[freetds] odbc isql

James K. Lowden jklowden at schemamania.org
Fri Apr 8 02:17:47 EDT 2005


I'm hacking the unixodbc isql and deciding if we want it included in our

My version is considerable simpler than the original: no html, no
delimited output, no fancy help.  More like tsql for odbc.  My objective
is to have a very simple (but complete) isql for FreeTDS that we can use
to test ODBC functionality.  I'm considering writing scripts against it
for regression testing.  

The motivation will be clear to anyone who's read the list this week.  The
just-released 0.63 odbc driver doesn't handle row-error-row situations
correctly: the second resultset is concatenated to the first one, and the
first resultset's final row is duplicated.  That's a data corruption
issue, and the mere fact that it crept into a release says the testing was
(by definition) not good enough.  I'm looking to make better testing
easier, on the theory that what's easy will happen more.  

Peter's code is GPL.  We can include it in source distributions (which is
all we offer).  I don't see a problem with binary distributions, either:
there's no rule against linking GPL and LGPL code, and in any case all the
sources are always available in ibiblio and sourceforge.  

I did plop Brian's readline hack into it from tsql.  Brian, if you don't
like that, I'll pull it.  

Any thoughts one way or the other?  


More information about the FreeTDS mailing list