CompuServe Thread

Release numbers?

19 messages in this thread
#29412From: SyndesisOct 20, 1992 1:18 PM
I'm looking for a list of the correspondences between the thirty-something version numbers and the real-world release numbers for AmigaDOS. I've lost track of which numbers go with which release. For example, there's V36, V37, V38 – V38 is 2.1, V37 is 2.0? and V36 is 1.3? Is there an even more detailed list of minor vrelease numbers in both 30-ish and 1.3.2 format that CBM has published somewhere?
#29426From: John Toebes/SYSOPOct 20, 1992 7:04 PM
I thought that there was a list of the release numbers in the last set of devcon notes. Unfortunately, mine have been loaned out at the moment and can't look them up. Is there something specific you are trying to get at? Most of the functionality should be better keyed off the presense of certain libraries (like gadtools…)
#29450From: Steve AhlstromOct 21, 1992 11:38 AM
John, Since I've been installing and using every intermediate release I've forgotten all the various version numbers. I can tell you tho (with some certainty) that the release version of 2.04 is 37.175 and the release version of 3.0 is 39.106. -sja
#29466From: SyndesisOct 21, 1992 10:32 PM
I'm getting ready to throw this darn machine out the window. Something's wrong, and I can't stand it. The GVP 040 board on loan to my local dealer has functioned flawlessly in both a stock A2000 rev 6, and a fully-loaded Toaster system. I can install InterChange Plus on my machine here without indicent. Installation proceeds smoothly on the dealer's machines, too – except on the A4000, when I try to start InterChange itself, either from the WB or CLI, it sometimes says "Out of memory" on a requester. Two-three attempts later, it launches fine. I'm running release 2.01 here. I'm thinking about buying a new A2000, either a complete machine or a new motherboard. Does CBM dev sales sell motherboards, or am I dreaming? I want to install from virgin 2.01 or 2.1 disks and slowly restore everything else. I think that should waste a good day or two, won't it?
#29475From: Dean BrownOct 22, 1992 1:10 AM
John, On the 4000 and the GVP '040 card make sure that the caches are turned off before loading your code. If it makes a difference then you probably have some self modifing code somewhere. (overlays can cause the same sorts of problems) Are you actually running 2.01 or 2.1? If 2.01 you really should upgrade to 2.04 at least. Dean DKB Software
#29535From: SyndesisOct 23, 1992 2:49 PM
I made very few changes to my code, and now it runs on the A4000. There were a few dusty places in the code that were still passing a "NULL" as the memory requirements in AllocRemember() calls. I changed them to MEMF_PUBLIC. Not having an A4000 at my side, I had to make several other changes at the same time, so I'll never know if that made a difference. I also made the GetScreenData() changes described in other messages. The "out of memory" message I saw was from *our* code, not WB. The function that performed the GetScreenData() was a void function, so it couldn't have been returning the "out of memory" error, I think it was the AllocRemember(). Which leads me to a more general question looming in my mind. By virtue of cut-and-paste editing, there's a lot of places in our code that ask for MEMF_PUBLIC memory for general, private allocations. I should change these to MEMF_ANY in hope for a future version of the OS that might give MMu protection of any non-public and non-chip requests. Are there other considerations I should re-think with the advent of the A4000, in regards to the MEMF_ flags?
#29540From: Dean BrownOct 23, 1992 6:51 PM
John, The only thing I can think of is to make sure that any memory allocations you do have the MEMF_PUBLIC flag set if the block of memory has any possibility of being passed to to another task to handle. If you don't then there is a 'possibility' that at some time in the future the other task won't be able to access that memory block. Other than that, I'd also strongly suggest you use the available OS functions to allocate any memory pertaining to graphics if such OS functions are available. In fact use the OS functions to allocate any OS structures if possible, even if it requires doing version checks to determine if the allocation functions exist. (There are a number of new allocation/initilization functions in 2.04, and even more in 3.0) Using the OS functions will isolate you from most, if not all, OS and hardware changes. Then again, you probably already know all this! <g> It appears that CBM is starting to accelerate the development of new OS's and new hardware. Designing according to the 'rules' is going to be even more important than ever was. Dean DKB Software
#29550From: SyndesisOct 23, 1992 9:44 PM
Almost all (by bulk) of the allocations I do need to be passed from program to program, that's the basis of InterChange. There's a few that don't. And we do use the MEMF_CHIP flag when appropriate.
#29619From: Vic WagnerOct 25, 1992 7:14 PM
John, I would change all MEMF_PUBLIC which isn't _required_ to be to MEMF_ANY. But then, we've been _assuming_ that MEMF_PUBLIC meant what it said in the ver 1.0 RKMs (memory which _may_ be referenced by more than one task or an interrupt routine….by OUR inference—not swapped out).
#29652From: SyndesisOct 26, 1992 1:22 PM
I think I'll worry about it when they add this protection. I think I'll know this moment when they have functions that let you change the memory attributes of a given piece of memory.
#29478From: Steve AhlstromOct 22, 1992 2:48 AM
If you're really running 2.01 you've got a LOT of problems! Get a release version of the OS (2.04 or 2.1).
#29534From: SyndesisOct 23, 1992 2:42 PM
Oh, wow, I'll update from the 2.1 disks we just got from CATS.
#29536From: SyndesisOct 23, 1992 2:50 PM
I found that my local dealer would perform the motherboard swap for about $266. That's a bargain.
#29465From: Steve Bennett/SYSOPOct 21, 1992 10:11 PM
RKM:Libraries, Third Edition, page 10 says: V30 – 1.0 V31 – 1.1 NTSC only V32 – 1.1 PAL only V33 – 1.2 V34 – 1.3 V35 – Special kickstart to support A2024 monitor V36 – 2.0 (old release on A3000) V37 – 2.04 (ROM release) As you've said, V38 is 2.1. Presumably V39 is 3.0 but as I haven't received my 3.0 disks yet, I can't say for sure. (Kindof odd, that – I ordered 2.1 as part of the same order and got it immediately… I wonder what's holding up my 3.0 disks… Rather frustrating too, as they have the autodocs for both 2.1 and 3.0 – there aren't any docs on the 2.1 disks. <sigh>) –>Steve Bennett
#29467From: SyndesisOct 21, 1992 10:32 PM
We haven't received 3.0 disks at Syndesis, either, and we're commercial status.
#29618From: Vic WagnerOct 25, 1992 7:14 PM
Our 3.0 disks and docs arrived yesterday (Saturday). I _knew_ there'd be a good reason to come into the office.
#29621From: Steve Bennett/SYSOPOct 25, 1992 8:00 PM
Heh… So did mine. <grin> –>Steve Bennett
#29624From: Michael McCormickOct 25, 1992 9:28 PM
Still waiting for mine. (Sigh) MikeM:) "Amiga means never having to say your sorry!"
#29674From: Mike ToddOct 27, 1992 12:18 AM
I can confirm that v39 = 3.0 … I've had the 3.0 release disks for a while now, and it's nice to know that it, and the reKick Kickstart, still work on my old 1000 🙂 Mike (from the UK)