CompuServe Thread

VP doesn't work with EGS

11 messages in this thread
#117849From: Roy S. LauferOct 13, 1993 4:19 PM
Clint, I have a problem with VP running when I have an EGS-workbench clone operating. My guess is that while VP is "initializing" it is polling the workbench screen to decide whether the computer it is running on is ECS or AGA. When it sees that MY EGS workbench screen can have a maximum of 256 colors it makes the reasonable assumption that I have an AGA chipset (which I don't have – I am using an A3000), and tries to set up it's display screen to be a 256 color one. The effort to try to open a 256 color screen on an ECS system leads to fatal error with a requester asking me if I would like to reboot the system! There must be simpler and less error prone method to poll the machine for its display capabilities. I realize that you're REALLY busy, but I do MISS my VP! BTW, there are other prgorams that have problems with an EGS WB, I will issue a list in th enear future for the interested bystanders out there <g>! // ** Commodore ** \X/ -The Bosnia of Computer Companies- -RSL-
#117864From: Clint Woeltjen@VRLIOct 13, 1993 5:24 PM
Roy: Are you running Version 3.0 or higher of AmigaDOS? If not, VistaPro will not assume that you have AGA, as it assumes that AGA will be running only with V39.0 and higher libraries. Currently that should only be a problem with developers who have legitimate beta copies of 3.0 running on non-AGA machines. VP will never attempt to open a 256 color screen anyway, (even if you have AGA) unless you specify 256-c in the GrModes panel. The main control screen in VP is 16 colors, even on AGA machines. I suspect there is another incompatibility with this WB Clone. Is this EGS WB Clone a commercial product? Shareware? Clint <VRLI>
#118189From: Roy S. LauferOct 15, 1993 10:49 PM
Clint, Guess who is a registered developer with access to OS 3.0 running on his A3000/25? <grin> The EGS workbench clone is a commercial piece of software which is being packaged with GVP's new Spectrum EGS board, and should be released in one form or another for other EGS style graphic boards in the near future. I would expect alot of A3000 owners to be using it to "upgrade" past the AGA capability of the A4000! One of these days, C= is very likely to release 3.x for non 1200 & 4000 owners, and I think that they MIGHT have similar problems! Good olde VP crashes right during the initiation portion when my WB is in 800x600x3 mode (with the capacity to go up to 256 colors listed in the viewmode's database). When I return my WB to native Amiga screen all is well. VP must be getting confused by looking at my WB screen and finding V39 libraries on an ECS system. I'm sure that I read about a reasonably simple method of polling a machine to detect whether it was AGA or ECS without checking the WB screen for maximum bit depth? Help! I miss my VP <g>! // ** Commodore ** \X/ -<Comment left to the reader's imagination>- -RSL-
#118200From: Clint Woeltjen@VRLIOct 15, 1993 11:33 PM
I now have reliable code to detect for AGA, which is not yet in VistaPro 3.05. However, I run with V39 mode turned on in debugging on my machine all of the time, and the only time I get into trouble is when I actually attempt to do an AGA function. The normal startup code knows nothing about V39 and should be able to open up just fine on any machine running V39 as long as the user doesn't try to open an AGA screen. VistaPro won't try to unless requested by the user. (In fact the code that opens the control panel should still run under version 1.1 of AmigaDOS.) The only V39 checks that exist are in the AGA specific code. That's why I suspect that the problem lies with GVP's EGS WorkBench and not V39. I have several beta testers who also run V39 on their 3000's and don't have any difficulties. VP will crash if you select HAM8 or 256-c on a non-AGA machine, but that doesn't seem to be your problem.
