On Wed, Dec 15, 2010 at 12:55, Robert Haas <robertmhaas@gmail.com> wrote:
>>> It seems like pg_read_binary_file() is good to have regardless of
>>> whatever else we decide to do here. Should we pull that part out and
>>> commit it separately?
>>
> The whole-file versions seem like a good idea - my only hesitation is,
> I'm not sure why we didn't include that functionality originally. It
> seems obviously useful, so does that mean that it was omitted on
> purpose for some reason?
I applied the attached patch extracted from Dimitri's work.
One difference is 'offset' argument is removed from 'whole' mode.
So, we'll have (path, offset, length) and (path) versions.
Checking with convert_and_check_filename is left as-is.
--
Itagaki Takahiro