Re: Latest EMACS on BZR trunk does not compile with MinGW

[Stefan Monnier]

>> How 'bout testing a functionality on which we rely during the build,
>> rather than looking for some indirect evidence?
> Chicken-and-egg problem: that test is there so that we could source
> nt/mingw-cfg.site file, which does what you want.

AFAIK it's not exactly a chicken-and-egg problem: yes, we usually rely
on autoconf to do "functionality tests" and here we'd need to do
a functionality test in order to decide how to run autoconf, but we can
perform functionality tests without autoconf.

One more thing: what is this test *really* trying to figure out?
You say that the "uname" output is not very useful, but I don't
understand why

   case "$(uname -a)" in
     *MINGW* | *MinGW* ) ...
   esac

wouldn't give the right answer.  IIUC we're trying to figure out whether
we're doing an "MSYS" build (whatever that means) and to do that we
should:
- discover that we're on a w32 system.
- discover that we're not doing a cygwin build.


        Stefan