This is the mail archive of the mailing list for the Java project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] for Review: SelectorImpl.wakeup() + Synchronization

>> Reading the JavaDoc, I think I see how they want us to implement interruption,
>> but I think they way we've implemented interruptible I/O on POSIX and Win32
>> precludes us from doing this and forces us to do the inverse: that wakeup()
>> uses Thread.interrupt() under the covers.
>I think I need to dig into chaos some time soon.

I'll redo the patch with the corrections you mentioned, leave AbstractSelector
alone and make the begin() and end() calls (which will do nothing). I'll
let you figure out what to do from there. If it's any consolation, the output
diffs beautifully between Sun's JRE and post-patch gcj.

I'm trying to get in the Sun people's heads, but can't figure out what they
want begin() and end() for (in AbstractSelector) or why they think Thread.interrupt()
should call wakeup() instead of the other way around. Even on Win32, it makes
more sense (to me) to leverage Thread.interrupt() for wakeup()....

-- Mohan

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]