Re: BUG #13601: bit as quoted column in output - Mailing list pgsql-bugs

From Thomas Munro
Subject Re: BUG #13601: bit as quoted column in output
Date
Msg-id CAEepm=1KrsRDhJOmnwKPkOmEf6ufjHQ3XE5dvKHe6PiZeQ7R_w@mail.gmail.com
Whole thread Raw
In response to BUG #13601: bit as quoted column in output  (vicky.soni@quipment.nl)
List pgsql-bugs
On Wed, Sep 23, 2015 at 6:35 PM, Vicky Soni  - Quipment India
<vicky.soni@quipment.nl> wrote:
> Hi Thomas,
>
>                           In this case, we are not able to retrieve correct datatype of returning function when
callingfrom Application. In our case it is DOTNET and we use NPGSQL. 
>
> Is this a bug? Can we expect fix for this or should we live with same?

I don't know anything about npgsql or C# sorry, but I assume that it
has the same information available to it.  I did notice from a very
quick look at that project's source[1] (if I am looking at the right
thing, which is not a given) that they treat bit(1) specially: they
use a C# bool for bit(1), but they use a BitArray class for any other
size of bit array.  At a guess, that might include values in a column
that contains bit strings of size 1 that was described as having type
bit with typmod -1, which means you'd get BitArray of size one
sometimes and bool other times, depending on whether Postgres happens
to provide typmod info.  Could that explain what you are seeing?

Otherwise, I would try raising an issue with a minimal reproduction
and description of the expected and actual behaviour over at the
npgsql project and see what they say.

[1] https://github.com/npgsql/npgsql/blob/develop/src/Npgsql/TypeHandlers/BitStringHandler.cs

--
Thomas Munro
http://www.enterprisedb.com

pgsql-bugs by date:

Previous
From: Haribabu Kommi
Date:
Subject: Re: Memory leak with PL/Python trigger
Next
From: Andreas Hauser
Date:
Subject: Re: BUG #13633: ERROR: invalid memory alloc request size