AmigaOS 3.1.4 withdrawn from market after Amiga

[gulliver]

The joke is on your hysteric claims, but I will answer you:

Kickstart 1.2 had a bug which prevented Zorro II hardware from autobooting. It took Commodore about two years to fix it.

The first Amiga 600s that were sold, could not use the pcmcia and the IDE interfaces. It took them several kickstart releases to fix both issues. Customers had to buy them separately (Kickstart 2.x era).

And these are the ones that I can remember right now, there are many others.

Also this was in a time when Commodore was in charge, and they were a multi million dollar company with plenty of resources.

[kolla]

None of the bugs you mentioned were potentially dangerous, as in the system appearing to work normally, but "randomly" doing things it should not do.

[gulliver]

Originally Posted by kolla

None of the bugs you mentioned were potentially dangerous, as in the system appearing to work normally, but randomly doing things it should not do.

They make the affected hardware unusable. That is serious.

[kolla]

Originally Posted by gulliver

They make the affected hardware unusable.

But they don't screw up your files.

The big problem with this bug, is exactly that it is so subtle.

I know one bug that is similarly dangerous, the max transfer bug. But it only affects certain IDE controller, and is easy to work around.

[gulliver]

If you are looking for data altering bugs then you can look at amigaos 3.5 and 3.9 that had issues with drive geometry.

Then any app in those Oses that calculated drive space reported negative/weird numbers.

It even broke the Installer that could not calculate properly the free drive space available.

Large drive support was a Titanic failure that never got fixed during the lifetime of 3.5 and 3.9. And was one of their main features.

[hth313]

So to keep with tradition, we should accept bad bugs, not get fixes for years, keep paying and also have the source hidden, hmmm... to prevent people from fixing it themselves?

Well, at least I can deny you the "keep paying" part. I am happy to use the old version I have until I can get AROS up on my machine.

[kolla]

We are talking about bugs in kickstarts, OS 3.5 and 3.9 did not come with kickstarts.

I have seen scripts for patching that rely on certain version of source file being available, and that abort when it's not the case. Some of these versions of patch still output an empty file instead of a patched file. With v46 shell, the patch script will continue, and potentially overwrite the file it attempts to patch with an empty file...

[kolla]

Originally Posted by hth313

So to keep with tradition, we should accept bad bugs, not get fixes for years, keep paying and also have the source hidden, hmmm... to prevent people from fixing it themselves?

That is Hyperion's modus operandi, yes.
For now, OS3 is in the exact same mess as OS4.

[gulliver]

Originally Posted by hth313

So to keep with tradition, we should accept bad bugs, not get fixes for years, keep paying and also have the source hidden, hmmm... to prevent people from fixing it themselves?

Well, at least I can deny you the "keep paying" part. I am happy to use the old version I have until I can get AROS up on my machine.

Customers will get bugfixes when they are ready.

You should not accept anything, since you said you are not paying, hence you are not a customer.

But you can use, patch and fix yourself whatever you want in any OS you like.

If you want open source you can go the AROS route, or even better, all the way up to Linux. It is your choice.

And you don't deny me of anything. I don't get anything from you.

[kolla]

Here is an example of a completely randomly chosen script from aminet...

ExtractModule FROM "DEVS:AmigaOS ROM Update" Module 19 TO T:Shell-Seg
spatch -oT:Shell-Seg.new -pShell.pch T:Shell-Seg
copy T:Shell-Seg.new to L:Shell-Seg

What happens here when extarctmodules fail, and spatch outputs an empty file T:Shell-Seg.new?

[kolla]

Originally Posted by gulliver

Customers will get bugfixes when they are ready.

The customers are ready, bring it on already.

Joking aside, all experience from OS4 says that Hyperion releases fixes when it fits them, not when the fixes are ready or when developers are ready.

[gulliver]

Originally Posted by kolla

The customers are ready, bring it on already.

Yes, I know and understand that people are impatient.

I expect it to be released very soon.

I cannot confirm any date because we are trying to make sure we have tested the update enough. And when can you say that enough is enough? More if you consider we are not thousands of developers and testers.

In my personal opinion (others may differ) we are almost ready, some little testing remains, but definately not much.

[gulliver]

Originally Posted by kolla

Joking aside, all experience from OS4 says that Hyperion releases fixes when it fits them, not when the fixes are ready or when developers are ready.

Well in our particular case (68k), it was exclusively our desition. Hyperion only took the reasonably expected time to put up the files online (just a few couple of days).

I dont know how things work with OS4.

[kolla]

Are updates only for "registered" buyers?

[gulliver]

Originally Posted by kolla

Are updates only for "refistered" buyers?

I assume that, since all material related to 3.1.4 has been released that way.

But that depends exclusively on Hyperion. We can of course, suggest them otherwise if we find there is a good reason for that.

[kolla]

Originally Posted by gulliver

More if you consider we are not thousands of developers and testers.

Not even thousands of customers. This could be done in a much more efficient and customer friendly way. Release early, release often.

[gulliver]

Originally Posted by kolla

This could be done in a much more efficient and customer friendly way. Release early, release often.

We would need more testers and developers for that.

[kolla]

Originally Posted by gulliver

We would need more testers and developers for that.

That is what community is for, we are here.
Just convince Hyperion to cut the NDA crap.

[gulliver]

Originally Posted by kolla

That is what community is for, we are here.
Just convince Hyperion to cut the NDA crap.

I wish I could, but I am afraid this litigation does not help.
I will try anyway.

[daxb]

Originally Posted by kolla

cd RAM:bla
delete #? all quiet

a script with the above will abort on all kickstarts, except on 3.1.4 where it will...

I am not concerned about my own scripts, but I do observe what others put I their scripts, and I have seen quite a few "optimistic" boot floppy startup-sequence files.

But such a script is broken by design. Nobody does it if he or she is not a complete moron. Delete is always dangerous by itself. If you use it in a script you need to be really careful. In nearly all situation when you delete something by a script you delete only known files. I delete stuff in many scripts and never would use your example code. This is complete independent to the known bug. In theory there are many possibilities that a script destroys something without the need of a Shell/OS bug. So if you want to be save don't use them.

Btw. I'm searching my HD for " cd " and until now nothing dangerous found. I'll tell if there is a hit.