This is an archive of CompuServe forum messages from 1985 to 1995, as saved in transcripts by John Foust.
Search Results (255 messages)
Well, 64MB is a fair estimation of how much you could actually push a 3000 up to _today_, which is all that really matters to most people 😎 Of course with that much memory, you might as well buy a UPS and make a 50 MB RAD: disk, and you…
Several C= engineers have said there's no practical way to get a 3000 working at the same speed as a stock 500/1000/2000. Beyond the clock speed there are other problems (like the fact that memory accesses are 32-bits at a time) that make it very difficult to get the speeds…
If you read the RKMs carefully (Libraries and Devices, to be specific), mutual exclusion has never been supported by the OS (makes you wonder why they document an unimplimented functoin to such detail!). In essence, you'll have to roll your own mutual exclusion mechanism. -Mike
Except for extreme situations, the compiler must pass our internal tests before we release it. This has ended up delaying releases (it happened twice in 5.0d that the release was delayed as new bugs were found), but we feel it's generally better to delay a release and ship a stable…
On a 3000, chip ram is accessed 32 bits at a time instead of 16 bits, which can make for a significant performance improvement. -Mike
Glad to hear we fixed it. 😎 Library source for 5.0d is now available – you should probably call our update dept. at (201)542-2121 and ask for it, I believe it's free of charge if you have the lib source from a previous 5.0 version. For now, we've just completely…
There's absolutely nothing Commodore could have done about Dpaint III not working under 2.0. EA forgot a WaitBlt() somewhere, and because of that it dies under 2.0 since certain operations are faster than they were under 1.3. -Mike
The bug you're referring to wasn't anywhere near as simple as you suggest – the operators used in the expression, the types of the variables involved, and what was on the left hand side of the expression (i.e., was it an =, a +=, etc.) all played a part in…
Try this: void testf (str, …) char *str; /* I prefer to use ANSI syntax BTW, even though its ugly */ { va_list arg; va_start (arg, str); vprintf (str, arg); va_end (arg); } The main point is use of vprintf instead of printf. There's a big difference between a function…
We do extensively internal (alpha) testing, as well as outside (beta) testing. Our internal tests consist of: running stress-test type suites, such as Plum Hall's stuff. recompiling _everything_ with the new compiler, including itself (called bootstrapping). running as much PD and shareware example code as we can find thorugh the…
None whatsoever. We did fix a number of compiler bugs – the only thing I can think of is that the u-EMACS code depended on particular code generation which obscured a bug in the code. If you can isolate the problem to a short example I'd be glad to take…
Yep, that's right Don. Software publishers (like Manx and SAS) have little or no control over how their distributors and dealers conduct business. You should always choose the product you want based on what the product does and what you need, not on how a particular dealer treats you. If…
That's an added feature to SAS C 5.1 – they now accept more arguments for registerized calling (like the #pragma stuff allows) of functions. BltBitMap had too many params to work with a #pragma in previous versions, which is why that line was commented out.
Please define 'disproportion' – that terms kinda vague to me. I have personally run into very little software which was truly broken under 2.0 (several have minor cosmetic problems which don't bother me all that much). What titles have you had problems with, and what were the problems? -Mike
C= automatically sent out KS 36.143 to all commercial developers. I can only conclude that either yours got lost in the shuffle, or Gold Disk isn't a commercial developer. I suspect that once a 'truly' final 2.0 is out then upgrades will be sent out to all current 3000 owners…
OK the problem is you have something like this: a () { b(); } b () { a(); } Now let's say you call function 'a' here from main(). It will in turn call function 'b'. Which will call function 'a', which will call function 'b', etc. etc. without end.…
You've got a case of runaway recursion going on. direction() calls maze(), but maze() also calls direction(). Those two will just keep ping-ponging back and forth until you run out of stack space, and then 'poof' you'll start overwriting stuff. -Mike
Nope – we treat uninitialized data in a, um, somewhat weird way, due to the way C historically had to treat uninitialized data. So instead of a BSEG or a section,bss, you have to use the global or bss directives for each individual uninitialized variable that you've got. -Mike
No, the proper format for BSS is: bss label,size Where 'label' is the name of the thing, and 'size' is the number of bytes it comprises. Debug files are created by the linker, ln, when you use the -g switch. It gets the debug info from the individual object modules,…
Change DATA to say DSEG. For a BSS section, you're going to have to individually declare each item there using the BSS or Global directives (described in the manual). This is not really a bug, but rather lack-of-a-feature. -Mike
Yes, it could be bad RAM – tough to say though. BTW, note also that some incompatabilities are due to speed & 32-bit memory on the 3000 and have nothing to do with 2.0 (the NewTek dynamic ham viewer falls into this category). -Mike
143 is fully public – current 3000's are being shipped with it. -Mike
This one particular item has been discussed publicly on Bix, in public forums there. In addition, Manx has not received any copies of 2.0 beyond release 36.143 (which is public). Developer non-disclosure agreements cover items received by the developer. Since Manx has not received any beta copies of 2.0 since…
#649-
Don't Go-Amigo!!
