- developers-list - monetdb.org

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/optimizer opt_replicator.mx, , 1.30, 1.31
by Stefan Manegold 03 Mar '08

03 Mar '08

On Mon, Mar 03, 2008 at 02:35:30PM +0000, Milena Ivanova wrote: > Update of /cvsroot/monetdb/MonetDB5/src/optimizer > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv4109/src/optimizer > > Modified Files: > opt_replicator.mx > Log Message: > Make sure the constant exists in var area of replicator. Just a simple question out of curiosity: It seems, the new variable v1 is only use in the new if-statement's body, right? Hence, why don't you also declare the variable v1 as "local" variable in the if-body? The is no need to have it as "global" variable in function newReplica(), is there? Stefan > Index: opt_replicator.mx > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/optimizer/opt_replicator.mx,v > retrieving revision 1.30 > retrieving revision 1.31 > diff -u -d -r1.30 -r1.31 > --- opt_replicator.mx 2 Mar 2008 12:08:46 -0000 1.30 > +++ opt_replicator.mx 3 Mar 2008 14:35:27 -0000 1.31 > @@ -438,7 +438,7 @@ > static int newReplica(MalStkPtr s, InstrPtr p, bit keep) > { > int i, j, k, c, bid, r; > - ValPtr v; > + ValPtr v, v1; > InstrPtr p1; > > assureSpace(); > @@ -446,7 +446,12 @@ > for(i = 0; i< p->argc; i++){ > j = p->argv[i]; > v = &s->stk[j]; > - c = defConstant(repl, v->vtype, v); > + c = fndConstant(repl, v); > + if ( c < 0 ){ > + v1 = VALnew(); > + VALcopy(v1,v); > + c = defConstant(repl, v1->vtype, v1); > + } > setArg(p1,i,c); > > if ((i > 0) || > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Monetdb-checkins mailing list > Monetdb-checkins(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/monetdb-checkins > -- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |

1 0

[Monetdb-developers] Visual Studio Solution/Project
by mobigital1 03 Mar '08

03 Mar '08

Does anyone know if it's possible to include a capability to generate a set of Visual Studio (2005 or 2008) project files for all components, together with solution file it would be really great to have. Because it would make it much easier to debug it on windows. perhaps someone already made such projects (even if they are outdated is ok too, I can modify to bring them up to date). -- View this message in context: http://www.nabble.com/Visual-Studio-Solution-Project-tp15640939p15640939.ht… Sent from the monetdb-developers mailing list archive at Nabble.com.

2 3

[Monetdb-developers] suggestions on debugging/tracing
by mobigital-nbl＠yahoo.com 02 Mar '08

02 Mar '08

Hi, I am trying to investigate this bug (https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1905467&group_…) which i submitted earlier. I was importing records with COPY TO a staging table in batches of 250K rows. Then copying them into final table. After completing 4,750,000 rows and importing a new batch thebelow exception occurs in GDKreallocmax at the last line of the snippetbelow: oldsize = GDK_MEM_BLKSIZE(s); with "s"=0x00000000100e0000" GDK_MEM_BLKSIZE macro tries to read 8 bytes prior to "s", which is 0x00000000100dfff8 and turns out to be unallocated. I am thinking the h->base of the heap is either free'ed at some point before being used or may be it is set to incorrect value? in any case i wanted to add some tracing to Heap to maintain some history of base change (and who did the change), so that when I hit another access violation during debugging I can use that to pinpoint why Heap base points to unallocated space. i wanted to see if anyone has already used something like that so that I can get a headstart quicker. if anyone have any suggestions on fixing it would be appreciated. Cheers.

1 0

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/optimizer opt_replicator.mx, , 1.28, 1.29
by Milena.Ivanova＠cwi.nl 02 Mar '08

