Re: [Jsynthlib-devel] Possible bug in Driver.java

[Hiroo H. <hir...@co...>]

Joe,

As I wrote on comment, I think the lines after '// Maybe you don't get
the expected path.' are used only for programming error of a Driver.
My take is to remove those lines and to show an error dialog to let users
send a bug report to us.

On Wed, 06 Apr 2005 17:04:05 +0000
Joe Emenaker <jo...@em...> wrote:

Joe> Around line 290 in Driver,java is the following method createPatches():
Joe> 
Joe> >    public IPatch[] createPatches(SysexMessage[] msgs) {
Joe> >         byte[] sysex = MidiUtil.sysexMessagesToByteArray(msgs);
Joe> >         IPatch[] patarray = DriverUtil.createPatches(sysex, getDevice());
Joe> >
Joe> >         // Maybe you don't get the expected patch!
Joe> >         // Check all devices/drivers again! Call fixpatch() if
Joe> > supportsPatch
Joe> >         // returns false.
Joe> >         // XXX Why don't we simply cause error? Hiroo
Joe> >         for (int k = 0; k < patarray.length; k++) {
Joe> >    ...
Joe> 
Joe> Due to a mistake on my part, I made a driver which didn't recognize the
Joe> sysex that I sent it. As a result, DriverUtil.createPatches() returned a
Joe> null which was then assigned to patarray[]. This caused an exception
Joe> when execution got to the patarray.length.
Joe> 
Joe> I didn't want to mess with this, since I didn't write it and I'm not
Joe> sure how it was intended to work.
Joe> 
Joe> Granted, it was a bug in my driver which exposed this bug.... but I
Joe> still think that this is a bug. It should handle this more gracefully.
Joe> What's the preferred behavior here? Should Driver.createPatches check
Joe> for a null patarray[], or should DriverUtil.createPatches be reworked so
Joe> as to guarantee that it never returns a null?
Joe> 
Joe> - Joe
-- 
Hiroo Hayashi