Permalink
Commits on Feb 16, 2005
  1. updated insert test to try multiple consecutive invocations of

    Query::insert, plus a re-bind (without re-prepare) and another insert
    test.
    committed Feb 16, 2005
  2. change statement handle reset logic to clear all local variables first

    before actually touching the oracle handle (in case the oracle operation
    fails, at least local state is correct); and changed the logic around
    Query::_prepared to accommodate multiple calls to Query::execute() (now
    we reset the prepared state only when the SQL is actually changed, since
    logically that's the only time we would actually need to re-prepare the
    statement handle).  also added some operator= overloads.
    committed Feb 16, 2005
  3. beginning of changes to support conversions to oracle-specific data t…

    …ypes,
    
    first by using a void* and casting to whatever the requested type is
    committed Feb 16, 2005
Commits on Mar 16, 2004
  1. version bump

    committed Mar 16, 2004
  2. after seeing one too many core files in the ORAPP::~Field dtor, I did

    a little research into Oracle 8i, 9i, and 10g to learn what valid
    inputs were into OCIHandleFree().  Apparently not all OCI_HTYPE's are
    OCIHandleFree()able, and so I've disabled the two cases of freeing a
    handle type that is not documented as a valid input into OCIHandleFree().
    committed Mar 16, 2004
Commits on Mar 13, 2004
Commits on Jan 21, 2004
  1. added a few more minor "test"s

    committed Jan 21, 2004
  2. added stdio.h to the include line for pre-gcc3.2 compiler toolchains (in

    order to get at vsnprintf()) and changed all references of
    std::string::clear to instead assign an empty string, since pre-gcc3.x
    compiler toolchains didn't have an API-complete implementation of the
    STL string object (clear() was not present).
    committed Jan 21, 2004
  3. commented out -O2 optimization because on gcc 2.96 it was causing a

    segfault (doesn't segfault without).
    committed Jan 21, 2004
Commits on Dec 18, 2003
  1. changed constructor to take name length to accommodate what we get back

    from OCIAttrGet(OCI_ATTR_NAME)
    committed Dec 18, 2003
  2. another bug from Fedor <fedor@oldconomy.com>, this one relating to th…

    …e fact that different versions of OCI null-terminate the field name when you OCIAttrGet(OCI_ATTR_NAME). Here we change field to accept the length of the name when instantiating the Field object.
    committed Dec 18, 2003
Commits on Dec 16, 2003
  1. typo

    committed Dec 16, 2003
  2. fixed bug whereby we can't assume that the vector<> returns NULL when…

    … you
    
    ask for an unpopulated/unallocated offset.  the symptom was that you would
    ask for r->name(some_OOB_value) and it would dereference a wild pointer,
    sometimes yielding garbage, sometimes crashing.
    committed Dec 16, 2003
Commits on Dec 14, 2003
  1. extraneous comments removed

    committed Dec 14, 2003
Commits on Dec 13, 2003
  1. and, voila, the license

    committed Dec 13, 2003
  2. typo

    committed Dec 13, 2003
  3. Initial (real) import.

    committed Dec 13, 2003
  4. oops.

    committed Dec 13, 2003
  5. Initial revision

    committed Dec 13, 2003