02 Mar '08

> Update of /cvsroot/monetdb/MonetDB5/src/optimizer > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv3358 > > Modified Files: > opt_replicator.mx > Log Message: > defConstant already performs a search for the constant value. > > This was my first version too, but defConstant sets a pointer and does not copy the value. The next call finds whatever has left in memory and breaks. Milena > Index: opt_replicator.mx > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/optimizer/opt_replicator.mx,v > retrieving revision 1.28 > retrieving revision 1.29 > diff -u -d -r1.28 -r1.29 > --- opt_replicator.mx 29 Feb 2008 15:43:20 -0000 1.28 > +++ opt_replicator.mx 2 Mar 2008 11:46:49 -0000 1.29 > @@ -438,20 +438,15 @@ > static int newReplica(MalStkPtr s, InstrPtr p, bit keep) > { > int i, j, k, c, bid, r; > - ValPtr v, v1; > + ValPtr v; > InstrPtr p1; > > assureSpace(); > p1 = copyInstruction(p); > for(i = 0; i< p->argc; i++){ > j = p->argv[i]; > - v = &s->stk[j]; > - c = fndConstant(repl, v); > - if ( c < 0 ){ > - v1 = VALnew(); > - VALcopy(v1,v); > - c = defConstant(repl, v1->vtype, v1); > - } > + v = &s->stk[j]; > + c = defConstant(repl, v->vtype, v); > setArg(p1,i,c); > > if ((i > 0) || > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Monetdb-checkins mailing list > Monetdb-checkins(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/monetdb-checkins >

1 0

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/optimizer opt_replicator.mx, , 1.23, 1.24
by Milena Ivanova 29 Feb '08

29 Feb '08

Romulo Goncalves wrote: > Update of /cvsroot/monetdb/MonetDB5/src/optimizer > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv18911 > > Modified Files: > opt_replicator.mx > Log Message: > If you stop you should free the structure that you allocked in REPstart... > > Compilation works and also the replicator test... > > > Index: opt_replicator.mx > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/optimizer/opt_replicator.mx,v > retrieving revision 1.23 > retrieving revision 1.24 > diff -u -d -r1.23 -r1.24 > --- opt_replicator.mx 29 Feb 2008 07:29:21 -0000 1.23 > +++ opt_replicator.mx 29 Feb 2008 13:13:20 -0000 1.24 > @@ -132,7 +132,7 @@ > opt_export str REPsemijoin(MalBlkPtr mb, MalStkPtr stk, InstrPtr pci); > opt_export str REPdump(int *ret); > opt_export str REPstart(MalBlkPtr mb, MalStkPtr stk, InstrPtr pci); > -opt_export str REPstop(void); > +opt_export str REPstop(MalStkPtr stk); > opt_export str REPsetRetain(int *ret, sht *p); > opt_export str REPsetReuse(int *ret, sht *p); > opt_export str REPsetCache(int *ret, sht *p); > @@ -242,8 +242,10 @@ > } > > > -str REPstop(void) > +str REPstop(MalStkPtr s) > { > + if( s->blk->replica ) > + GDKfree(((ReplicaRec*)s->blk->replica)->vmap); > replref--; > return MAL_SUCCEED; > } > In principle yes, but in this case I don't think so. The vmap is the map to replicated variables, that can be used next time the same query is executed, which is after rep.stop. Pls, revert Milena > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Monetdb-checkins mailing list > Monetdb-checkins(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/monetdb-checkins >

1 0

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/modules/mal tablet.mx, , 1.114, 1.115
by Romulo Goncalves 29 Feb '08

29 Feb '08

Niels Nes wrote: > Update of /cvsroot/monetdb/MonetDB5/src/modules/mal > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv3888/src/modules/mal > > Modified Files: > tablet.mx > Log Message: > correctly check for null values > > > Index: tablet.mx > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/modules/mal/tablet.mx,v > retrieving revision 1.114 > retrieving revision 1.115 > diff -u -d -r1.114 -r1.115 > --- tablet.mx 18 Feb 2008 19:29:07 -0000 1.114 > +++ tablet.mx 18 Feb 2008 19:57:37 -0000 1.115 > @@ -1264,9 +1264,8 @@ > e = s; > if ( s == NULL || > (s == e && fmt->nullstr[0] == 0) || Would not be better here to do only fmt->nullstr because if fmt->nullstr was initialized with NULL this if clause will crash the mserver... Romulo > - (!quote && strcasecmp(s,fmt->nullstr) == 0) || > - (quote && strncasecmp(s+1, fmt->nullstr, e-s) == 0) > - ) { > + (s != e && !quote && strcasecmp(s,fmt->nullstr) == 0) || > + (s != e && quote && strncasecmp(s+1, fmt->nullstr, e-s) == 0)) { > adt = fmt->data; > memcpy(adt, ATOMnilptr(fmt->adt), fmt->nillen); > } else { > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Monetdb-checkins mailing list > Monetdb-checkins(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/monetdb-checkins

