For those of us that are writing text mode apps in OS/2 and have noticed
the "missing keyboard" bug when executing java apps using the "-norestart"
switch on the command line... take note. I wrote the first part of the
following text to one of the JDK developers at IBM and his response is in the
second half. Needless to say, it's been A LONG time comming that someone has
admitted a bug in the JDK. I'm not beating anyones brow, actually I'm quite
happy and pleases that my bug reports and complaints actually got a response
at all. After all, TEXT MODE programming is frowned upon in the Java universe.
Little to they know, it's quite prevalent... <G> At least in my part of the
woods...
--- MY TEXT ---
Jerry McBride wrote:
> Andrew, this HAS to be a bug in the JDK. It's been present since 1.1 went ga. I
> can duplicate this on any of three platforms (os/2 3.0, 3.0 Connect and Warp
> Server 4.0) and have read posts from people using Warp 4.0 with the same
> problem reading key strokes from the console.
>
> Any ideas?
--- THE RESPONSE ---
Yes, this is a bug, already fixed in 1.1.6
The problem was that InputStream.available() was returning the wrong
value, so the BufferedReader was trying to read some
huge amount of data, presuming it was already there.
If you typed in 8000 characters (type 255 'a's, press enter,
then press PF3, Enter 32 times) the program will terminate.
My purpose for posting this in the NetRexx is to make aware, anyone, working
with NetRexx and Java and TEXT MODE applications that there is a rather nasty
bug using STDIN and someone has (hopefully) fixed it.
Now, back to work...
--
/-------------------------------------\
| Jerry McBride |
| |
| (
[hidden email] |
\-------------------------------------/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To unsubscribe from this mailing list ( ibm-netrexx ), please send a note to
[hidden email]
with the following message in the body of the note
unsubscribe ibm-netrexx <e-mail address>
Locked
