On Fri, Apr 22, 2011 at 9:16 AM, Geoffrey Myers <geof@xxxxxxxxxxxxxxxxxxxxx> wrote:
Vick Khera wrote:Totally agree. Still, the question remains, why not leave it as SQL_ASCII?The database's enforcement of the encoding should be the last layer that does so. Your applications should be enforcing strict utf-8 encoding from start to finish. Once this is done, and the old data already in the DB is properly encoded as utf-8, then there should be no problems switching on the utf-8 encoding in postgres to get that final layer of verification.
Well, if your data is supposed to be UTF-8 encoded, then any of those characters with invalid encoding in UTF-8 could reasonably be viewed as data errors. Leaving the database in SQL-ASCII allows those errors to continue accumulating, which will make a switch in the future even harder. If the lack of being able to check encoding errors at the database level doesn't bother you, and you're fine with risking bigger pain later in order to avoid pain now, then I see no compelling reason to move away from SQL_ASCII.
-Eric
