Feature Requests item #1043573, was opened at 2004-10-09 16:18
Message generated for change (Comment added) made by mr-meltdown
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=482471&aid=1043573&group_id=56967

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
>Status: Closed
Priority: 5
Private: No
Submitted By: Fabian (mr-meltdown)
Assigned to: Niels Nes (nielsnes)
Summary: Testweb: SQL tests show minor differences due to MapiClient

Initial Comment:
For some tests (e.g. marcin2) in the SQL repo
significant changes are to be seen, as the MapiClient
utility shows different error warnings than it used to.

The actual change is visible in a query like:

-- this is some explanation
SELECT * FROM WRONG QUERY;

MapiClient used to bail about an error, reporting the
wrong SQL query.  However, nowadays, it bails but shows
the comment instead of the query.

Either some tests have to be reapproved, or the
original behaviour of MapiClient should be restored. 
It might be evident that that behaviour is much more
useful in most cases.

----------------------------------------------------------------------

>Comment By: Fabian (mr-meltdown)
Date: 2009-12-01 12:01

Message:
The intention of this request has been rereported by skinkie recently since
this behaviour even got worse for some situations.

----------------------------------------------------------------------

Comment By: Niels Nes (nielsnes)
Date: 2004-10-16 09:43

Message:
Logged In: YES 
user_id=43556

This bug is releated to not having a proper 'sql'
communication protocol. 
So I'll move this feature request to the proper section. 

----------------------------------------------------------------------

Comment By: Niels Nes (nielsnes)
Date: 2004-10-10 19:48

Message:
Logged In: YES 
user_id=43556

MapiClient shouldn't print the query. The sql server should
include the
query in the error result.

----------------------------------------------------------------------

Comment By: Stefan Manegold (stmane)
Date: 2004-10-09 17:19

Message:
Logged In: YES 
user_id=572415

Just for info:

Emails are only sent, if the test results differ between two
subsuquent days,
i.e., if a test that used to be work (be green), now shows
differences (orange/red), or vice versa. If all tests show
the same behaviour as the day before, no emails are sent ---
even if all tests (still) "fail significantly"!

We might consider disabling the respective tests in the
release branch, but I'd higly recomment to keep them running
in the main trunk, to keep us reminded of the current
situation/status.
(But as I said, my role concerning testing has been
"reduced" to at most been an advisor ... ;-)


----------------------------------------------------------------------

Comment By: Fabian (mr-meltdown)
Date: 2004-10-09 17:04

Message:
Logged In: YES 
user_id=963970

I'm aware of it difficult to solve, but a test should give a
"green" mark.  If that is not possible, then I should say,
disable the whole test.  This way we keep getting massive
email traffic.

----------------------------------------------------------------------

Comment By: Stefan Manegold (stmane)
Date: 2004-10-09 17:00

Message:
Logged In: YES 
user_id=572415

AFAIK, the current behaviour of MapiClient is actually the
"original" one: in case of an error, the first line after
the last (successful) statement/query is returned as the
query that cause the error. In case there is a comment
between the last succesful statement/query and the failing
statement/query, the comment is not ignored but returned as
the query casuing the error.

If I recall correctly from what Sjoerd mentioned, this is
not easily to be fixed...

The stable output is not the "original" one, but rather the
"desired" one.
The tests used to show "major" differences, until I found
this too annoying, and made Mfilter threat them as "minor"
differences just before releasing 4.4.0/2.4.0.

I'd vote for keeping the tests and stable output as is, to
remind us of the problem --- but since I'm no longer in
charge of testing, other opinions/solutions are welcome, too ;-)



----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=482471&aid=1043573&group_id=56967

------------------------------------------------------------------------------
Join us December 9, 2009 for the Red Hat Virtual Experience,
a free event focused on virtualization and cloud computing. 
Attend in-depth sessions from your desk. Your couch. Anywhere.
http://p.sf.net/sfu/redhat-sfdev2dev
_______________________________________________
Monetdb-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/monetdb-bugs

Reply via email to