- developers-list - monetdb.org

[Monetdb-developers] Error in CVS / sql
by Stefan de Konink 21 Aug '08

21 Aug '08

Due to a bug in Monet5, I am forced to run a more recent version, compiling the sql cvs module has the following result: libtool: compile: gcc -DHAVE_CONFIG_H -I. -I../../../../src/backends/monet5 -I../../.. -I../../../../src/backends/monet5 -I../../include -I../../../../src/backends/monet5/../../include -I../../common -I../../../../src/backends/monet5/../../common -I../../storage -I../../../../src/backends/monet5/../../storage -I../../server -I../../../../src/backends/monet5/../../server -I/opt/cherokee/include/MonetDB5/atoms -I/opt/cherokee/include/MonetDB5/compiler -I/opt/cherokee/include/MonetDB5/kernel -I/opt/cherokee/include/MonetDB5/mal -I/opt/cherokee/include/MonetDB5/optimizer -I/opt/cherokee/include/MonetDB5/scheduler -I/opt/cherokee/include/MonetDB5 -I/opt/cherokee/include/MonetDB -I/opt/cherokee/include/MonetDB/mapilib -I/opt/cherokee/include/MonetDB -I/opt/cherokee/include/MonetDB/common -I/opt/cherokee/include/MonetDB/gdk -DLIBSQL -Wall -Wextra -std=c99 -g -Werror-implicit-function-declaration -Werror -Wpointer-arith -Wdeclaration-after-statement -Wundef -Wp,-D_FORTIFY_SOURCE=2 -c sql_bpm.c -fPIC -DPIC -o .libs/lib_sql_la-sql_bpm.o cc1: warnings being treated as errors ../../../../src/backends/monet5/sql_bpm.mx: In functie ‘send_part’: ../../../../src/backends/monet5/sql_bpm.mx:272: fout: impliciete declaratie van functie ‘RMTfindconn’ ../../../../src/backends/monet5/sql_bpm.mx:272: fout: assignment makes pointer from integer without a cast ../../../../src/backends/monet5/sql_bpm.mx:276: fout: impliciete declaratie van functie ‘RMTinternalput’ ../../../../src/backends/monet5/sql_bpm.mx:276: fout: assignment makes pointer from integer without a cast ../../../../src/backends/monet5/sql_bpm.mx: In functie ‘get_part’: ../../../../src/backends/monet5/sql_bpm.mx:339: fout: assignment makes pointer from integer without a cast ../../../../src/backends/monet5/sql_bpm.mx:343: fout: impliciete declaratie van functie ‘RMTinternalget’ ../../../../src/backends/monet5/sql_bpm.mx:343: fout: assignment makes pointer from integer without a cast make[7]: *** [lib_sql_la-sql_bpm.lo] Fout 1 make[7]: Map '/mnt/dv/skinkie/monetdb/sql/Linux/src/backends/monet5' wordt verlaten make[6]: *** [all] Fout 2

2 3

Re: [Monetdb-developers] MonetDB5/src/modules/atoms url.mx, , 1.36, 1.37
by Fabian Groffen 15 Aug '08

15 Aug '08

On 14-08-2008 15:44:44 +0000, Lefteris Sidirourgos wrote: > Update of /cvsroot/monetdb/MonetDB5/src/modules/atoms > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv29402 > > Modified Files: > url.mx > Log Message: > Also deleted trailing spaces, alligned text (removing tabs...). > Fixed documentation/comments. Hi Lefteris, Please remember that the MonetDB code style uses tabs, not evil spaces. Unfortunately the Pathfinder sources do not comply to the MonetDB code style (in more than just spaces and tabs).

