I've committed the patch that allows unicode binds in 0.91. I've also added a CAST() to the "tablename" column when we query the information schema, to work around the nvarchar/ntext issue that for utterly mysterious reasons your setup does not get (I get it on both linux and OSX).
At this point I can barely run FreeTDS on OSX at all as my 0.82 build is hosed and it no longer builds - I can use the binaries though. 0.91 continues to be plagued with MemoryError and completely broken encoding on OSX, and I also get many core dumps on Linux - more tickets have been posted to Pyodbc with much more detail including ODBC and FreeTDS logfiles. I can get basic unicode and ORM tests running with 0.91 and Linux, though. I've wasted many days trying to get all of these libraries to work predictably and I'm out of time. Overall the status of ODBC on unix is extremely concerning to me as there doesn't seem to be anyone around to help with anything. On Sep 12, 2011, at 3:27 PM, Victor Olex wrote: > I confirm successful build and run time on both 64-bit and 32-bit > Linux (RHEL5) againt SQL Server 2008. > > It has turned out that Pyodbc needed the CPLUS_INCLUDE_PATH set to > where my version of UnixODBC was prefixed to prior to building. Best > bet is to remove all prior versions including distro packaged. If you > don't the Pyodbc may pickup header sources from there. By the way > their setup.py seems to have issues and can leave behind certain files > despite "clean" command. Also mind any leftovers in site-packages. > > Playing with compiler settings for the unixODBC, namely various > combinations of the CPPFLAGS="-DBUILD_LEGACY_64_BIT_MODE - > DSIZEOF_LONG_INT=8" allowed me to align the libs to the Pyodbc such > that the segfault got eliminated but in the end the flags proved > unneccesary once the above mentioned include path was properly set. > > On Sep 9, 10:54 pm, Victor Olex <[email protected]> wrote: >> Pyodbc has changes as late as this month. Misery likes company. When I >> tried to install this same stack on 64 bit architecture I am getting >> Segmentation fault errors for some queries. Easysoft has some info on >> ODBC in 64bit athttp://www.easysoft.com/developer/interfaces/odbc/64-bit.html >> There are some variables that can be set wrt to word length when >> building unixODBC. Hopefully I can find the right setup. >> >> On Sep 9, 1:47 pm, Michael Bayer <[email protected]> wrote: >> >>> 2.1.9 from the github link you showed me. Just to make sure I also just >>> cloned the tip from code.google.com and tried that one too. >> >>> Has pyodbc made adjustments to changes in FreeTDS 0.91 ? >> >>> On Sep 9, 2011, at 1:15 PM, Victor Olex wrote: >> >>>> What version of pyodbc? >> >>>> On Sep 9, 11:08 am, Michael Bayer <[email protected]> wrote: >>>>> On Sep 9, 2011, at 10:48 AM, Michael Bayer wrote: >> >>>>>> It also makes me less than comfortable unconditionally emitting a u'' >>>>>> for a bound parameter as it appears to cause problems. >> >>>>> I've also checked my SQL server, all databases including master are >>>>> configured at SQL Server 2008 compatibility... >> >>>> -- >>>> You received this message because you are subscribed to the Google Groups >>>> "sqlalchemy" group. >>>> To post to this group, send email to [email protected]. >>>> To unsubscribe from this group, send email to >>>> [email protected]. >>>> For more options, visit this group >>>> athttp://groups.google.com/group/sqlalchemy?hl=en. > > -- > You received this message because you are subscribed to the Google Groups > "sqlalchemy" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/sqlalchemy?hl=en. > -- You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en.
