Andres Freund <> writes:
> On 2017-03-14 19:34:12 -0400, Tom Lane wrote:
>> It seems bizarre that you chose to spell the new configure symbol as

> I went back-and-forth about this a number of times.  We have a bunch of
> symbols defined with HAVE__ as a prefix (and some with HAVE_GCC__) - and
> more of the nearby code seems to use __ rather than _.  I don't really
> know why we started doing that, but it's far from new..

> Any idea why we introduce __ stuff?

The nearby stuff is describing features that have a specific name that
includes a leading underscore or two, like __builtin_unreachable().
That doesn't seem to apply here, so I wouldn't add extra underscores.

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to