[Monetdb-developers] check-in "killed" pf/tijah
hej, one of the check-ins yesterday more or less "killed" pf/tijah. the test-set is now completely red for our module. since i also checked in one line of code yesterday, i tested already whether it was the cause of the trouble, but undoing my check-in does not solve anything. so, i suppose there must be some other reason. -henning
Hi Henning, do you use ws_create()? -- If that's the case the comment in pathfinder.mx was incorrect and the proc ws_create() -> ws_create(0) was still needed. Jan On 05/07/2008 10:11 AM, Henning Rode wrote with possible deletions:
hej,
one of the check-ins yesterday more or less "killed" pf/tijah. the test-set is now completely red for our module. since i also checked in one line of code yesterday, i tested already whether it was the cause of the trouble, but undoing my check-in does not solve anything. so, i suppose there must be some other reason.
-henning
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers
-- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
But here we are not talking about the algebra backend, but about MPS.
From what I saw all fail with a message like:
QUERY = let $opt := <TijahOptions ERROR = !ERROR: leftfetchjoin(param 2): invalid BAT. # the original non-normalized output was: ERROR = !ERROR: interpret_params: leftfetchjoin(param 2): invalid BAT. !ERROR: reverse(param 1): evaluation error. # the original non-normalized output was: !ERROR: interpret_params: reverse(param 1): evaluation error. !ERROR: execute MIL failed. # the original non-normalized output was: !ERROR: CMDtijah_query: execute MIL failed. = !ERROR: CMDtijah_query: operation failed. On Wed, May 7, 2008 at 11:25 AM, Jan Rittinger <rittinge@in.tum.de> wrote:
Hi Henning,
do you use ws_create()? -- If that's the case the comment in pathfinder.mx was incorrect and the proc ws_create() -> ws_create(0) was still needed.
Jan
On 05/07/2008 10:11 AM, Henning Rode wrote with possible deletions:
hej,
one of the check-ins yesterday more or less "killed" pf/tijah. the test-set is now completely red for our module. since i also checked in one line of code yesterday, i tested already whether it was the cause of the trouble, but undoing my check-in does not solve anything. so, i suppose there must be some other reason.
-henning
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers
-- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao... _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers
On 05/07/2008 10:54 AM, Lefteris wrote with possible deletions:
But here we are not talking about the algebra backend, but about MPS.
From what I saw all fail with a message like:
QUERY = let $opt := <TijahOptions ERROR = !ERROR: leftfetchjoin(param 2): invalid BAT. # the original non-normalized output was: ERROR = !ERROR: interpret_params: leftfetchjoin(param 2): invalid BAT. !ERROR: reverse(param 1): evaluation error. # the original non-normalized output was: !ERROR: interpret_params: reverse(param 1): evaluation error. !ERROR: execute MIL failed. # the original non-normalized output was: !ERROR: CMDtijah_query: execute MIL failed. = !ERROR: CMDtijah_query: operation failed.
This is probably a problem of Tijah that it cannot handle certain functions if some loading (see test loadthesis) did not work. -- Jan Rittinger Database Systems Technische Universität München (Germany) http://www-db.in.tum.de/~rittinge/
exactly, though i must admit that the error message in this case is far from being helpful. -henning Jan Rittinger wrote:
On 05/07/2008 10:54 AM, Lefteris wrote with possible deletions:
But here we are not talking about the algebra backend, but about MPS.
From what I saw all fail with a message like:
QUERY = let $opt := <TijahOptions ERROR = !ERROR: leftfetchjoin(param 2): invalid BAT. # the original non-normalized output was: ERROR = !ERROR: interpret_params: leftfetchjoin(param 2): invalid BAT. !ERROR: reverse(param 1): evaluation error. # the original non-normalized output was: !ERROR: interpret_params: reverse(param 1): evaluation error. !ERROR: execute MIL failed. # the original non-normalized output was: !ERROR: CMDtijah_query: execute MIL failed. = !ERROR: CMDtijah_query: operation failed.
This is probably a problem of Tijah that it cannot handle certain functions if some loading (see test loadthesis) did not work.
ok, using ws_create(0) instead of ws_create() helped already a lot. now i only get bat-property warnings in two cases which did not appear before. thanks -henning Jan Rittinger wrote:
Hi Henning,
do you use ws_create()? -- If that's the case the comment in pathfinder.mx was incorrect and the proc ws_create() -> ws_create(0) was still needed.
Jan
On 05/07/2008 10:11 AM, Henning Rode wrote with possible deletions:
hej,
one of the check-ins yesterday more or less "killed" pf/tijah. the test-set is now completely red for our module. since i also checked in one line of code yesterday, i tested already whether it was the cause of the trouble, but undoing my check-in does not solve anything. so, i suppose there must be some other reason.
-henning
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javao...
_______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers
participants (3)
-
Henning Rode -
Jan Rittinger -
Lefteris