Hi Guys, Tim, I wonder if some of the code I used in ManpWin could be of use here. Although I have left out much of Fractint's code, I do have a reasonable working version under windows XP. Have you made any progress with my source code? I'm just interested to see how that is progressing. At least there is a 32 bit windows true colour platform to start from and maybe it is easier to start with simple working code and port the rest of the Fractint stuff from there, rather than trying to get the 16 bit code working. Just an idea. Seeya, Paul. ---------------------------------------------------------- Paul de Leeuw Computers NSW Central Coast, Australia Email: pdeleeuw@deleeuw.com.au www: < http://www.deleeuw.com.au> ABN 72 360 822 562 ---------------------------------------------------------- -----Original Message----- From: fractdev-bounces+pdeleeuw=telstra.com@mailman.xmission.com [mailto:fractdev-bounces+pdeleeuw=telstra.com@mailman.xmission.com] On Behalf Of Tim Wegner Sent: Sunday, 22 October 2006 2:30 AM To: Fractint developer's list Subject: Re: [Fractdev] ap math zooming in WinFract Jonathan wrote:

To add to the frustration of fixing the ap math zooming in WinFract, it actually works for the manzpower type.

I can't boot my old computer which has my Fractint environment. However I'm pretty sure I hosed it with an ill-advised Microsoft update, and I do have a backup, and can probably recover. Of more concern is that we're moving our kitchen into our dining room because of our house addition project, which is, shall we say, a slight disruption :-) Despite both of these issues, I'll make an effort to look at Winfract this weekend. Sooner or later I need to get MSC working in XP. I also have watcom installed, I'm nearly good to go with MSC except environment variables which collide with Watcom. I may just get rid of Watcom, I'm not using it. At least I got the DSL working, no thanks to the phone company, who came out and wired my splitter wrong while debugging a noise on the phone line. I fixed it myself once I realized that the phone line was OK outside the house. Two different DSL guys came out, saw the outside line was OK, didn't bother to look inside the NID (Network Interface Device) and left. I may sound a bit grumpy, but actually life is pretty good, just complex :-) Tim _______________________________________________ Fractdev mailing list Fractdev@mailman.xmission.com http://mailman.xmission.com/cgi-bin/mailman/listinfo/fractdev

Jonathan asked:

Is the BN math faster than BF?

BN is like integer math, only with a huge integer. It's fixed point with a lot of digits. It is faster than BF math. BF math also manages the exponent and so has more dynamic range. I'm at a disadvantage here because my development environment is not accessible so I can't look to check my facts ... I think fractint starts out with BN math and then uses BF math later when the zoom depth is greater.

What is Fractint doing different that allows it to work?

I don't know the answer to this.

Would there be any advantage to switching to Watcom for continued support of DOS Fractint?

The main advantages are that Watcom is now actively maintained open source so anyone could intstall the environment and help, and it supports both legacy memory modes such as we use and windows development. Theoretically one should be able to compile the DOS fractint and the WIndows 3.1 Winfract. But I wouldn't necessarily recommend that you look into Watcom, because it might take some work to port to Watcom (mostly figuring out overlay support) , and when you are done, you'd have what you started with - a medium model Fractint. We have a whole lot of choices of development environments we have investigated and could investigate. Having said that, I don't mean you shouldn't look into it, just trying to explain what Open Watcom is. The SciTech guys are maintaining Watcom. I recall their helping us in Fractint's early days. Changing subjects ... My symptoms on my old HP Pavillion with my Fractint environment are that it said the bios failed a checksum and it took me into the bios. I set the time and date and proceeded with the boot. It made it a fair way into the boot process and hung with a screeching sound. It boots OK to safe mode, and it boots a CD with Bart's PE (Windows XP) environment. I restored earlier images to no avail, so I don't think it's Windows update gone bad. The PC can't be too horribly broken. I'll pursue it a little while longer, but what I should really do is get MSC going on my XP machine. That won't be a big job. I haven't lost any data from my old machine because I am a backup fanatic, and in any case, I can still boot to modes that allow accessing the data. Unfortunately with these complications I can't quickly investigate the Winfract problems to help you out. Tim

Tim,

<begin off topic> I had a feeling I had dealt with this computer hang before. The computer (HP Pavilion 9680c vintage 1998) has a DAT tape drive connected via an adaptec AHA-1520 ISA (yup ISA!) SCSI adapter. I had scribbled on the adaptec documentation "disable COM1, enable VGA interrupt, reserve IRQ 10 for ISA", and yup, when I reset those in BIOS, that did it. Apparently the BIOS hiccoughed and reset itself, and the results were a hung computer. </off topic>

That would do it. Luckily you wrote it down. Who would remember that?

Does Tortoise work with CVS as before? Probably not, I'll most likely have to configure it to use keys rather than passwords, so I'm not home free yet. Have you committed anything to CVS since the fractalus migration, or is what you want me to see before that?

I am running pageant from the Startup directory so that the keys are always in place. Yes, I've committed to fractalus since the migration. However, what needs to be looked at all happened prior to the migration. Jonathan

Paul wrote:

I guess that explains why the computer date you send these emails from is two days in the future. ;-}

Yikes! Actually it is the other way around. The clock on my "new" (actually not so new) PC was two days off. I have XP set to synch with the internet, but I guess it doesn't attempt setting the day. Thanks for pointing that out. Now let's see if this message has the right date. Sending this 7:15 PM CDT. Tim

Jonathan wrote:

I don't see anywhere that this could happen.

I think you are right, I answered too quickly. I'm still going on memory (having failed to learn from experience not to trust it :-), but I think bn math is only accessed via a debugflag switch, and then only a few fractal rtypes (mandelbrot/julia) that were implemented using it. I think it turned out there was no real need for bn math API, though in theory it was a bit faster than bf math. Does that sound right? Tim

Jonathan wrote:

Yes, up until the point where you added the USEBN define to fractalp.c. Perhaps you added it to test the bn math and forgot to comment it out before you committed the changes.

Exactly right, I certainly didn't want to commit that change, so it's OK to undo it. Tim