Hi all; Overall the experience moving LedgerSMB to UTF-8 was not as bad as we expected. There were a number of cascading failures. Keep in mind we had already made a number of significant changes to the software in important areas first, but I don't expect that these would be particularly problematic. The key aspects have been:
1) Database encodings (this is a big one since Debian initdb's in C locale which is not UTF-8) 2) File encodings (chart of accounts, translations). These can be converted with appropriate 3) Output encodings. binmode STDOUT ':utf8'; needs to be set in all appropriate spots but where non-text is being sent out, one needs to reset this to an appropriate binmode. There are a couple challenges I think you will face: 1) I doubt Dieter will be willing to accept patches to his version to do this and so at some point you are kind of on your own. the binmode patches will be the most troublesome in this area. 2) It is likely that when you begin this process you will end up with some issues in output encodings being set to the wrong thing. This is really where about 70-80% of the time in LedgerSMB doing this has been spent. Best Wishes, Chris Travers _______________________________________________ SQL-Ledger mailing list [email protected] http://lists.ledger123.com/mailman/listinfo/sql-ledger
