Posted 31 January 2011 - 02:52 PM
Posted 31 January 2011 - 03:51 PM
The first two are from our alphabet, the rest are in order of other alphabets throughout the UniCode character map.
Is it any wonder that a program has a problem sorting things the way we expect.
When it was just the 26 English characters, upper and lowercase it wasn't too bad.
I am not sure that the problem with not finding a name or finding an unrelated one is related to umlauts, but is sure seems to be.
There was something like that about a year ago with enhanced characters and I seem to remember that exporting to a GEDCOM and importing it into a new database (OR dragging and dropping everyone in the database) eliminated the problem in the new database.
I don't know why though.
Posted 31 January 2011 - 04:23 PM
Posted 31 January 2011 - 04:41 PM
I recall this symptom (as Alfred mentions) going way back to this thread:
Or I will type in a name and it goes to some name that doesn't come close to what I typed.
--- "GENEALOGY, n. An account of one's descent from an ancestor who did not particularly care to trace his own." - Ambrose Bierce
--- "The trouble ain't what people don't know, it's what they know that ain't so." - Josh Billings
K e V i N
Posted 01 February 2011 - 11:01 AM
Thank you for this insight. I passed this along to Bruce and Mike in case they were not aware of this discussion. Confirming this issue is in our tracking system.
I daresay that this is a problem that may be out of the hands of the RM developers as it may be embedded in DISQLite3 which they use to integrate the SQLite database engine with the application. What's worse is that the DSQLite3 developers are dependent on third party developers of the libxslt library to resolve a memory leak in the library before they can compile what should be a fix in DISQLite. See this about sorting German text using DIXml; the problem is likely common to DISQLite3.