2 2

[Monetdb-developers] Announcement: new minor release of MonetDB Clients package
by Sjoerd Mullender 12 Aug '08

12 Aug '08

The MonetDB team at CWI is pleased to announve a bug fix release for the MonetDB Clients package. A number of bugs in the ODBC driver have been fixed: - fixes for the handling of UTF-8 encoded queries and query parameters; - fixes for dealing with wide characters. There is also a fix for mclient: - the timer (-t/--time option) could overflow. There are new Windows installers for the ODBC driver, and there are new RPMs for Fedora 8 and 9. All of this for both 32 and 64 bit architectures. There is also a new Super Source Ball. -- Sjoerd Mullender

1 0

[Monetdb-developers] Some questions
by Elad . 11 Aug '08

11 Aug '08

Hi there! A few questions: 1. Is it possible to run the server module without a visible console on Windows? 2. When inserting large bulks using the ODBC driver, the server process terminates without warning when the bulk is too large to fit in memory (on a 32 bit machine) and the ODBC driver throws a strange exception. I assume this will also happen with large SELECT queries as well. Is there a way to uniquely identify this scenario and/or avoid it? 3. What is the best way of determining whether the server process is up and running? 4. What is the best way to set up my own version of Monet? I want to take a look at the code and perhaps package the server a little differently (create a different MSI, change the EXE file name, etc). Which IDE would be best to do this? Thanks! _________________________________________________________________ Got Game? Win Prizes in the Windows Live Hotmail Mobile Summer Games Trivia Contest http://www.gowindowslive.com/summergames?ocid=TXT_TAGHM

1 0

[Monetdb-developers] Announcing MonetDB support for Ruby on Rails
by Michalis Polakis 10 Aug '08

10 Aug '08

Hello folks, I developed the MonetDB adapter for Rail's ActiveRecord. It will enable you to work with MonetDB databases in Ruby on Rails. You are more than welcome to try the code, give me some feedback, and maybe help me maintain it. To get the code visit : http://rubyforge.org/projects/monetdb-ror Thanks, Michalis Polakis

1 0

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/modules/mal/Tests remote02.stable.out, , 1.12, 1.13
by Stefan Manegold 10 Aug '08

10 Aug '08

FYI: the last 4 checkins on this file (mainly) consisted of alternatingly removing and adding (the same) part of the output: http://monetdb.cvs.sourceforge.net/monetdb/MonetDB5/src/modules/mal/Tests/r… http://monetdb.cvs.sourceforge.net/monetdb/MonetDB5/src/modules/mal/Tests/r… http://monetdb.cvs.sourceforge.net/monetdb/MonetDB5/src/modules/mal/Tests/r… http://monetdb.cvs.sourceforge.net/monetdb/MonetDB5/src/modules/mal/Tests/r… http://monetdb.cvs.sourceforge.net/monetdb/MonetDB5/src/modules/mal/Tests/r… Most of the log messages unfortunately give hardly any hint as to whether (and if so, why) this was intended or whether this was by "accident" / "sloppyness" ... Stefan On Sun, Aug 10, 2008 at 11:19:31AM +0000, Fabian wrote: > Update of /cvsroot/monetdb/MonetDB5/src/modules/mal/Tests > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv30947/Tests > > Modified Files: > remote02.stable.out > Log Message: > restore (accidently?) wiped out test output > -- | 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 |

2 1

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/mal mal_debugger.mx, , 1.232, 1.233 mal_recycle.mx, , 1.134, 1.135
by Sjoerd Mullender 08 Aug '08

08 Aug '08

On 2008-08-08 18:30, Stefan Manegold wrote: > Update of /cvsroot/monetdb/MonetDB5/src/mal > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv22332/src/mal > > Modified Files: > mal_debugger.mx mal_recycle.mx > Log Message: > > acknowledge the type that functions return (for good reasons); > for BATcount(), it is size_t . @@ -789,8 +789,8 @@ > > static void > printBATproperties(stream *f, BAT *b){ > - stream_printf(f, " count=%d lrefs=%d ", > - (int) BATcount(b), BBP_lrefs(ABS(b->batCacheid))); > + stream_printf(f, " count=%d lrefs=" SZFMT " ", > + BATcount(b), BBP_lrefs(ABS(b->batCacheid))); > if( BBP_refs(ABS(b->batCacheid)) - 1) > stream_printf(f, " refs=%d ", BBP_refs(ABS(b->batCacheid)) - 1); > if( b->batSharecnt) > I think you got this backwards: it should be count=SZFMT , lrefs=%d. -- Sjoerd Mullender

1 0

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/mal/Tests tst052.mal, , 1.1, 1.2 tst052.stable.out, , 1.1, 1.2
by Stefan Manegold 05 Aug '08

