On Thu, Mar 31, 2005 at 09:53:57PM +0200, Peter Eisentraut wrote: >That in turn forces the user to set up each ODBC drivers twice, once for > each driver manager. Not true; the driver managers are supposed to share config files. > Well, the above is mostly true because you can build the program one way > and the driver the other way and it might still work, but who really > knows? Hrm? I periodically test binary compatibility between UnixODBC and iODBC before uploading new versions of UnixODBC. In the past, I've also specifically continued to build some drivers against libiodbc2-dev instead of unixodbc-dev for testing purposes, but I think I dropped that due to lack of solid 64-bit support in iODBC at the time. > Should we somehow declare one or the other as the preferred driver > manager, thus making it easier for users and perhaps developers? Christian and I have discussed this in the past. He has repeatedly asked for someone to take over maintenance of iODBC; he's offered it to me, but my answer was that if no one else wanted to maintain it, I would kill it off instead. The main reason for not killing it off is indeed the wretched circular build-dep with Qt. I'm not willing to drop the Qt-based GUI tools from UnixODBC, and the Qt maintainers were not willing to fix Qt's inane build-dependency on ODBC. This would surely make Qt<->UnixODBC the largest, ugliest circular build-dep in Debian if not fixed somehow. > - Both myodbc (MySQL ODBC driver) and psqlodbc (PostgreSQL ODBC driver) > build against unixODBC. and freetds (MSSQL/Sybase ODBC driver). -- Steve Langasek postmodern programmer
Attachment:
signature.asc
Description: Digital signature