Issue 122115 - Record mixup/repetition when loading chunks from SQL-DB
Summary: Record mixup/repetition when loading chunks from SQL-DB
Status: CONFIRMED
Alias: None
Product: Base
Classification: Application
Component: code (show other issues)
Version: 3.4.1
Hardware: All All
: P2 Normal with 2 votes (vote)
Target Milestone: ---
Assignee: AOO issues mailing list
QA Contact:
URL:
Keywords: regression
Depends on:
Blocks:
 
Reported: 2013-04-21 13:28 UTC by Jean-Louis Fuchs
Modified: 2013-05-04 09:33 UTC (History)
6 users (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments
Test file (8.89 KB, application/vnd.oasis.opendocument.text)
2013-04-21 13:28 UTC, Jean-Louis Fuchs
no flags Details
Test database (24.31 KB, application/vnd.oasis.opendocument.database)
2013-04-21 13:29 UTC, Jean-Louis Fuchs
no flags Details
Screenshot Mac (107.71 KB, image/png)
2013-04-21 13:29 UTC, Jean-Louis Fuchs
no flags Details
Screenshot Solaris (87.50 KB, image/png)
2013-04-21 13:30 UTC, Jean-Louis Fuchs
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description Jean-Louis Fuchs 2013-04-21 13:28:44 UTC
Created attachment 80566 [details]
Test file

Hello

When trying to print only a subset of records, the wizard can mixup the records. The result will also be wrong.

In the example the ID column is continuos, but as you see when selecting rows 1-63 the row 63 is repeated.

How to reproduce the problem:

* Open the file Etiq-Test.odt
* Do a mailing (File -> Print ; Output in: file)
* Choose records for the mailing with the mouse (ex. 1 to 10): Select ID
1, move to ID 10 press shift key and click the mouse.
* Move down with the mouse till ID 63, and move up, suddenlly multiple identical record are displayed (see screenshot)

-> Moving up isn't always necessary, but it helps triggering the behavior. It's impossible to select certain ranges.

Tested on MacOSX, Solaris and I think also Linux (not by me). 

Versions tested: 3.4.1 and a build of 3.5 from october I found on the web.

It would be great if the fix can be done in 3.4.1, so we can send a patch to our client.

Best,
  Jean-Louis
Comment 1 Jean-Louis Fuchs 2013-04-21 13:29:16 UTC
Created attachment 80567 [details]
Test database
Comment 2 Jean-Louis Fuchs 2013-04-21 13:29:35 UTC
Created attachment 80568 [details]
Screenshot Mac
Comment 3 Jean-Louis Fuchs 2013-04-21 13:30:04 UTC
Created attachment 80569 [details]
Screenshot Solaris
Comment 4 hanya 2013-04-21 14:30:45 UTC
This problem also happens on Table Data View to edit the database records that can be opened from Base.
Comment 5 Matthieu Riolo 2013-05-01 13:00:41 UTC
Dear Ladies and Gentlemen

I successfully tested the version 3.0.0, 3.2.0 and 3.3.0 on Linux. The bug does not appear in those versions. The first appearance of this bug was discovered in the version 3.4.0 and is still present in 3.4.1.

Is there a way to find out in which revision the bug has been included? 

Sincerly, Matthieu Riolo
Comment 6 Ariel Constenla-Haile 2013-05-01 13:07:33 UTC
(In reply to comment #5)
> Dear Ladies and Gentlemen
> 
> I successfully tested the version 3.0.0, 3.2.0 and 3.3.0 on Linux. The bug
> does not appear in those versions. The first appearance of this bug was
> discovered in the version 3.4.0 and is still present in 3.4.1.
> 
> Is there a way to find out in which revision the bug has been included? 

Please try with pre-Apache 3.4.0 Beta. It is more likely that this got broken before Oracle donated the code to the ASF; if so, you will have to search the history in the Mercurial repository.
Comment 7 Matthieu Riolo 2013-05-01 14:50:47 UTC
(In reply to comment #5)
> Dear Ladies and Gentlemen
> 
> I successfully tested the version 3.0.0, 3.2.0 and 3.3.0 on Linux. The bug
> does not appear in those versions. The first appearance of this bug was
> discovered in the version 3.4.0 and is still present in 3.4.1.
> 
> Is there a way to find out in which revision the bug has been included? 
> 
> Sincerly, Matthieu Riolo



After some further testing I found some additional information. It looks like that the display of the data from the database loads its content in small chunks. This can be seen if you look at the "total number of rows"-display. Once it loads a new chunk, you can scroll back up to see the problem in action. The problem can be reproduced in a fresh database as well.

That means that any database with more than about 100 records should have this bug...
Comment 8 Matthieu Riolo 2013-05-01 16:25:29 UTC
I finally did even more research and I found out that the problem does not happen with all connectiontypes. The problem does not occur if the database was built (New > Database > Connection with existing database) from a textfile or a spreadsheet. I also tried out the mysql connection and there I had the same problem as with the default databasetype.
Comment 9 dyo 2013-05-02 06:57:17 UTC
(In reply to comment #8)
> I finally did even more research and I found out that the problem does not
> happen with all connectiontypes. The problem does not occur if the database
> was built (New > Database > Connection with existing database) from a
> textfile or a spreadsheet. I also tried out the mysql connection and there I
> had the same problem as with the default databasetype.

The same problem with a connection to Postgresql database.
Comment 10 Raphael Bircher 2013-05-03 13:12:02 UTC
It's a regression between 3.3 and 3.4. LO does not have this bug but according Jean-Louis there is no issue at LibreOffice Bugzilla. There are two big chances:

1) The bugfix is in a CWS in the old repository from OOo who is not integrated because it stand as "Ready for QA"
2) The Regression happend because of the replacement of libraries.

Add keywords
Change Priority to P2
Set Bug to confirmed
Comment 11 Ariel Constenla-Haile 2013-05-03 13:21:53 UTC
(In reply to comment #10)
> It's a regression between 3.3 and 3.4

Did you check with Oracle 3.4.0 Beta?

> 2) The Regression happend because of the replacement of libraries.

if the bug is reproducible in the Beta, IP clearance has nothing to do; on the other hand, I doubt it touched the Base related code.
Comment 12 Raphael Bircher 2013-05-04 09:33:08 UTC
The OpenOffice.org 3.4 beta from Oracle shows the same behavure. so it has nothing to do with the IP clearance of Apache OpenOffice. So there is maybe a chance that the fix is in a CWS who is "Ready for QA"