Re: [Firebird-devel] CNCT_user and CNCT_host

2016-09-19 Thread Jiří Činčura 
> things. But for os user name - no, not ok.

Can you elaborate?

--
Mgr. Jiří Činčura
Independent IT Specialist

--
Firebird-Devel mailing list, web interface at 
https://lists.sourceforge.net/lists/listinfo/firebird-devel

Re: [Firebird-devel] CNCT_user and CNCT_host

2016-09-19 Thread Alex Peshkoff 
On 19.09.2016 20:14, Jiří Činčura wrote:
>> Historical reasons.
>> This values were passed in CNCT since pre-firebird times therefore
>> current solution was OK for old clients.
> But there's inherently nothing wrong with using DPB, right? Because i.e.
> isc_dpb_process_id and isc_dpb_client_version are only in DPB to name a
> few (I don't know any matching CNCT, to be precise), so I would rather
> keep it around one place in code.
>

DPB is OK for process id, version of the client and a lot of other 
things. But for os user name - no, not ok.


--
Firebird-Devel mailing list, web interface at 
https://lists.sourceforge.net/lists/listinfo/firebird-devel

Re: [Firebird-devel] CNCT_user and CNCT_host

2016-09-19 Thread Jiří Činčura 
> Historical reasons.
> This values were passed in CNCT since pre-firebird times therefore 
> current solution was OK for old clients.

But there's inherently nothing wrong with using DPB, right? Because i.e.
isc_dpb_process_id and isc_dpb_client_version are only in DPB to name a
few (I don't know any matching CNCT, to be precise), so I would rather
keep it around one place in code.

-- 
Mgr. Jiří Činčura
Independent IT Specialist

--
Firebird-Devel mailing list, web interface at 
https://lists.sourceforge.net/lists/listinfo/firebird-devel

[Firebird-devel] [FB-Tracker] Created: (CORE-5351) left join with UDF where clause return different results in 2.5.5 and 3.0

2016-09-19 Thread JIRA 
left join with UDF where clause return different results in 2.5.5 and 3.0
-

 Key: CORE-5351
 URL: http://tracker.firebirdsql.org/browse/CORE-5351
 Project: Firebird Core
  Issue Type: Bug
Affects Versions: 3.0.0
Reporter: Erik Jõeveer


Hello,

FB3 returns wrong results on left join selects with specific UDF condition (no 
problems with FB1.5 and FB2.5). 
It seems to be similar with this issue: 'left join with COALESCE where clause 
return different results in 2.5.5 and 3.0 RC2' 
(http://tracker.firebirdsql.org/browse/CORE-5150).

Here is example with Firebird standard ib_udf library function STRLEN , but it 
can be reproduced with other UDF functions too (we having same problem with our 
proprietary UDF-s).


CREATE TABLE TEST_TABLE1 (
ID INTEGER NOT NULL,
TESTTABLE2_ID INTEGER
);
ALTER TABLE TEST_TABLE1 ADD PRIMARY KEY (ID);
COMMIT;

CREATE TABLE TEST_TABLE2 (
ID INTEGER NOT NULL,
GROUPCODE VARCHAR(10)
);
ALTER TABLE TEST_TABLE2 ADD PRIMARY KEY (ID);
COMMIT;


INSERT INTO TEST_TABLE1 (ID,TESTTABLE2_ID) VALUES (1,100);
INSERT INTO TEST_TABLE2 (ID,GROUPCODE) VALUES (100,'A');
COMMIT;

DECLARE EXTERNAL FUNCTION strlen
CSTRING(32767)
RETURNS INTEGER BY VALUE
ENTRY_POINT 'IB_UDF_strlen' MODULE_NAME 'ib_udf';
COMMIT;

-  NEXT QUERY SHOULD RETURN EMPTY RESULT SET BUT IT RETURNS 1 RECORD:

SELECT T1.ID
FROM TEST_TABLE1 T1
LEFT JOIN TEST_TABLE2 T2 ON (T2.ID=T1.TESTTABLE2_ID)
WHERE (STRLEN(T2.GROUPCODE)=0)


  ID

   1



Intrestingly same query with conditon 'STRLEN(T2.GROUPCODE)=1' or with 
condition 'STRLEN(T2.GROUPCODE)=2' works as expected (first returns 1 record 
and second returns no records).


-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: 
http://tracker.firebirdsql.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

   

--
Firebird-Devel mailing list, web interface at 
https://lists.sourceforge.net/lists/listinfo/firebird-devel

Re: [Firebird-devel] Foreign key error even there should not be one (maybe, Weird problem in one database)

2016-09-19 Thread Tommi Prami 
Weird part was that I actually pumped the data to the empty database and it
behaved the same.

This is more than less puzzling.

And the Backup and restore did not help,. which also, I think, should
recreate the indexes and so on...

-Tee-



On Thu, Sep 15, 2016 at 10:58 AM, Thomas Steinmaurer 
wrote:

> > Whole story here :
> >
> > https://plus.google.com/105626236962310493174/posts/3UZB71NjUhA
> >
> > Copy of message
> > - - - - - - - -  - - - - - - -  -
> > Is there a tool (or way) to debug database step by step.
> >
> > Now we have weird situation where delete from one database fails, on
> > foreign key error, which sounds like it would be trivial thing, but seems
> > so far it is not.
> >
> > I was just thinking that if it would be our application code I would be
> > able to debug through all of the related code and see if there is the
> > culprit or not.
> >
> > We suspect it has somethign to do with specific data on the rows. because
> > some of them delete perfectly fine. This happens in this one database
> only.
> > I've done backup restore for it, I ran gfix stuff on it. I even pumped
> the
> > data in empty DB.
> >
> > it should be problem of database structure (trickers and such), data or
> the
> > Firebird. and most likely it problem of first two.
> >
> > I was just thinking that is there an tool I could see step by step all
> the
> > operations in the DB before the foreigh key error is raised. So I could
> > step by step see the order the trickers and so fires and what they do.
> >
> > Or if someone at Firebird project could check if it is indeed the
> Firebird
> > bug. I think we could send the DB, if it he/she keeps it safe and keeps
> it
> > secret :) 
>
> If you re-create the FK constraint - which re-creates an index behind the
> scene - solves the problem, then some sort of data/index corruption might
> be an explanation.
>
>
> Thomas
>
> 
> --
> Firebird-Devel mailing list, web interface at
> https://lists.sourceforge.net/lists/listinfo/firebird-devel
>
--
Firebird-Devel mailing list, web interface at 
https://lists.sourceforge.net/lists/listinfo/firebird-devel