<p dir="ltr">On 13 Oct. 2016 05:28, "Tom Lane" <<a href="mailto:tgl@sss.pgh.pa.us">tgl@sss.pgh.pa.us</a>>
wrote:<br/> ><br /> > I wrote:<br /> > > Albe Laurenz <<a
href="mailto:laurenz.albe@wien.gv.at">laurenz.albe@wien.gv.at</a>>writes:<br /> > >> Tom Lane wrote:<br />
>>>> I'm okay with adding PGDLLEXPORT to the extern, but we should update<br /> > >>> that
commentto note that it's not necessary with any of our standard<br /> > >>> Windows build processes. (For
thatmatter, the comment fails to explain<br /> > >>> why this macro is providing an extern for the base
functionat all...)<br /> ><br /> > >> Here is a patch for that, including an attempt to improve the
comment.<br/> ><br /> > > Pushed with some further twiddling of the comment.<br /> ><br /> > Well, the
buildfarmdoesn't like that even a little bit. It seems that<br /> > the MSVC compiler does not like seeing both
"externDatum foo(...)" and<br /> > "extern PGDLLEXPORT Datum foo(...)", so anything that had an extern in<br /> >
a.h file is failing. There is also quite a bit of phase-of-the-moon<br /> > behavior in here, because in some cases
somefunctions are raising errors<br /> > and others that look entirely the same are not.<p dir="ltr">Pretty sure we
discussedand did exactly this before around 9.4. Will check archives...<p dir="ltr">Yeah. Here's the thread.<p
dir="ltr"><a
href="https://www.postgresql.org/message-id/flat/27019.1397571477%40sss.pgh.pa.us#27019.1397571477@sss.pgh.pa.us">https://www.postgresql.org/message-id/flat/27019.1397571477%40sss.pgh.pa.us#27019.1397571477@sss.pgh.pa.us</a><p
dir="ltr">Ithink the discussion last time came down to you and I disagreeing about Microsoft droppings too ;)
