Bug in Hijack, I think. Help Mark.

Hi Mark,

As you may rememebr I mentioned this bug quite a while ago: when skipping to the next song, I get a W on next to the hdd icon, and I hear the disks spinning up. This happen almost always when I skip to teh next track, and sometimes even when the player goes to the next track in the playlist without me skipping anything, in the normal playback.
I thought this happened due to a bug in the software but now I think it depends on Hijack. Today I installed beta13 and I noticed I did not have the problem after the update. So I played around with it and never had the W sign. Then I installed Hijack 283 and I started to have the problem again. So a couple of hours ago I reinstalled the factory kernel, beta 13 plain, and I don't have the probloem anymore. I've been skipping back and forth for a couple of hours now,without having the W sign at all.
Maybe somebody else can confirm this if they want to try?

My configuration is:
Mark 2a, 60Gb (30+30) v. 2.00b13 .

Thanks

I got W three times now, for a much shorter amount of time, and still it is much much less frequent than when I have hijack installed. Maybe just less memory available for caching purposes when Hijack is in the kernel?

I'm having that problem too...and from time to time I also get pauses of a few seconds between tracks with the W-symbol on. During that time there's no error-messages on the serial-output or any strange noises from the HDs, so I also believe something's wrong with caching.
And it happens more often with Hijack... see this thread
(If you wonder about my last message there: after those 2 days it happened again... but the behaviour really seems to change with every new installation of the empeg-software...somehow...)

I see.
Tony, can I forward my msg to "Programming" too?
Thanks

Of course.

>I got W three times now, for a much shorter amount of time,
>and still it is much much less frequent than when I have hijack
>installed. Maybe just less memory available for caching purposes
>when Hijack is in the kernel?

Yes, I suppose that must be the issue. Now that frame-pointers are back in the Hijack kernels, they are much larger than they once were. I've never figured out the size, but maybe 100-200KB added size for a stock Hijack. This could be "just enough" to put it over the next-track readahead threshold for average sized tunes.

Cheers