On Sat, May 29, 2021 at 10:44:14AM -0400, Tom Lane wrote:
> What I had in mind was to *only* add pgwin32_setenv, not setenv.c.
> There's no evidence that any other modern platform lacks setenv.
> Moreover, there's no issue in these branches unless your platform
> lacks setenv yet has GSS support.
I have been finally able to poke at that, resulting in the attached.
You are right that adding only the fallback implementation for
setenv() seems to be enough. I cannot get my environment to complain,
and the code compiles.
--
Michael
