On Sat, Apr 26, 2008 at 7:04 AM, Matthew Toseland
wrote:
> I'm reverting this before releasing 1144. You can put it back afterwards, but
> it needs to be rewritten a bit: FileChannel.read() is not guaranteed to
> complete the operation (unlike RandomAccessFile.readFully()). Also there are
I'm reverting this before releasing 1144. You can put it back afterwards, but
it needs to be rewritten a bit: FileChannel.read() is not guaranteed to
complete the operation (unlike RandomAccessFile.readFully()). Also there are
a few RuntimeException's that you might want to catch explicitly.
According to a comment filed by sdiz on bug #2301:
"FileChannel.read()/write() on windows do serialized read/write.
Better nio for windows is on java 7 todo list. But I don't think they can
finish it on schedule."
So do we need to have multiple cache files to workaround this win-specific
bug?
According to a comment filed by sdiz on bug #2301:
FileChannel.read()/write() on windows do serialized read/write.
Better nio for windows is on java 7 todo list. But I don't think they can
finish it on schedule.
So do we need to have multiple cache files to workaround this win-specific
bug?
I'm reverting this before releasing 1144. You can put it back afterwards, but
it needs to be rewritten a bit: FileChannel.read() is not guaranteed to
complete the operation (unlike RandomAccessFile.readFully()). Also there are
a few RuntimeException's that you might want to catch explicitly.
On Sat, Apr 26, 2008 at 7:04 AM, Matthew Toseland
[EMAIL PROTECTED] wrote:
I'm reverting this before releasing 1144. You can put it back afterwards, but
it needs to be rewritten a bit: FileChannel.read() is not guaranteed to
complete the operation (unlike RandomAccessFile.readFully()). Also
