Proposed platform assumption changes

[Nico Williams]

On Wed, Feb 1, 2012 at 1:31 AM, Chris Hecker <checker at d6.com> wrote:
> I find it surprising that a lot of people chiming in are so willing to
> give up the currently working build support for one of the (admittedly
> sadly) most used compilers in the world in exchange for pretty minor
> syntactic sugar.  Seems crazy to me, but I guess I'm biased as one of
> the (apparently) few windows developers who builds krb5 instead of just
> links to prebuilt binaries.  However, even if I did just link to it,
> I've found (and reported) a bunch of minor bugs, and so fixing those
> locally with the proposed mingw way to get designated initializers would
> require a completely different toolchain.  Just doesn't seem worth it
> for this minor of a win.

It's not just syntactic sugar for me: it's sugar that cscope
[accidentally] (and other tools, no doubt) understands, which makes it
much easier to read a large code base.

But then, I'm not proposing that MSVC support be dropped.  I'm
proposing that since this isn't about syntactic sugar for humans but
rather about tools finding assignments to struct fields, ugly macro or
other workarounds for MSVC are worthwhile.

Nico
--