Benkei Gaiden - Suna no Shou (J) [lazy bugtesters]

[Eyedunno]

Benkei Gaiden: Suna no Shou is one of my favorite SFC RPGs, but for some reason, levelling up to lv. 77 results in a freeze in ZSNES (and a slew of other emulators!). I asked about this here maybe four (!) years ago, and pagefault said he'd check it out, but I never saw another reply, or any other info for that matter, so I guess I'll ask again (both here and on Bugzilla, to try and cover my bases).

ROM details:

---------------------Internal ROM Info----------------------
File: Benkei Gaiden - Suna no Shou (J).smc
Name: BENKEI -SUNA- Company: Sunsoft
Header: None Bank: LoROM
Interleaved: No SRAM: 64 Kb
Type: Normal + Batt ROM: 12 Mb
Country: Japan Video: NTSC
ROM Speed: 200ns (SlowROM) Revision: 1.0
Checksum: Good 0xBEC1 CRC32: 2743017D
MD5: 5F579309272216100A6CFF92DE297479
--------------------------Database--------------------------
Name: Benkei Gaiden - Suna no Shou
Country: Japan Revision: 1.0
Port 1: Gamepad Port 2: Gamepad
Genre 1: RPG Genre 2: Turn Based

---------------------Internal ROM Info----------------------
File: Benkei Gaiden - Suna no Syou (J).smc
Name: BENKEI -SUNA- Company: Sunsoft
Header: SWC Bank: LoROM
Interleaved: No SRAM: 64 Kb
Type: Normal + Batt ROM: 12 Mb
Country: Japan Video: NTSC
ROM Speed: 200ns (SlowROM) Revision: 1.0
Checksum: Good 0xBEC1 CRC32: 2743017D
MD5: 5F579309272216100A6CFF92DE297479
--------------------------Database--------------------------
Name: Benkei Gaiden - Suna no Shou
Country: Japan Revision: 1.0
Port 1: Gamepad Port 2: Gamepad
Genre 1: RPG Genre 2: Turn Based

Both ROMs result in the same error. In 1.40 and 1.42, the game freezes (but I can still access the ZSNES menu). In the latest (5/14) WIP, ZSNES itself crashes.

My computer specs are as follows:
Dell Dimension 2400
Celeron 2GHz
1GB RAM
Edit: Windows XP, ZSNES Windows version (but I've replicated the bug before in older DOS versions and a Linux version as well)

However, I had the same problem long ago on my old P2-333, and I was also able to replicate it on my friend's computer (specs unknown), so it should be replicable for anyone.

Steps to replicate:
1. Open the savestate (which is at level 76, and just close enough to 77 that killing just about anything will get you there).
2. Get a random encounter.
3. Kill everything (just mashing A should be enough) and scroll through the messages (again with A).
4. Voila. Crash.

Another link to the savestate, just in case

[Deathlike2]

I'd just like to point out that both ROMs are exactly the same (they are just renamed differently).

[Eyedunno]

Actually, one has a header, and one doesn't. I don't know if that makes a difference or not. I assume not, but I pasted both anyway, just in case.

[Deathlike2]

I'm only saying... the header doesn't matter unless you're using IPS patches anyways.

[Eyedunno]

Oh, hey guys, I'm resurrecting this old thread (that references a REALLY old thread), but I just noticed the release of ZSNES 1.51, and I wanted to see if there was anything new on this game.

There isn't, but I found some info on a Japanese forum. You can choose a male or female main character at the beginning, and if you choose the male, the game will crash at level 77. The female character can apparently be maxed out (at level 80) with no problems.

So, while I suspected it for years, I now have confirmation that this is a bug in the code of the game itself, not in ZSNES. This is strange, because this is a quality RPG otherwise, and it's surprising that they wouldn't have noticed a thing like that. I can only guess the play testers didn't power level like I did.

Just thought I'd let everyone know as a public service announcement of sorts.

Now I think I might play the game again, this time as the female character, and maybe write an English walkthrough or something.

Edit: I made an additional discovery. By entering 7E108A50 as a cheat code while the main character is leading the party, then saving, one can automatically max out his level and thereby bypass the glitch.

[grinvader]

This proves that bugtesters used a similar code and didn't check each level.

Anyway if it was happening in the real thing there's no real issue, is there ? Thread renamed to make it explicit.

[pagefault]

Savestates are a bane because they contain a CPU state I am unable to use any savestate anymore since the CPU has drastically changed. So I would assume the bug has been fixed now or it will still be addressed.

Looking at the SPC it has bogus data in it, the SPC core has been changed so I don't think it's a bug anymore.

It's not about being 'lazy' it's the sheer amount of bugs we have to sort through and test, we can't test them all at once.

[grinvader]

It's a bug on real hardware, Pagefault. Non-issue.