>>>> Can we get the BIT type working now that 7.1 is branched?
I did some work on the BIT types a couple months ago. According to
my notes, the following issues are still outstanding before they
can be said to work at all:
Bit and hexstring literals are not handled in a reasonable fashion;
the scanner converts them to integer constants which is bogus.
Probably they need to be converted to some generic 'UNKNOWNBITSTRING'
pseudo-type that can later be coerced to a specific bitstring type.
I didn't touch this because it seems to open up the Pandora's box
of unknown-constant handling, for which we do not have a good
general solution.
SQL92 sez we need a position() function for bitstrings.
Need a regression test for bit types.
scalarltsel() and friends need to cope with bit/varbit types in
order to make good use of indexes on bitstrings.
pg_dump does not handle BIT/VARBIT lengths properly (pjw may have
fixed this by now).
regards, tom lane