05 Aug '08

On Tue, Aug 05, 2008 at 06:22:32AM +0000, Martin Kersten wrote: > Update of /cvsroot/monetdb/MonetDB5/src/mal/Tests > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv9480 > > Modified Files: > tst052.mal tst052.stable.out > Log Message: > simplify test. > Martin, wouldn't it have been better to keep the check whether type wrd is handled correctly on both 32-bit and 64-bit systems and approve the (correct!) 32-bit specific output instead of simply removing this check? Stefan > > U tst052.mal > Index: tst052.mal > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/mal/Tests/tst052.mal,v > retrieving revision 1.1 > retrieving revision 1.2 > diff -u -d -r1.1 -r1.2 > --- tst052.mal 4 Aug 2008 06:57:09 -0000 1.1 > +++ tst052.mal 5 Aug 2008 06:22:30 -0000 1.2 > @@ -7,6 +7,5 @@ > bat.insert(b,7,127:wrd); > bat.insert(b,15,32768:wrd); > bat.insert(b,31,2147483647:wrd); > -bat.insert(b,63,9223372036854775807:wrd); > > io.print(b); > > U tst052.stable.out > Index: tst052.stable.out > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/mal/Tests/tst052.stable.out,v > retrieving revision 1.1 > retrieving revision 1.2 > diff -u -d -r1.1 -r1.2 > --- tst052.stable.out 4 Aug 2008 06:57:09 -0000 1.1 > +++ tst052.stable.out 5 Aug 2008 06:22:30 -0000 1.2 > @@ -32,10 +32,9 @@ > bat.insert(b,7,127:wrd); # 5 _9:void := BKCinsert_bun(b:bat[:int,:wrd], _10:int, _11:wrd) > bat.insert(b,15,32768:wrd); # 6 _12:void := BKCinsert_bun(b:bat[:int,:wrd], _13:int, _14:wrd) > bat.insert(b,31,2147483647:wrd); # 7 _15:void := BKCinsert_bun(b:bat[:int,:wrd], _16:int, _17:wrd) > - bat.insert(b,63,9223372036854775807:wrd);# 8 _18:void := BKCinsert_bun(b:bat[:int,:wrd], _19:int, _20:wrd) > - io.print(b); # 9 _21:void := IOprint_val(b:bat[:int,:wrd]) > -end main; # 10 > -#---------------------------------# > + io.print(b); # 8 _18:void := IOprint_val(b:bat[:int,:wrd]) > +end main; # 9 > +#-------------------------# > # h t # name > # int wrd # type > #---------------------------------# > @@ -43,7 +42,6 @@ > [ 7, 127 ] > [ 15, 32768 ] > [ 31, 2147483647 ] > -[ 63, 9223372036854775807 ] > > # 08:54:36 > > # 08:54:36 > Done. > > > ------------------------------------------------------------------------- > 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-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 |

2 1

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/mal mal_profiler.mx, , 1.108, 1.109
by Stefan Manegold 05 Aug '08

