On Wed, Jul 30, 2008 at 08:02:16AM +0000, Martin Kersten wrote:
> Update of /cvsroot/monetdb/MonetDB5/src/modules/atoms
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv22933/atoms
>
> Modified Files:
> blob.mx str.mx xml.mx
> Log Message:
> Silenceing the compiler on windows.
> Strings and BATcount are always less then 4GB
Even if so, int (signed) allows only 2^31 - 1, i.e., "2GB".
Moreover, in the 64-bit case (in particular with 64-bit OIDs), neither the
length of strings nor the tuple (BUN) count of BATs is formally restricted
to 2^31 - 1 (or 2^32 - 1) --- in fact, breaking through this limit is one of
the reasons for / features of the 64-bit port of MonetDB --- how efficiently
such larger strings and BATs are handled is a different question ...
That's why I used to guard all down-cast that I added with an assertion to
detect overflows ...
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 |
On Mon, Jul 28, 2008 at 06:10:05AM +0000, Martin Kersten wrote:
> Update of /cvsroot/monetdb/MonetDB5/src/modules/kernel
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv6684/kernel
>
> Modified Files:
> batcalc.mx
> Log Message:
> Oid unary accums are limited to '+'
Since there is IMHO no unary '+' operator defined in MAL, this is unused
("dead") code.
(In fact, I could not tell what the purpose/semantic of a unary '+' operator
should be...)
Moreover, the current code (for good?) also disables unary "--" & "++" for
type OID; hence, the current code is equivalent to a simple
" case TYPE_oid: throw(MAL, "batcalc.@1", "unexpected operator"); break;"
and could/should be simplified like this.
In fact, the part of the code that has been modified, i.e., the
"CMDunaryAccum@1" functions, are not used at all, yet, only the
"CMDunary@1" are used, but is actually the one that triggers
".\..\..\..\..\src\modules\kernel\batcalc.mx(923) : warning C4146: unary minus operator applied to unsigned type, result still unsigned"
I'd suggest to disable (remove?) the unused ("dead") code fo not (yet?) used
"CMDunaryAccum@1" and cleanup / simplify the code for "CMDunary@1" as
suggested above.
Stefan
> U batcalc.mx
> Index: batcalc.mx
> ===================================================================
> RCS file: /cvsroot/monetdb/MonetDB5/src/modules/kernel/batcalc.mx,v
> retrieving revision 1.152
> retrieving revision 1.153
> diff -u -d -r1.152 -r1.153
> --- batcalc.mx 21 Jul 2008 09:28:55 -0000 1.152
> +++ batcalc.mx 28 Jul 2008 06:10:02 -0000 1.153
> @@ -937,10 +937,22 @@
>
> switch(BATttype(bn)){
> case TYPE_oid: if ( strcmp("@2","+")== 0){
> - @:accumUnary(oid,@2)@
> - } else
> - throw(MAL, "batcalc.@1", "unexpected operator");
> - break;
> + oid* p = (oid*)Tloc(b,BUNfirst(b));
> + oid *q = (oid*)Tloc(b,BUNlast(b));
> + oid *o = (oid*)Tloc(bn,BUNfirst(bn));
> +
> + for(; p<q; p++, o++){
> + /* this code is better as x?nil:y */
> +#ifdef NULLTST
> + if(*p == oid_nil )
> + *o = oid_nil;
> + else
> +#endif
> + *o = *p;
> + }
> + } else
> + throw(MAL, "batcalc.@1", "unexpected operator");
> + break;
> case TYPE_bte: @:accumUnary(bte,@2)@ break;
> case TYPE_sht: @:accumUnary(sht,@2)@ break;
> case TYPE_int: @:accumUnary(int,@2)@ break;
>
>
> -------------------------------------------------------------------------
> 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 |
Hi Stefan,
Thanks for cleaning up the code. However, it is referred in MIL, so I
think
this change will break loading the module. This BAT should be used by
xrpc_handle_file_request() for security check. I should add code to
do this,
so I'd suggest to undo this change. What do you think?
