PING: WIN-28: fix POSIX file descriptor numbering assumptions

[Adam Megacz]

Hey, I never got a response on this. May I commit?

  - a


Bryce> I recall seeing a patch from Jeff a long time ago to fix
Bryce> FileDescriptor to be more portable and not make the POSIX
Bryce> file-descriptor-is-an-int assumption, but it never got checked
Bryce> in.

Tom> As I recall it used a RawData to represent the file descriptor.
Tom> I think we should resurrect that.

I have his changes -- they're part of the half-working mingw tree I
inherited. Unfortunately those files are way out of date;
diff-and-patch isn't going to help much.

I looked into how much stuff this change would touch, and it turns out
that it impacts a lot of stuff:

  java/io/natFileDescriptorEcos.cc
  java/io/natFileDescriptorPosix.cc
  java/io/natFileDescriptorWin32.cc
  java/io/FileDescriptor.java
  java/net/PlainDatagramSocketImpl.java
  java/net/natPlainDatagramSocketImpl.cc
  java/net/PlainSocketImpl.java
  java/net/natPlainSocketImpl.cc

I'm worried that if I go through making these changes, a) I'll break
something, and b) it's going to take a really really long time to get
this all tested and approved (sorry, I'm trying really hard to get XWT
out the door *and* finish mingw support for 3.1, and there just aren't
enough hours in the day...).

Since there is no such thing as a 'mingw64', and probably won't be for
a while, is it okay if I commit WIN-28 with a FIXME noting that we
need to switch to RawData before porting to Win64?

  - a

-- 
"If I put copyrighted material into the 'chewy nougat center', I can
 claim the crunchy chocolate coating is an 'Access Control
 Mechanism'."                                     --lynx_user_abroad

Why Windows NT/2k/XP is unreliable: www.zappadoodle.com