Re: Unicode support
"ZIGLIO, Frediano, VF-IT" <[email protected]>
| Newsgroups | gmane.comp.db.unixodbc.devel |
|---|---|
| Message-ID | <[email protected]> |
Well... it's easier to say... if you compile DM with a given SQLWCHAR type every program that use the DM and every driver that DM load have to be compiled with same options. That is if you use for instance a PHP script that use php-odbc to connect to a MySQL db: - PHP - unixODBC - MySQL have to use same SQLWCHAR type! Regards Frediano Ziglio -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of SHANKAR Uma Sent: martedì 18 agosto 2009 21.28 To: Development issues and topics for unixODBC Subject: Re: [unixODBC-dev] Unicode support Is there some clear documentation what this build time option is? I would like to read up before deciding whether to make code changes by using iconv() or use this build option. I am also a bit unclear on the Unicode driver part. We are towards the end of our development cycle, and this is holding things up. Given that we have had several exchanges since y;day, here are what I seem to have as options: 1) Use iconv() for linux to convert to UCS-2 2) Use SQLWCHAR 3) Use the build time option. Given the above 3 options, it seems to me that the build time option would be the easiest given where we are now. If it is, then could you explain the Unicode driver comment that you made? Like I said we are building the unixODBC drivers ourselves, so we do have control over how we build it. -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Nick Gorham Sent: Tuesday, August 18, 2009 3:11 PM To: Development issues and topics for unixODBC Subject: Re: [unixODBC-dev] Unicode support SHANKAR Uma wrote: > Another question Nick. I was up on the unixodbc website, and for the 2.0.9 release , it says the following was fixed: > - Add build time option to select wchar_t UNICODE (4 bytes) as opposed > to signed short UNICODE (2 byte) > > I believe we build unixODBC ourselves. Does this mean that I could actually build unixODBC on unix boxes for the different sizes of wchar_t on unix v/s windows? And this would require no code changes on my part? And I could use wchar_t's the way it is being used now? > > Please clarify. > Yes, but if you wanted to connect to unicode ODBC driver they would also need to use 4 byte UNICODE. Can be done, but beware the driver. -- Nick _______________________________________________ unixODBC-dev mailing list [email protected] http://mailman.unixodbc.org/mailman/listinfo/unixodbc-dev _______________________________________________ unixODBC-dev mailing list [email protected] http://mailman.unixodbc.org/mailman/listinfo/unixodbc-dev _______________________________________________ unixODBC-dev mailing list [email protected] http://mailman.unixodbc.org/mailman/listinfo/unixodbc-dev