Regards,
Jennie
On Jul 23, 2008, at 12:37 , Stefan Manegold wrote:
> Update of /cvsroot/monetdb/pathfinder/runtime
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv9875/
> pathfinder/runtime
>
> Modified Files:
> xrpc_server.mx
> Log Message:
>
> clean-up: removing unused global variable
>
>
> U xrpc_server.mx
> Index: xrpc_server.mx
> ===================================================================
> RCS file: /cvsroot/monetdb/pathfinder/runtime/xrpc_server.mx,v
> retrieving revision 1.82
> retrieving revision 1.83
> diff -u -d -r1.82 -r1.83
> --- xrpc_server.mx 17 Jul 2008 21:28:28 -0000 1.82
> +++ xrpc_server.mx 23 Jul 2008 10:37:49 -0000 1.83
> @@ -194,7 +194,7 @@
> static MT_Sema xrpc_commit_sema;
> static xrpc_commit_t *xrpc_commit_active = NULL;
>
> -BAT *xrpc_trusted = NULL, *xrpc_admin = NULL, *xrpc_user = NULL;
> +BAT *xrpc_trusted = NULL, *xrpc_admin = NULL;
> char *xrpc_hostname = NULL;
> int xrpc_port = 0;
>
> @@ -889,7 +889,6 @@
> @:xrpc_bat(admin,str)@
> xrpc_admin = xrpc_bat;
> @:xrpc_bat(user,str)@
> - xrpc_user = xrpc_bat;
>
> CMDmy_hostname(&xrpc_hostname);
> MT_init_sema(xrpc_commit_sema,0, "XRPC_COMMIT_SEMA");
>
>
> -------------------------------------------------------------------------
> 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-pf-checkins mailing list
> Monetdb-pf-checkins(a)lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/monetdb-pf-checkins
On Tue, Jul 22, 2008 at 03:48:02PM +0000, Sjoerd Mullender wrote:
> Update of /cvsroot/monetdb/MonetDB5/src/mal
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv5497
>
> Modified Files:
> mal_recycle.mx
> Log Message:
> Fix so that initialization happens *exactly* once.
> The initialization of a per-process (a.k.a. global) varirable should
> be governed by a per-process variable, not a per-thread or
> per-something-else variable. Especially not if that variable isn't
> actually used to indicate whether initialization took place (see
> OPTrecycleImplementation in src/optimizer/opt_recycler.mx where
> version was set to -1 which broke the use here.
>
This fixes most of the numerous SQL tests that were failing on Windows (with
M5 server, only) --- thank you very much for caring and taking care!
Stefan
>
> U mal_recycle.mx
> Index: mal_recycle.mx
> ===================================================================
> RCS file: /cvsroot/monetdb/MonetDB5/src/mal/mal_recycle.mx,v
> retrieving revision 1.120
> retrieving revision 1.121
> diff -u -d -r1.120 -r1.121
> --- mal_recycle.mx 22 Jul 2008 06:27:53 -0000 1.120
> +++ mal_recycle.mx 22 Jul 2008 15:47:58 -0000 1.121
> @@ -244,10 +244,13 @@
>
> void RECYCLEversion(Client cntxt, MalBlkPtr mb)
> {
> + static int recycleLockInitialized = 0;
> int i;
>
> - if (mb->version == 0)
> + if (!recycleLockInitialized) {
> MT_init_lock(recycleLock,"recycleLock");
> + recycleLockInitialized = 1;
> + }
> if ( mb->version && mb->version != recycleVersion) {
> mal_set_lock(recycleLock,"recycle");
> /* all re-used instructions are dropped */
> @@ -257,8 +260,8 @@
> (void) cntxt;
> #endif
> for(i=1; i<mb->stop; i++)
> - if (getInstrPtr(mb,i)->recycle)
> - getInstrPtr(mb,i)->recycle = -1;
> + if (getInstrPtr(mb,i)->recycle)
> + getInstrPtr(mb,i)->recycle = -1;
> mal_unset_lock(recycleLock,"recycle");
> }
> mb->version = recycleVersion;
>
>
> -------------------------------------------------------------------------
> 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 |
Martin,
prior to this checkin, this test used to work fine (i.e., create no
differences) both with Stable & Current.
After this checkin to the Stable branch (and its propagation to the Current
trunk), the test does fail as follows:
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/MonetDB5/.mTests103/G…http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB5/.mTests103/…
Has the output been incorrectly approved?
Or is the output as expected, but some related code buggy?
I would be very thankful, if you could check & fix this problem.
Stefan
On Sat, Jul 19, 2008 at 05:25:27PM +0000, Martin Kersten wrote:
> Update of /cvsroot/monetdb/MonetDB5/tests/gdkTests/Tests
> In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv8526/tests/gdkTests/Tests
>
> Modified Files:
> Tag: MonetDB_5-6
> append_dense_tail.stable.out
> Log Message:
> Approved.
>
>
> U append_dense_tail.stable.out
> Index: append_dense_tail.stable.out
> ===================================================================
> RCS file: /cvsroot/monetdb/MonetDB5/tests/gdkTests/Tests/append_dense_tail.stable.out,v
> retrieving revision 1.5.2.1
> retrieving revision 1.5.2.2
> diff -u -d -r1.5.2.1 -r1.5.2.2
> --- append_dense_tail.stable.out 10 Jul 2008 18:34:35 -0000 1.5.2.1
> +++ append_dense_tail.stable.out 19 Jul 2008 17:25:25 -0000 1.5.2.2
> @@ -80,10 +80,10 @@
> # h t # name
> # str str # type
> #-------------------------#
> -[ "hdense", "1" ]
> [ "hseqbase", "0@0" ]
> -[ "tdense", "0" ]
> +[ "hdense", "1" ]
> [ "tseqbase", "0@0" ]
> +[ "tdense", "0" ]
>
> # 09:01:01 >
> # 09:01:01 > 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 |