Message #726
We did used to have 'maintenance contracts', which gave you a year of free updates, with one free upgrade guaranteed (i.e., if an upgrade didn't come out in a year, you got the next one whenver it came out). This turne dout to be an unbelievable hassle. When 5.0 came…
No, it's not DP:TAS's fault. The feature that broke it was added late in the 2.0 beta test cycle, and clearly the code in DP:TAS was totally clean given the amount of information available at the time of its development. Almost no one anticipated that the 2.0 release would have…
You need to use the mod function (%) not multiply (*) to get the range right. -Mike
ran() should only be used if you need a floating point random number. If you're interested in integer stuff (which is the usual case), you say something like: foo () { int val; val = rand(); } rand() is similar to ran(), except that it returns an integer between 0…
What's the array decl and the code referencing the array look like? It sounds like you're not allocating the array correctly & you end up referecing uninitialized stuff. -Mike
I'm sure I'll be accused of being 'unprofessional' (or worse..) for saying this but…… Certainly it sounds like Go's been jerking you around a bit, and it's understandable that you're upset about that. But it makes no sense to me that SAS will send a free 5.1 to Go who…
Of course if you're talking about anything after 36.143, then that's a different story….. -Mike
What developer agreement? 2.0 is a public release now, right up to the 36.143 revision, so there isn't any problem in discussing it in open forums. People have been talking about 2.0 problems (and good points!) quite candidly in the open forums on Bix. You sound like you're still using…
What version are you running? 36.143 is quite stable – I haven't had a 2.0 caused crash since going from .141 -> .143. What programs aren't working for ya? Commodore meanwhile has confirmed that the next rev of 2.0 will make copper clipping optional, with the default disabling it, so…
I don't know the specific details of your case Bob, but in general we will give free updates if you buy a new package within a month or so of an upgrade. Buying a package predicated on a new release coming out, and said release being delayed, is fairly grey…
Well, I guess that why I said I just suspected it 😎 So it sounds even easier than I had thought to get a license. -Mike
It's a blanket-deal, unlimited copies for a particular company (for a year I think). -Mike
BTW, I have a strong suspicion that you may have to be a commercial developer to get a workbench license… -Mike
Call CATS and tell 'em you want a Workbench License and they'll send you all the necessary forms etc. you need. It's pretty easy, and I don't think it costs much (they like to encourage workbench licenses, from what I can tell). I doubt they care where they get the…
Free update offers are generally done within a month or so of the _actual_ release date – rumored ones generally don't count 😎 As it was, you had the use of 3.6a for the months prior to the 5.0 release, and after being able to use it for that period…
Huh?
From the description, it sounds more like Go-Amigo's caught in the middle more than anything else. I don't know any distributors who like to handle update stuff. I'm not sure that Manx has an 'official' policy regarding buying an 'old' version right before or after a new release comes out.…
Would you care to comment on that Tom? Saying that 2.0 is a disappointment without citing specific problems is kind of an empty statement. -Mike
Yup, Leo does top notch stuff, and he's a genuine nice guy too 😎 Dunno if I could survive with just a bicycle and no car though! -Mike
My reply over in AmigaTech to this says pretty much the same thing, with one addition: it's usually the manufacturer/publisher who takes care of upgrades and not distributors/dealers, and if you buy a package within a month of so of new update becoming available, the publisher (SAS in this case)…
Hmm, I've heard rumors of a 'Dpaint IV', and I could definitely see that being a year away. But I sure hope they don't wait a year to fix the 2.0 problem – if they do I'll be forced to find some alternative to DPIII in the near future i.e.…
#89-
#Disney Studio
Message #131
Most developers have remarked upon the fact that so much software _does_ work under 2.0 – such a major upgrade of an OS usually has many more problems. Considering the vast improvement of 2.0 over previous versions and the level of compatability that Commodore has managed to achieve, I think…
Leo programmed fully within the guidelines set forth by Commodore. What he got caught by was a 'feature' added to 2.0. This feature could not have been anticipated by Leo in any way, shape, or form. Apparently when Commodore added the feature (Copper list clipping), they didn't realize that a…
Lattice C 5.0 & Go-Amigo
Message #151
Um, it hardly sounds like GO-AMIGO is ripping you off. It sounds more like a combination of mis-communication, mis-information, and general problems with SAS taking the compiler away from Lattice. True, the GO-AMIGO people shouldn't have told you you were getting 5.1 if they knew that you wouldn't be getting…
EA has acknowledged that fills are broken under 2.0. Basically, doing a fill on any object more complex than a box has about a 25% chance of just hanging – the "I'm busy" pointer stays up forever, and the fill never completes. Apparently there's a missing WaitBlt in their code…
Oh sorry – I TAS is an abbreviated form of DP:TAS, which is how Leo Schwab prefers to refer to it. Guess I should've made that clearer! -Mike
Actually, the majority of software works pretty well under 2.0, except for some minor cosmetic problems (people have a tendency to write one pixel into the window borders, for example). I now use 2.0 exclusively, except when I use dpaint III (which is broken under 2.0), and in that case…
Page 1 of 6