Re: BIT datatype (Fixed)

Bruce Momjian wrote:

> > > here is an updated version of the bit type with a bugfix and all the necessary
> > > SQL functions defined. This should replace what is currently in contrib. I'd
> > > appreciate any comments on what is there.
> >
> > I'm hoping to get a chance to look at it fairly soon. It's already
> > been committed to the tree...
>
> The bits never even got cold before it was applied. :-)

Thanks! I spent a day with a debugger to find a silly error, but it all does seem to
behave now. I could not use the name BIT for the example type, as it is reserved. I also
noticed that I could not use BITS(6) for a user-defined type to specify a length. A last
problem is that there are two types of output routines, one returning a bit string in hex
format, and another returning it in binary format. I seem to recall that there once was a
discussion on adding a C-string type for situations like this, so that one output-type
would be the default, but the second could be made available through a function. I also
didn't do anything to make these indexable.

Any suggestions or complaints, please let me know.

Adriaan