1 0

Re: [Monetdb-developers] load runtime incorrectly
by mobigital-nbl＠yahoo.com 28 Feb '08

28 Feb '08

FYI, this issue turned out to be due to the linking to non-debug MSVCR90.lib in: 1. iconv project. - had to make sure that MDFlags in iconv makefile also changed to /MDd when debug was requested. 2. pthreads - changed /MD to /MDd in flags to each .c compilation to .obj, also added /MDd in final linking and removed explicit linking with libcrt.lib - because /MDd already does it with debug crt. > additional info: > > it seems even after DEBUG build, my executable is looking for non-DEBUG version > of MSVCR90.dll > > --------------------------- > mserver5.exe - Unable To Locate Component > --------------------------- > This application has failed to start because MSVCR90.dll was not found. > Re-installing the application may fix this problem. > --------------------------- > > when i copy the set of MSVCR90.dll and manifest into debug directory, i get the > error shown below.(my prior message) > > here are my build commands that i use for all modules: > > nmake /nologo BITS64=1 NEED_MX=1 DEBUG=1 HAVE_ICONV=1 HAVE_MONETDB4=1 > HAVE_MONETDB5=1 HAVE_PYTHON=1 "prefix=%BUILDTOOLS_PREFIX%" > > nmake /nologo BITS64=1 NEED_MX=1 DEBUG=1 HAVE_ICONV=1 HAVE_MONETDB4=1 > HAVE_MONETDB5=1 HAVE_PYTHON=1 "prefix=%BUILDTOOLS_PREFIX%" install > > > > > Does anyone know why i may be getting this error: > > > > Runtime Error! > > > > Program: C:\monetdb\build\bin\mserver5.exe > > > > R6034 > > > > An application has made an attempt to load the C runtime library incorrectly. > > Please contact the application's support team for more information. > > > > ----------- > > also is there a 'clean' target for make? > > i don't see it anywhere. > > > >

1 0

[Monetdb-developers] FYI: reversed nightly testing order: Current, Stable
by Stefan Manegold 27 Feb '08

27 Feb '08

Dear all, just to keep you posted: As of tonight, nightly MonetDB testing will (again) first run the "Current" development head and then the "Stable" release branch. (With the release out for just over a week, now, "urgent" bug-fixing and TestWeb clean-up has settled, making early-morning "Stable" testing results less needed. Instead, increased feature development activities ask for early-morning "Current" testing results.) Best regards from A'dam, Stefan -- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |

1 0

Re: [Monetdb-developers] load runtime incorrectly
by mobigital-nbl＠yahoo.com 27 Feb '08

27 Feb '08

additional info: it seems even after DEBUG build, my executable is looking for non-DEBUG version of MSVCR90.dll --------------------------- mserver5.exe - Unable To Locate Component --------------------------- This application has failed to start because MSVCR90.dll was not found. Re-installing the application may fix this problem. --------------------------- when i copy the set of MSVCR90.dll and manifest into debug directory, i get the error shown below.(my prior message) here are my build commands that i use for all modules: nmake /nologo BITS64=1 NEED_MX=1 DEBUG=1 HAVE_ICONV=1 HAVE_MONETDB4=1 HAVE_MONETDB5=1 HAVE_PYTHON=1 "prefix=%BUILDTOOLS_PREFIX%" nmake /nologo BITS64=1 NEED_MX=1 DEBUG=1 HAVE_ICONV=1 HAVE_MONETDB4=1 HAVE_MONETDB5=1 HAVE_PYTHON=1 "prefix=%BUILDTOOLS_PREFIX%" install > > Does anyone know why i may be getting this error: > > Runtime Error! > > Program: C:\monetdb\build\bin\mserver5.exe > > R6034 > > An application has made an attempt to load the C runtime library incorrectly. > Please contact the application's support team for more information. > > ----------- > also is there a 'clean' target for make? > i don't see it anywhere. > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2008. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Monetdb-developers mailing list > Monetdb-developers(a)lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/monetdb-developers >

1 0

[Monetdb-developers] load runtime incorrectly
by mobigital-nbl＠yahoo.com 27 Feb '08

27 Feb '08

Does anyone know why i may be getting this error: Runtime Error! Program: C:\monetdb\build\bin\mserver5.exe R6034 An application has made an attempt to load the C runtime library incorrectly. Please contact the application's support team for more information. ----------- also is there a 'clean' target for make? i don't see it anywhere.

1 0