Some assertion in BATsetcount where the count is larger then the capacity triggered by some select (in the BSM queries). Occurs only with imrints enabled, but I have not yet managed to understand why, how, and where what actually goes wrong ... to be continued once there is time ...

----- Original Message -----

Hai Stefan,

What problems is imprints causing in the sciql branch?

Jennie

On Jun 14, 2013, at 10:29, Stefan Manegold wrote:

...

Changeset: 3841d3bd5fd7 for MonetDB URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=3841d3bd5fd7 Modified Files: gdk/gdk_select.c Branch: sciql Log Message:

for now, disbale imprints in sciql branch, until we find the time to debug suspicious errors

diffs (12 lines):

diff --git a/gdk/gdk_select.c b/gdk/gdk_select.c --- a/gdk/gdk_select.c +++ b/gdk/gdk_select.c @@ -1344,7 +1344,7 @@ BATsubselect(BAT *b, BAT *s, const void * ii) it is not an equi-select, and * iii) is not var-sized. */

• 	use_imprints = 1;
  

• 	use_imprints = 0; /* temporarily disabled in sciql branch */
  

  } bn = BAT_scanselect(b, s, bn, tl, th, li, hi, equi, anti, lval, hval, maximum, use_imprints);

---

checkin-list mailing list checkin-list@monetdb.org http://mail.monetdb.org/mailman/listinfo/checkin-list