Soo... how strict are we about changes between RC and final?
We've got a couple of small outstanding patches that take
freeciv-modpack, a headline new feature in 2.3.x, from being completely
unusable on Windows (crashes on launch) to being basically functional as
far as it's been tested (bug #18354, patch #2812).
The code patches are only in modinst.c itself. So they can't affect the
client or server; it seems the worst that could happen is that
freeciv-modpack is broken on some platform on which it was previously
working. (In my test on Ubuntu Linux, it still works as well as it ever
did, but that's hardly exhaustive.)
However, freeciv-modpack is arguably of most use to Windows users, so
that could be a risk worth taking. (Looking at the patches, the risk
seems small anyway.)
Conceivably, freeciv-modpack could start failing to _build_ on some
obscure platform, I suppose, but that seems very unlikely.
Additionally, the latter patch touches configure.ac; but only if MINGW32
is defined. So the risk seems confined to the Windows build process,
which (approximately) only needs to work for cproc (most Windows users
use executables), and clearly does (since it's his patch).
Freeciv-dev mailing list