Re: [hatari-devel] Lethal Xcess load state

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 06/18/2018 12:10 PM, Vincent Rivière wrote:
> On 18/06/2018 à 10:38, Eero Tamminen wrote:
> > So you're saying that there's some difference in Hatari startup
> > state from one run to another, when nothing has been changed in
> > Hatari configuration / setup.
> Exactly.
>
> > Are you using a machine configuration with RTC (MegaST/STE, TT, Falcon),
>
> I use STe, I believe RTC is off.
>
> (I would have preferred using ST, but Lethal Xcess fails due to other 
> sync scroll bug in Hatari).
>
> For memories, I use pretty recent Hatari HG on Ubuntu 18.04 Bionic Beaver.
>
> > or using "--video-timing random" setting ("VideoTiming = 0" in your 
> > Hatari config file)?
> No, I use fixed wake state (default).
>
> > Nicolas mentioned that the game is a bit buggy.
>
> I can easily believe that.
> But it has always worked flawlessly on my real STe.
>
> And I can't find any good reason to see different result after loading 
> snapshots, depending on different Hatari runs.
>
> Of course, I can understand that, as any other software, Hatari has 
> bugs. In that case, they must be acknowledged, and considered as "to be 
> fixed some day".

Sure, I just wanted to make sure it's not something to do with
Hatari config.


> It just looks like something in the emulated STe hardware is not 
> properly reinitialized on loading snapshot. That strange palette bug is 
> very visible. By thinking twice, that menu screen uses rasters. Maybe it 
> is just Timer B which is not properly restarted. Or issue with 
> interrupts, acknowledges, etc.

Is this with EmuTOS, real TOS or both (their startups differ a bit)?

And could you provide the snapshot?


	- Eero