This is the mail archive of the
java-patches@gcc.gnu.org
mailing list for the Java project.
Re: FYI: Patch: Java NIO
- From: Mohan Embar <gnustuff at thisiscool dot com>
- To: Michael Koch <konqueror at gmx dot de>, Anthony Green <green at redhat dot com>
- Cc: java-patches at gcc dot gnu dot org
- Date: Thu, 16 Oct 2003 06:56:53 -0500
- Subject: Re: FYI: Patch: Java NIO
- Reply-to: gnustuff at thisiscool dot com
Anthony,
>> I commited a patch to provide more NIO functionality. Hopefully all slowly
>> begins to really work.
>
>This patch causes the following build regression for Linux-cross-mingw32 builds...
I emailed Michael privately about this. There are several possible
solutions:
- split natPipeImpl.cc into natPipeImplWin32.cc and natPipeImplPosix.cc
à la natFileWin32/Posix.cc, etc. I think we'd want to factor NoNet
into this too, right Michael? I asked Michael if he wanted to submit a
patch for this or wanted me to do it, but haven't heard back from him
yet. I've also been too busy to whip up a patch for this. I think this
is the best solution.
- a quicker and dirtier solution would be to define _Jv_Pipe in both
win32.h and posix.h and provide implementations for these in
both win32.cc and posix.cc. I was so relieved to eradicate all of
these, though, so doing this would be a step backward.
If you want a hack to unbreak the build for MinGW only:
- Add the following includes:
#include <io.h>
#include <fcntl.h>
- Use _pipe instead of pipe
- Add the parameters 4096 and _O_BINARY to the _pipe call.
-- Mohan
http://www.thisiscool.com/
http://www.animalsong.org/