Can't break on _bank > 15

[+Karl G]

I'm using Stella 6.2.1 on Mac, and I've been having trouble getting certain breakpoints to trigger for a while, but hadn't taken the time to try to narrow it down until now. This is with my 128K WIP game, which uses DFSC bankswitching. I think, although I'm not positive, that the times my breakpoints won't trigger are when they reside in bank numbers > 15.  For example, this doesn't work in a case where I know the game hits the code in question:

 

> breakif {_bank==$10 && pc==LoadDungeon}

 

Trying it this way also doesn't work:

 

> breakif {_bank==16 && pc==LoadDungeon}

 

Finally, I tried this just to break if the bank is accessed at all:

 

> breakif _bank==$10

 

And for completeness:

 

> breakif _bank==16

 

None of these trigger in a case where I know for 100% certainty that the code in this bank is being run.

 

Any possible ideas as to what could be the issue? The source isn't public, but I can make it available privately if needed.

[+stephena]

It's hard to say without trying it myself.  I will say that the bases could be an issue.  So the default base is hex, and when you say 16, that's $16.  If you want decimal 16, you need to use '#16'.

 

Honestly, it might be better if you could provide a sample ROM privately.  So let's follow this up in PM if want to go that route.

[+Karl G]

Okay, instead of asking you to wade through my byzantine code, I made a simplified example with a minimal DFSC template.

 

If right difficulty = B, then the routine LowBankColor is run from bank 1 every frame, which sets the background color to green.

 

If right difficulty = A, then the routine HighBankColor is tin from bank 1F every frame, which sets the background color to red.

 

I am able to cause Stella to break when LowBankColor is run from bank 1, but not when HighBankColor is run from bank 1F.

 

dfsc-example.zip

[+stephena]

Confirmed to be a bug in 6.2.1, but seems to be fixed in the latest internal code (what will become 6.3).  We hope to do a new release over the next few weeks.  I can send you a personal macOS copy now if you like.

[+Karl G]

Thanks! I can probably wait, or compile my own dev version if I need it sooner, though. I'm just glad it wasn't just me being stupid this time.

[+stephena]

Just following up on this for anyone else reading the thread.   This is a bug in 6.2.1, and we've confirmed it is fixed in the upcoming 6.3 release.