Re: VS 2015 support in src/tools/msvc

On Wed, Mar 23, 2016 at 8:45 PM, Petr Jelinek <petr(at)2ndquadrant(dot)com> wrote:
> On 23/03/16 08:17, Michael Paquier wrote:
>>
>> On Mon, Mar 21, 2016 at 3:43 AM, Andrew Dunstan <andrew(at)dunslane(dot)net>
>> wrote:
>>>
>>>
>>>
>>> On 03/20/2016 12:02 AM, Michael Paquier wrote:
>>>>
>>>>
>>>> On Sun, Mar 20, 2016 at 8:06 AM, Andrew Dunstan <andrew(at)dunslane(dot)net>
>>>> wrote:
>>>>>
>>>>>
>>>>>
>>>>> Still to do: the non-perl pieces.
>>>>
>>>>
>>>> The patch to address locales is the sensitive part. The patch from
>>>> Petr is taking the correct approach though I think that we had better
>>>> simplify it a bit and if possible we had better not rely on the else
>>>> block it introduces.
>>>
>>>
>>>
>>>
>>> OK, the please send an updated set of patches for what remains.
>>
>>
>> Here you go:
>> - 0001 fixes the global declarations of TIMEZONE_GLOBAL and
>> TZNAME_GLOBAL to be WIN32-compliant. I got bitten by that in the ECPG
>> compilation.
>> - 0002, support of VS2015 in the scripts of src/tools/msvc
>> - 0003, this is necessary in order to run the regression tests,
>> details are here:
>>
>> http://www.postgresql.org/message-id/CAB7nPqTDiXxS8CdL8mOiVh6qFQ-1qV9mKN0AyjzYBvzv6WC0dA@mail.gmail.com
>
> So that's basically what Andres did? Interesting that we now actually really
> need it. I was in favor of doing those changes in any case.

Yes, that's what Andres did. I am just attaching it here to allow
Andrew to test this patch set appropriately.

>> - 0004 is the fix for locales. This is actually Petr's work upthread,
>> I have updated comments in the code though and did a bit more
>> polishing. This still looks like the cleanest solution we have. Other
>> solutions are mentioned upthread: redeclaration of struct _locale_t in
>> backend code is one.
>
> Thanks for polishing this.
>
> I think this is ready to be committed, but the 0003 might be somewhat
> controversial based on the original thread.

I thought that the end consensus regarding 0003 was to apply it, but
we could as well wait for the final verdict (committer push) on the
other thread. And well, if this is not applied, some of the gin tests
will complain about "right sibling of GIN page is of different type" a
couple of times.
--
Michael