#23455 - wintermute - Tue Jul 13, 2004 9:44 am
http://www.work.de/nocash/gba.htm
unfortunately without the debugger but it does run ROM based images this time round
#23456 - caitsith2 - Tue Jul 13, 2004 11:14 am
Well, there is a debugger version, but even the homebrew developers have to pay some money for it. $15 US to be exact, which gives just the debugging feature, without source level debugging (unless the game was written completely in assembler), and without profiling.
#23460 - wintermute - Tue Jul 13, 2004 3:30 pm
wow, cool. I missed that bit.
Shame he didn't do this a long time ago but what the heck.
/me buggers off to register
#23461 - FluBBa - Tue Jul 13, 2004 3:52 pm
Nice, would have been cool with the profiler also but what can one ask for for $15? =)
_________________
I probably suck, my not is a programmer.
#23463 - dagamer34 - Tue Jul 13, 2004 4:14 pm
Hey, does it support linking? If so, then we can finally get some decent multiplayer testing around here!!!
_________________
Little kids and Playstation 2's don't mix. :(
#23469 - tepples - Tue Jul 13, 2004 5:47 pm
Pardon my l33t, but this is g00d n3wz0rz!
| Martin Korth wrote: |
| It is counting obvious programming bugs (such like attempts write into ROM, to access mis-aligned memory addresses, etc.) It's also adding 100 penalty bugs if the program doesn't support low power battery saving, the final rating is then: |
"Attempts to write into ROM"? I wonder what this will do to the people who test PogoShell or any PogoShell program that writes to the flash card ASIC's bankswitch registers in order to "exit to shell". I guess one would just have to be careful not to choose "Exit" during a No$ session.
However, I'd propose using a different scale that fans of a certain popular dance game would be more familiar with: - 0 bugs/s: Perfect
- 1-15 bugs/s: Great
- 16-199 bugs/s: Good
- 200-999 bugs/s: Boo
- More bugs/s: Miss
This would distinguish between game logic bugs that happen a few times a second (Great) and game engine bugs that happen once a frame (Good). Then a "combo" counter would count the number of Perfect seconds since the last Good, Boo, or Miss.
If you really need a profiler, then build profiling machinery into your program at the cost of a few scanlines of CPU time. When optimizing my code for speed, I've been able to get away with just counting scanlines that a function uses or even just changing the backdrop color after each major step of a computation.
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.
#23475 - Abscissa - Tue Jul 13, 2004 6:44 pm
That's awesome. But how useful would a non-source-level debugger be to C devlopers? Would there really be any point at all? I've stepped through a disassembly of my code in MappyVM and it didn't seem to help much ('Course, I'm really only familiar with x86 and 6502-style Asm)
#23479 - col - Tue Jul 13, 2004 9:12 pm
| dagamer34 wrote: |
| Hey, does it support linking? If so, then we can finally get some decent multiplayer testing around here!!! |
I just tested the free version, and yes it does support multiplayer, but its not so great for testing multiplayer code.
It seems to offer a 'perfect' environment where all the connected gbas run at the same speed and are in sync - thats not what you get with hardware !
maybe the full pro debugging version can emulate the syncronisation problems that real gbas have?
(and simulate corrupted packets...)
These real world hardware issues are what makes multiplayer code so difficult.
cheers
Col
#23482 - dagamer34 - Tue Jul 13, 2004 10:12 pm
| col wrote: |
| dagamer34 wrote: | | Hey, does it support linking? If so, then we can finally get some decent multiplayer testing around here!!! |
I just tested the free version, and yes it does support multiplayer, but its not so great for testing multiplayer code.
It seems to offer a 'perfect' environment where all the connected gbas run at the same speed and are in sync - thats not what you get with hardware !
maybe the full pro debugging version can emulate the syncronisation problems that real gbas have?
(and simulate corrupted packets...)
