I'm replying on my message.
There is obviosly problem with low SP values in EXE headers. There is 0000 on 10h offset in fury.exe.

I think there is same problem in grand prix (gp.exe)

I changed SP initial value to 0040h and the game works.

But still there is a problem.

Guess you are not using 0.73

I'm using 0.73.

More precisely one built on 20.02.2010.

Download it from www.dosbox.com

I thought that daily builds from http://builds.tharos-online.de/ are reliable.
I tried latest build and had the same problem.
As I know version on dosbox.com is basic 0.73 version without further changes mentioned in logs.

they are not

they are reliable, but not supported.
You are running into an intentional error

OK. Is that issue with IRET replaced with RETF planed to be solved?

what are you talking about ?
execute using retf is the correct way and that is what we do

Maybe I misunderstood something. I realized that this caused the problem, but obviously it doesn't. Then, what causes the problem with low SP values?

can you finally get 0.73 from www.dosbox.com ??

the game sets very low stack values and we weren't 100% sure on how to handle the stack wrapping in that case.
So we made an exit condition there so we will get feedback on games that have that unusual low stack values in order to see which way of handling them makes them all work.
Hence an intentional error

OK. Thanks.
I'll report about it if it happend again.

P.S. I don't need to download anything from www.dosbox.com because i have changed initial SP value in exe header and the game works. I was just interested in the issue.

Same issue with Lotus: The Ultimate Challenge

Same "fix" for that "bug".

"just" the e_exit to a log_msg ?

If it is assumed that it was added to annoy people who use a development
build but can't post a useful bug report, then yes.

well bug reports come in a lot of flavours. 😉

Qbix wrote:

{...} Hence an intentional error

Some call them cowboys; these be words of a King.