Re: [hatari-devel] OS X performance problem

Eero,

The status bar update fixed the problem. It is great to have it back! Thanks for making the change so that the status bar is usable on OS X. I do notice that the LED simply turns on and off now. In some of the earlier development versions and possibly 1.7, the drive LED actually pulsed to simulate drive activity. However, since the track readout is there, that gives me the same information. I used to look at the pulsing drive LED to confirm that the disk was being read. I did see one brief pulse, but that could have been my imagination.

Thanks again for working around the limitations in the OS X SDL 1.2 port. SDL 2 may well be better (For instance, I know it supports real OS X fullscreen mode), but I am sure that is a longer-term project. Considering the main developers are Linux programmers (as I recall), I would have been shocked if the SDL Linux port was sub-par. It is disappointing how the Mac SDL port turned out though.

@Nicolas - I tried Lethal Xcess in STE mode and whether I had the blitter turned on or not, I had the flickering effect. However, it appears to work fine in STF mode. As you mentioned, it seemed to work most of the time in STF mode. In any case, it sounds like fixing the root cause will wait until after 1.8. I cannot speak for anyone else, but that would be fine with me.

I have only tried a few STX files so far, but everything has worked perfectly.

Bob C

On May 29, 2014, at 9:28 AM, Eero Tamminen <oak@xxxxxxxxxxxxxx> wrote:

Hi,

On torstai 29 toukokuu 2014, Eero Tamminen wrote:
On torstai 29 toukokuu 2014, Bob Carpenter wrote:
Do you get a problem if you use just HD, or drive LED:
--statusbar=no --drive-led=yes

?

I just tried this and Hatari was very sluggish again.

Both when not using floppy and when having just drive led?

It is very strange because the drive LEDs and the status bar worked
until the most recent commits.

If it has change recently it means that either your SDL version changed,
or somehow also LED update range changed.

I'll look into it.

FDC track information caused statusbar to be updated
on every screen update. I've changed the code to update
statusbar only when some of the information shown by
it actually changes.

There should have been no changes on how often overlay
LED is updated. Could you re-verify that it didn't really
regress?

- Eero