These real world hardware issues are what makes multiplayer code so difficult.
cheers
Col |
That's funny, I can't get anything to work. Does it emulate the sending of an multiboot ROM from the master GBA to slaves?
You think you can send that binary to my e-mail? dagamer34@yahoo.com
_________________
Little kids and Playstation 2's don't mix. :(
#23483 - Miked0801 - Tue Jul 13, 2004 10:32 pm
Yep, it doesn't emulate send errors - but those are easy to "Create" Just send crap across the link 1 in 1000 or so sends and see if you recover - this is how I tested it. And yes, even with source level stuff, it is still very useful for debugging with a bit of practice. I was using it before he go the source level stuff working. You still get function labels which is what you need. Go Martin!
#23486 - col - Wed Jul 14, 2004 12:51 am
| Miked0801 wrote: |
| Yep, it doesn't emulate send errors - but those are easy to "Create" Just send crap across the link 1 in 1000 or so sends and see if you recover - this is how I tested it. And yes, even with source level stuff, it is still very useful for debugging with a bit of practice. I was using it before he go the source level stuff working. You still get function labels which is what you need. Go Martin! |
The problems I have been having are related to the drifting (a)synchronisation - so i guess i'm stuck there.
But I'm really impressed by this emulator. Even just the fact that it starts up with realistic gba colours is enough to make it a useful tool. I wish i could afford to buy the full version :)
cheers
Col
#23491 - sgeos - Wed Jul 14, 2004 3:24 am
| col wrote: |
| The problems I have been having are related to the drifting (a)synchronisation - so i guess i'm stuck there. |
When (if) you seed your RNG with some version of time, calculate a fairly large random number. Make sure that each unit uses a different number. Each time you wait for vblank, generate a number from 0 to your big number. If the result is 0 then wait two vblanks instead of one. Not exactly drifting, but it will introduce unpredictable asynchronisation.
-Brendan
#23493 - MumblyJoe - Wed Jul 14, 2004 4:23 am
I am the best!
Apparently my latest few get 0 bugs :P
Although granted one of my first projects from ages ago did very very poorly... but I have learnt alot since then.
_________________
www.hungrydeveloper.com
Version 2.0 now up - guaranteed at least 100% more pleasing!
#23495 - Abscissa - Wed Jul 14, 2004 5:05 am
| MumblyJoe wrote: |
I am the best!
Apparently my latest few get 0 bugs :P |
Yay! Mine too! Although it did complain about battery drain and the checksum.
#23500 - FluBBa - Wed Jul 14, 2004 9:01 am
| tepples wrote: |
| "Attempts to write into ROM"? I wonder what this will do to the people who test PogoShell or any PogoShell program that writes to the flash card ASIC's bankswitch registers in order to "exit to shell". I guess one would just have to be careful not to choose "Exit" during a No$ session. |
Or try to use the RTC which also requires writing to the ROM area...
Apparently my emulators generate a lot of other errors as well ;)
And it seems it doesn't emulate the way I use the blending registers to do sprite priorities and the sleep mode.
Might be a while before I switch on second thought.
_________________
I probably suck, my not is a programmer.
#23510 - col - Wed Jul 14, 2004 2:17 pm
| sgeos wrote: |
| col wrote: | | The problems I have been having are related to the drifting (a)synchronisation - so i guess i'm stuck there. |
When (if) you seed your RNG with some version of time, calculate a fairly large random number. Make sure that each unit uses a different number. Each time you wait for vblank, generate a number from 0 to your big number. If the result is 0 then wait two vblanks instead of one. Not exactly drifting, but it will introduce unpredictable asynchronisation.
-Brendan |
I guess this is getting way off topic :)
Anyhow, the drifting i'm talking about is +/- a handfull of cycles, so randomly waiting for an extra game tick dosn't help.
