You read them wrong. It is running stable under other AmigaOS combinations but not with a Kickstart 3.2 ROM. Obviously, you will not understand what you can not understand so there is no point in continuing the hardware topic any longer.

He reported strange Auto-Config conflicts and sizing errors with Kickstart 3.2 but the other Kickstart versions worked just fine.

My DKB2632 tool had a resident memory option which worked under OS2.x but not OS3.x. It seems these resident memory tools can potentially fail with every update to the exec.library.

Hi everyone !
I finally took the plunge and updated my 3.2.1 install to 3.2.2.1 with 3.2.2 ROMs.
The update went fine (contrary to my previous attempt with custom 3.2.1 ROMs).

Still have an issue though :
textedit s:startup-sequence only shows the first 10 lines.
Also, for some reason, deficons doesn't see startup-sequence as a text file / script, but does see my startup-sequence backup as one (and the other scripts in S:). And it does show the content of my backup SS properly.

Thank you for your help and much appreciated work ! :)
Cheers ! :-)

EDIT : after adding a couple of lines in the SS (with ed), now TextEdit seems its entire content. Still, this is not very reassuring ;-) And now, deficons does detect its filetype properly. Strange.

@torturedutopian:

It seems to me that you SS got corrupted and a non-ascii character was present after the first 10 lines. This would cause both symptoms that you saw, though it appears to be fixed now.

Possible ! Although it was the default SS from 3.2.2.1...

I had 2 years of CS classes in high school (IBM 1130) and at university (1975) my first programming class was on a rebranded Xerox Sigma 7 (Samsung SDS). Naturally, at the low course levels, punch cards were used -- you punched them, then handed them to the CS person on duty in the basement, and later retrieved your cards and program output. On one particular failure, either the card reader, or the computer itself, had flipped a bit in one of the codes and caused the error.

A re-run of the cards worked the way it should have; needless to say, sh:bangheadt happens.

This behaviour is explainable. And not so scary. In an attempt to make the startup time of larger files faster the syntax highlighting is now done in a seperate task. As soon as it has highlighted past the visible portion it will render and continue highlighting the rest of the document in the background. The doucment is there and as you see if you edit it or scroll around in it it will be rendered.

I think your problem is tied to the fact that deficons doesn't detect it as a script. This couple with a possible bug in TextEdit that may not cause the render if the highlighting fails. See if the startup-sequence has somehow lost its script flag (eg use the workbench information window)

@Boemann, FYI, the recoverable alerts with 47.111 and Apollo V4 are not reproducible when I use Thor’s poolmem utility.