[Libreoffice-bugs] [Bug 50575] ODBC sqlite3 timestamps as primary column: show empty data / crashes / ...

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=50575

--- Comment #10 from Terrence Enger <lo_bugs at iseries-guru.com> 2012-07-11 16:47:47 PDT ---
Created attachment 64119
  --> https://bugs.freedesktop.org/attachment.cgi?id=64119
screenshot, different result but still wrong

( Yes, I know that this bug is RESOLVED NOTOURBUG, but I had already started to
prepare this comment. )


In the environment described in the description of fdo#51987 (short version:
Linux, master pulled yesterday), my results are changed but still wrong.

First, the results from isql, which I take to be right:

    $ isql bug_047520
    +---------------------------------------+
    | Connected!                            |
    |                                       |
    | sql-statement                         |
    | help [tablename]                      |
    | quit                                  |
    |                                       |
    +---------------------------------------+
    SQL> select * from byTs
    +-----------+---------------------------------+-----------+
    | nr        | ts                              | word      |
    +-----------+---------------------------------+-----------+
    | 1         | 2012-04-06 12:34:56.654321      | Friday    |
    | 2         | 2012-04-05 13:45:57.123456      | Thursday  |
    | 3         | 2012-04-04 15:20:22.980         | Wednesday |
    | 4         | 2012-04-04 15:20:22.98          | Wednesday |
    +-----------+---------------------------------+-----------+
    SQLRowCount returns 4
    4 rows fetched

But Base shows rows with nr 1, 2, and 2, and the bottom of the Table
Data View indicates that the table has three rows.  Attaching
screenshot of the Table Data View.


@LEM ...

I have built the driver, sqliteodbc-0.87, with debug symbols to make
it easier to look into this bug.  But the "dynamic typing" used by
sqlite makes me doubt the possibilty of a robust solution.  And now I
see that fdo#38811 "default to SQLite not HSQLDB in Base" has been
declined; I wish I had been following that bug report.  These two
"downers" just about kill my interest, unless you want to tell me
otherwise.

-- 
Configure bugmail: https://bugs.freedesktop.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.