#118369From: Roy S. LauferOct 16, 1993 9:06 PM
The new AGA checking algorithm worked! VP now runs as nicely as ever! My guess is that it isn't the checking for V39 alone (otherwise others would have found this problem long ago), it's the combination of an ECS system (an A3000 in my case) running V39, AND also having a workbench screen that is capapable of displaying 256 colors. The previous versions of VP considered this (now rare, but soon to be less rare) to be too much like an AGA system, and proceeded to crash, attempting to open a nonexistant AGA screen on my A3000! I'm glad to have the problem fixed. <Now I'll have to get Innovatronics to "fix" CanDo – it has a similar problem.> I also want to thank you here in public for your help with this problem! I know of at least one person on the forum that has not been thrilled with my negative attitude regarding C=, but it's guys like you in this industry, and the forum, that keep me loyal to my old A3000. I'd love to clone you and a couple of other guys here and let them loose on West Chester (can you just see a few hundred Ben William clones overrunning C='s executive boardroom, reducing it to chop sooey<g>)! I just want to reiterate that I (along with a lot of people) appreciate the hard work and care that you take in producing products for the Amiga, and your tireless help when an occassional problem pops up! <Mark – It's guys like this ^ that will bring the Amiga back to full throttle, not a chorus of people whistling the praise of C= on the forum!>
#118412From: Clint Woeltjen@VRLIOct 17, 1993 12:42 AM
It could be SAS. VistaPro 3.05 won't run on 1.3 machines, while the SAS version (no changes) runs quite nicely. Several other crashing problems have disappeared with the SAS version. There isn't any code being executed at the time of the crash that was V39 dependent, so that can't be the problem, you can't crash code that you don't run. Again, VP NEVER opens an AGA screen unless the user specifically selects an AGA mode. (I like to get my bugs straight, it was another bug, not _that_ one! <grin>) I want my customers up an running, using VP! I don't know what plans VRLI has for 3.1, if they want to wait for 4.0 (probably still a year off) or to come out with an interim speedup with a few enhancements "update" release. That will be up to them to decide. I hope to get ahold of Brick's code which supports a couple of additional video cards and get that incorporated asap.
#118891From: Roy S. LauferOct 19, 1993 6:09 PM
Well… whatever the reason, I'm glad to have VP working like the olde pre-EGS days. SAS/c is a nice little compiler, eh (I like the way it is smart enough to open libraries, and close them automatically without the programmer writing the specific code when functions requiring specific library calls are used). Keep up the great work!
#118899From: Clint Woeltjen@VRLIOct 19, 1993 7:12 PM
SAS 6.3 is definitely a big improvement over the last version of Manx, but being a compiler writer, I cringe whenever I look at the code it generates. I keep promising myself that if I ever win the Lottery, that I will devote myself to writing a decent compiler/code generator for the Amiga. <grin> But SAS doesn't do bad considering that they actually have to make a profit selling them, and can't just write them as an artistic excersize.
#119067From: Roy S. LauferOct 20, 1993 4:14 PM
Clint, I would assume that writing "tight" optimizing code for a compiler is a "black art"! All things considered, I am happier with the integration of SAS/C as opposed to the rather primitive condition that Manx is (and unfortunately will always be) in. I hope that SAS/C 6.5 will be better still! (I'm gonna have to learn C++ one of these days!) // ** Commodore ** \X/ -<Comment left to the reader's imagination>- -RSL-
#119185From: Clint Woeltjen@VRLIOct 21, 1993 5:25 AM
Roy: SAS is at least working in the right direction. When I was writing compilers, I never understood why people had such problems with codegeneration and optimization. I always found the parsing to be the biggest pain (which was fortunate, as my partner on my team was just the opposite). We used a technique using two pseudo-languages, and optimizing for particular features at each language "boundary". This allowed us to keep track of what was in each pseudo register across rather long pieces of code, so when it came time to generate the actual assembly, there was very little of the "reload this register" phenomenon. Also, we could catch the "last occurance" so that we didn't need to transfer a register to a temp for a calculation if we knew that it wasn't going to be used any more. This allowed for "register variables" across small segments of code within a procedure, and then a new set of register variables later on in the same procedure. (Without the user telling us.) I was sorry that DRI dropped the Language division when it did, as the multiple front end/codegen/multiple assembler project was basically finished, but never saw the marketplace. A few years later we tried to see if we could license the code for the Amiga, but no-one there could tell us what happened to the source. And no-one had any interest in figuring it out. <sigh> The compiler had front ends for "C", Pascal, Basic and Fortran, and final assemblers for 8086/80286 and 68000. It would have been a nice addition to the Amiga repitoire back in '85-86. Clint <Chaocity><VRLI>
#119248From: Roy S. LauferOct 21, 1993 7:57 PM
Clint, I agree, it would've been nice to have even more compiler choices for the Amiga platform. Unfortunately code compilers is one of MANY categories that are experiencing a "thinning" lately. It's a good thing SAS/c is still around! As a guy that wrote for both CPU's do you share the "brain dead" opinion I have heard in the past about the Intel 80×86 series (atleast until the 486 series), especially when it came to memory management?