>>>>> "M" == M -A Lemburg <mal@stripped> writes:
M> I've downloaded the latest MyODBC driver and iODBC manager and
M> tested them for memory leakage. Unfortunately, the leakage
M> I reported months ago is still there. Even worse, it seems to
M> have increased.
M> The strange thing is that the leakage only occurs on Linux,
M> not with the latest driver for Win95, yet both platforms run the
M> same test code.
M> I also checked the test code with other drivers directly and
M> through the iODBC manager: neither the drivers, nor the iODBC
M> manager leak memory and this makes me pretty convinced that the
M> leakage must be located somewhere in the Unix part of the
M> MyODBC driver.
M> I looked around the code a bit, but all I found was that
M> a malloc in prepare.c is done via my_malloc(), but freed via free(),
M> instead of my_free(). Changing it didn't result in any improvement
This was a bug (but it should only bit you if you compiled myodbc with
the -DSAFEMALLOC flag set)
M> Can anybody comment on this ?
If you can you make a full example that shows some memory leaks I
could probably find out what's wrong!