Re: Fix Windows socket error checking for MinGW

On Mon, Aug 19, 2013 at 10:18 AM, Michael Cronenworth <mike(at)cchtml(dot)com> wrote:
> On 08/19/2013 09:11 AM, Andrew Dunstan wrote:
>> I already gave you a sufficient identifier for the commit. In case you're not
>> aware, git is quite happy dealing with small commit identifiers. If you do "git
>> log -1 a099482" you should get the details you require.
>
> It should cross your mind that people have more than one computer and may reply
> using a mobile phone that doesn't have git loaded on it.
>
> When I'm back at my Postgres development system I'll gladly look at the git history.
>
> P.S. The tone of your replies is very hostile and unwelcome.

Uh... I don't think Andrew has said anything terribly hostile. He
claims that your patch reverts a change made by another recent patch,
and he's right. He doesn't mean that you intentionally did that by
basing your patch off the wrong git commit; he means that the changes
in your patch happen to be the opposite of part of what was in that
previous patch. That means that either the previous patch was wrong,
or yours is.

I'm not sure why it should be Andrew's job to care about which
machines you have git installed on. If you can't look up the SHA he
provided on the platform you're currently using, that's your problem,
not his. This is not exactly time-critical, so you could have just as
easily waited until you were back at your laptop before replying.

TBH, I think Andrew is going above and beyond the call of duty by
developing and testing a patch for the problem you reported.

And FWIW, I also agree that we should avoid putting in direct
references to WSA* constants in various portions of the code base.
We've worked pretty hard to paper over to the differences between
Windows and everything else, and the reward of that labor is that 95%
of the code we write doesn't have to care about which operating system
it's running on. That's a good thing, because many developers - such
as myself - do not have Windows development environments set up, and
even those that do will not welcome an increased need to test a
proposed patch on multiple platforms. I think everyone here accepts
that some level of additional effort is going to be needed, on an
ongoing basis, to support Windows. But keeping that to a minimum is
important.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company