Uploaded a new iteration of the patch, reflecting the comments so far, here:
http://cr.openjdk.java.net/~jlahoda/8080679/webrev.02/full/
A delta patch from previous iteration:
http://cr.openjdk.java.net/~jlahoda/8080679/webrev.02/delta/
Any feedback on this would be appreciated!
Thanks for the comments!
Jan
On 26.6.2015 10:45, Alan Bateman wrote:
On 26/06/2015 08:58, Jan Lahoda wrote:
:
The native method readKeyEvent seems to do a FindClass per key event.
Maybe this is from the upstream project but I would think it would be
more efficient to cache this in a global ref. It would also be more
I will work on this.
Related to this is that you probably want the wide ReadConsoleInput
rather than ReadConsoleInputA.
efficient if the INPUT_RECORD were just returned and avoid the upstream
to create the object but that might be too much to change when you are
I am sorry, but I don't follow - I don't have much experience with
native programing for JDK. Are there some more details on what I
should do?
If the subset of the KEY_EVENT_RECORD that you need will fit in a jlong
then you could return that to avoid the upcall. I don't think this is
important for the initial push, I think the main thing to fix here is
the FindClass usage.
-Alan
