/xmldocs/empty.xml", buffer=0x0, s=0x0) at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:1479 #5 0x00002aaab9bb71d9 in shred (docBAT=0x127c8c8, location=0x1006370 "/export/scratch0/lsidir/xmldocs/empty.xml", buffer=0x0, s=0x0, percentage=0, serFun=0x0, serCtx=0x0, collLock=0x427f8890) at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:1985 #6 0x00002aaab9bb7422 in CMDshred_url (docBAT=0x127c8c8, location=0x1006370 "/export/scratch0/lsidir/xmldocs/empty.xml", percentage=0x427f8880, collLock=0x427f8890, verbose=0x427f88a0 "") at /home/lsidir/develop//current/pathfinder/runtime/shredder.mx:2013 #7 0x00002aaab9b70753 in CMDshred_url_unpack524513197 (argc=6, argv=0x427f8850) at pf_support.glue.c:52 You have to try 3-4 times to shred the same empty doc before you get the segmentation fault. The segmentation fault is caused when the call back function "shred_end_document" is called by xmllib2, so it might be some leftovers from previous runs that have not been freed correctly. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=482468&aid=2017862&group_id=56967 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Monetdb-bugs mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/monetdb-bugs
