Charles Wilson wrote:
> On 10/24/2011 8:53 AM, Earnie wrote:
>> Earnie Boyd wrote:
>>> We have a fallibility in that we have named our API interface w32api.
>>> In order to recover should we rename it to winapi or should we go with
>>> w64api? I would rather see winapi instead.
>> Ping. Opinions needed.
> I don't think there is any need to rename it at all. So what if the name
> says "w32api" for historical reasons -- just go ahead and add the new decls.
Ok, we'll just leave it. I already see #if _WIN64 used anyway.
> I think that renaming the api will cause lots of breakage which will
> take a lot of effort to fix and will cause pain for our users and
> ourselves -- because 'w32api' is actually embedded in the gcc specs for
> cygwin's compiler and the msys compiler (-I/usr/include/w32api/,
> -L/usr/lib/w32api/) as well as in the build process for both cygwin and
> msys. binutils might "know" about it as well. And never mind the churn
> in mingw-get's catalogues...every single mingw32-*-dev entry will need
> to be updated, and most of the msys-*-dev ones as well.
It's just a shame that this wasn't thought about by Cygnus back in the
day. Or maybe back then they had plans for w64api when it came about.
> If you must rename it, then winapi is the better choice, even tho RMS
> will complain ("avoid using 'win' as a shortened form of the OS from
> Redmond, because it implies a positive connotation for that non-free
Yes, I know the rant.