05 Aug '08

On Mon, Aug 04, 2008 at 11:46:50PM +0000, Martin Kersten wrote: > Update of /cvsroot/monetdb/MonetDB5/src/mal > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv18736 > > Modified Files: > mal_profiler.mx > Log Message: > Use a larger type for the diskspace Martin, thank for taking care of this one, and my apologies for not having suggested a mu simple solution in my bug report. In detail: I'm in fact afraid that your fix / new solution will not even compile on Windows. The reason is, that the C function getDiskSpace() internally (still) collects / calculates the occupied disk space in a variable of type "lng" --- this is (and was) good & correct, as even on 32-bit systems, MonetDB's (total) database size can grow beyond 2 GB --- however, C function getDiskSpace() now returns a "size_t" (while it used to correctly return the internal "lng"). Not only is "size_t" only 32 bit on 32-bit systems, while "lng" is 64-bit also on 32-bit systems, but it is also unsigned while lng is signed. Hence, the Windows compilers will (at least) complain about the possible loos of data when implicitly truncating the 64-bit (signed) "lng" to a 32-bit (unsigned) "size_t" in the return statement of C function getDiskSpace() on 32-bit systems. Moreover, it is a bit "odd" to use unsigned type "size_t" in the undelying C function, but signed type "wrd" in the MAL command --- why not also use type "wrd" in C -- we have it defined (for exacly this reason) as any other "three-letter-type" we use in GDK ... In fact, the IMHO correct (and simpler/simplest) solution would have been to leave the C function getDiskSpace() return a "lng" as it (correctly) used to do, and simply change to MAL command signature and wrapper to directly pass this "lng" through as "lng" instead of truncating it (on both 32- & 64-bit systems) to an "int". Again, my apologies for not explicitly pointing out this simple solution in my bug report. Stefan ps: If you wish, I could take care of fixing this ... > U mal_profiler.mx > Index: mal_profiler.mx > =================================================================== > RCS file: /cvsroot/monetdb/MonetDB5/src/mal/mal_profiler.mx,v > retrieving revision 1.108 > retrieving revision 1.109 > diff -u -d -r1.108 -r1.109 > --- mal_profiler.mx 4 Aug 2008 06:56:27 -0000 1.108 > +++ mal_profiler.mx 4 Aug 2008 23:46:48 -0000 1.109 > @@ -213,7 +213,7 @@ > mal_export void clearTrace(void); > mal_export BAT *getTrace(str ev); > > -mal_export lng getDiskSpace(void); > +mal_export size_t getDiskSpace(void); > #define DEBUG_MAL_PROFILER > > #endif > @@ -550,7 +550,7 @@ > log("%d,\t", mb->profiler[pc].obytes); > } > if (profileCounter[PROFdisk].status) { > - log("%d,\t", getDiskSpace()); > + log(SZFMT",\t", getDiskSpace()); > } > if (profileCounter[PROFaggr].status) > log2("%d,\t%d,\t", mb->profiler[pc].counter, mb->profiler[pc].clk); > @@ -950,7 +950,7 @@ > TRACE_id_stmt = TRACEcreate("id", "statement", TYPE_int, TYPE_str); > TRACE_id_ibytes = TRACEcreate("id", "ibytes", TYPE_int, TYPE_wrd); > TRACE_id_obytes = TRACEcreate("id", "obytes", TYPE_int, TYPE_wrd); > - TRACE_id_diskspace = TRACEcreate("id", "diskspace", TYPE_int, TYPE_lng); > + TRACE_id_diskspace = TRACEcreate("id", "diskspace", TYPE_int, TYPE_wrd); > TRACE_init = 1; > mal_unset_lock(mal_contextLock, "trace"); > } > @@ -1132,7 +1132,7 @@ > BUNins(TRACE_id_obytes, &evnt, &mb->profiler[pc].obytes, FALSE); > } > if (profileCounter[PROFdisk].status) { > - lng disk= getDiskSpace(); > + size_t disk= getDiskSpace(); > BUNins(TRACE_id_diskspace, &evnt, &disk, FALSE); > } > eventcounter++; > @@ -1147,7 +1147,7 @@ > if( mb->profiler) return; > mb->profiler = (ProfPtr) GDKzalloc(mb->ssize*sizeof(ProfRecord)); > } > -lng > +size_t > getDiskSpace() > { > BAT *b; > > > ------------------------------------------------------------------------- > 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-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

Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/modules/atoms rmtobj.mx, , 1.4, 1.5
by Sjoerd Mullender 04 Aug '08

04 Aug '08

Fabian wrote: > Update of /cvsroot/monetdb/MonetDB5/src/modules/atoms > In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv22761 > > Modified Files: > rmtobj.mx > Log Message: > For what its worth, apply my diffs after merging Martin's changes. I > basically had the same fixes. > > > @@ -86,11 +86,11 @@ > int > RMTOtoString(str *retval, int *len, str handle) > { > - int hl = (int) strlen(handle)+1; > + int hl = (int)strlen(handle) + 1; > if (*len < hl) { > if (*retval != NULL) > GDKfree(*retval); > - *retval = GDKmalloc(hl); > + *retval = GDKmalloc(sizeof(char) * hl); > } > *len = hl; > memcpy(*retval, handle, hl); I've seen this kind of thing quite often, but I have to ask, why? The standard says about the sizeof operator: "When applied to an operand that has type char, unsigned char, or signed char, (or a qualified version thereof) the result is 1." So why bother multiplying with sizeof(char)? -- Sjoerd Mullender

2 1