[unixODBC-dev] Lots of warnings, and an outright error, in 2.2.14 on 64-bit

Nick Gorham nick at lurcher.org
Mon Feb 16 16:14:50 GMT 2009


Tom Lane wrote:

>Peter Harvey <pharvey at peterharvey.org> writes:
>  
>
>>>Any interest in fixing these?  I gather that odbctest won't even be
>>>there anymore in 2.2.15, but...
>>>      
>>>
>
>  
>
>>I am willing to fix anything you find in unixODBC-Test.
>>    
>>
>
>Attached is a patch against 2.2.14 that fixes most of the more severe
>compiler warnings in a 64-bit build, as well as the error for
>SQLExtendedFetch.  I'm afraid you might have to do a bit of disassembly
>to pull out the pieces to where the code lives now, but it's a starting
>point anyway.
>
>Note there are several things that look to be outright bugs
>independently of platform word width, such as uninitialized variables
>and use of "==" where "=" was presumably meant.  I did not by any
>means fix all of the uninitialized-variable warnings, only a couple
>that were very obviously wrong.  I would urgently recommend that someone
>look at the rest of them.
>
>Also, there are quite a few reference-to-undeclared-function warnings,
>which are something else that's a severe portability hazard on 64-bit
>platforms.  I didn't try to clean those up but it'd be a good idea.
>
>Lastly, I gave up trying to fix Drivers/nn/ after realizing that the
>package does not provide the complete source code (.y and .l files
>are missing).  Isn't that a GPL violation?
>  
>
I had missed the .y from the Makefile.in, though I need to check if it 
builds with bison now.

The patch seems to have been stripped, can you send it to me directly 
and I will pass it on to Peter if he wants it.

Thanks.

-- 
Nick


More information about the unixODBC-dev mailing list