Cheers
Col
#23518 - Abscissa - Wed Jul 14, 2004 4:56 pm
| col wrote: |
| sgeos wrote: | | col wrote: | | The problems I have been having are related to the drifting (a)synchronisation - so i guess i'm stuck there. |
When (if) you seed your RNG with some version of time, calculate a fairly large random number. Make sure that each unit uses a different number. Each time you wait for vblank, generate a number from 0 to your big number. If the result is 0 then wait two vblanks instead of one. Not exactly drifting, but it will introduce unpredictable asynchronisation.
-Brendan |
I guess this is getting way off topic :)
Anyhow, the drifting i'm talking about is +/- a handfull of cycles, so randomly waiting for an extra game tick dosn't help.
Cheers
Col |
Instead of using an extra vblank, you could just perform a few no-ops, or a short do-nothing loop using a small randomized number as the number of iterations. (Make sure to turn compiler optimizations off, I assume a good compiler would rip out no-ops and such.)
#23525 - tepples - Wed Jul 14, 2004 5:26 pm
GSM Player's font looks like crap because the method I used to scale a 6x8 pixel font to 12x16 requires mosaic, and unlike VBA 1.7.2, No$gba freeware 1.9 doesn't seem to emulate mosaic.
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.
#24050 - dovoto - Tue Jul 27, 2004 12:45 am
Does anyone know how to get to the shareware/homebrew version? I would like the homebrew version as it seems worth the 15$ but i cant seem to find a link to it. I emaild but, of course, i get no responce.
_________________
www.drunkencoders.com
#24051 - keldon - Tue Jul 27, 2004 12:58 am
send the money first; and give details for him to contact back, then he'll give you the "link" =)
#24081 - Miked0801 - Tue Jul 27, 2004 6:08 pm
Lol. I wish he wouldn't go hide. He needs to setup a better business front for sure... :)
#24110 - tepples - Wed Jul 28, 2004 2:18 am
Are the restrictions on non-U.S. PayPal accounts so strict that Mr. Korth couldn't just start taking payments through one of those?
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.
#24121 - DiscoStew - Wed Jul 28, 2004 5:28 am
There have been numerous times where he got scammed, one in particular who emailed saying he sent the money. He took the guy's word for it and gave him a reg key, and that reg key ended up on the internet. He is just probably being very cautious, and I wouldn't blame him. One of these days I'll send some money so that I can get my own copy.
_________________
DS - It's all about DiscoStew
#24189 - wintermute - Thu Jul 29, 2004 11:11 am
| tepples wrote: |
| Are the restrictions on non-U.S. PayPal accounts so strict that Mr. Korth couldn't just start taking payments through one of those? |
A lot of people still don't trust Paypal
http://www.paypalsucks.com/
I checked with my bank yesterday about the cost of transferring funds to Martin's account directly - turns out to be $15. Bearable if you're going for the full single user license at $750, bit crap for the shareware version.
I did ask Martin if he would take Paypal but I've had no response as yet.
Anyone got any suggestions on the best way to pay him?
Last edited by wintermute on Fri Jul 30, 2004 4:18 pm; edited 1 time in total
#24247 - ecurtz - Fri Jul 30, 2004 4:14 pm
Any volunteers to collect PayPal and do a bulk transfer to his account with the list of names? I'd do it, but I'm a bit scatterbrained about keeping track of this sort of thing. I'd gladly send $15 plus a portion of the bank fee to any of the regulars here though...
#24250 - wintermute - Fri Jul 30, 2004 5:07 pm
I've just emailed Martin to ask if he'd be happy with a bulk registration like this. If I get a reply in the affirmative I'd be happy to make the arrangements. Please feel free to PM me with your email if you're interested in registering this way.
#24256 - wintermute - Fri Jul 30, 2004 6:46 pm
Update, just received a mail from Martin saying he's in the process of setting up a paypal account - he's just registered and going through the verification process so hopefully that will be up and running soon. He's suggested mailing cash as an alternative so I might go that route instead.