On 12/15/2010 10:17 AM, Tom Lane wrote:
> Alvaro Herrera<alvherre@commandprompt.com> writes:
>> Excerpts from Andrew Dunstan's message of mié dic 15 02:08:24 -0300 2010:
>>> That didn't work. But git bisect says it's this commit that's to blame:
>>> <https://github.com/postgres/postgres/commit/e710b65c1c56ca7b91f662c63d37ff2e72862a94>
>> Hmm I wonder if this is reproducible in a non-Windows EXEC_BACKEND
>> scenario.
> I'm pretty sure I tried the no-flat-files code in that scenario while
> writing it. But it might be worth trying that again. You'd think
> though that if EXEC_BACKEND were sufficient to provoke it, all Windows
> builds would fail. I'm still mystified by what is the difference
> between Andrew's non-working installation and working mingw builds.
>
>
This is a new installation of Mingw. The buildfarm animals were set up
years ago, with substantially older versions of Mingw. SO ISTM that
either we have tickled a new bug of theirs or their new setup has
tickled a bug of ours.
cheers
andrew