Here is a patch:

MemPatch

Its a quick and dirty hack I just knocked together in 5 mins. It basically patches the 68k AllocMem(), and AllocVec() functions and causes them to allocate an extra 8 bytes at the start and end of the buffer. It also places the requested allocation size just before the returned buffer in AllocVec() which some badly written software use.

I don't recommend installing this patch as a matter of course, only install it if you find it makes a particular piece of software you need start working.

Its a very ugly hack, and I would hope that someone could take this patch and make a per-application patch instead of a global patch like this one. (in other words, I hope this becomes obsolete very soon :)

Someone told me that it got Final Writer running under OS4 Final, although your mileage may vary.

(Oh, and in case someone was wondering, it wasn't the missing size field from AllocVec that prevents FW from working; it just happily overwrites the ends of its buffers).

_________________
Playstation Network ID: xeron6

Oh yes! If you're putting it in your startup-sequence, DO NOT forget the "Run NIL:" since you can't quit the patch and it'll halt the bootsequence forever.

_________________
Playstation Network ID: xeron6

@xeron
That sounds like a good compliment to MemTrailer, which can be got from the Aminet. I found that helped stop Wordworth from crashing under Amithlon/WinUAE.

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

@ChrisH

Having looked up memtrailer, I don't recommend running mempatch and memtrailer at the same time.

_________________
Playstation Network ID: xeron6

I already wrote a program called MemHack (actually it's the ADDCHIPMEM code from SetPatch) which does almost the same thing but it's a little more robust as it does not peek memory areas it didn't allocate. Checking for a 'os4x' signature, that's what MemPatch does, is a bit dangerous as it can cause unexpected visits from India.

_________________
Stephan Rupprecht - AmigaOS4 developer

@srupprecht

Yes, I'm fully aware that its not the best of ideas, but I figured the worst case would be a recoverable DSI from a 68k program that was already running when MemPatch was run. In most cases the check is fine, especially if MemPatch was run at the start of startup-sequence (not much 68k code running then). It seemed a better idea than just adding 16 to the allocation size and subtracting 8 from the buffer address without any kind of check, since that would probably be a much worse "crash and burn".

I wasn't aware of MemHack.

Edit: And as far as I'm aware "ADDCHIPMEM" doesn't get FinalWriter running, which MemPatch does.

Last edited by xeron on 29-Dec-2006 at 05:09 PM.
Last edited by xeron on 29-Dec-2006 at 05:08 PM.
Last edited by xeron on 29-Dec-2006 at 05:04 PM.

_________________
Playstation Network ID: xeron6

@xeron

I wrote MemHack yesterday, so you couldn't know it SetPatch does not solve the problem because MemHack is an extended version which also emulates MEMF_FAST (SP only patches MEMF_CHIP allocations).

_________________
Stephan Rupprecht - AmigaOS4 developer

If this patch does what it claims, it will allow older GoldED version to run under OS4, too, including the free GoldED AIX version. They require an AmigaOS3-style implementation of AllocVec().

This is great news. Can anyone else confirm that FinalWriter97 does indeed work with this?

@mpiva

I`ve got FW to load documents can`t save crashes on printing.
I`ll keep playing though.

OOPS! Buggered up installation of FW so ignore me sorry.

Cheers Frank

Last edited by Frank on 29-Dec-2006 at 06:46 PM.

@mpiva
No problems with FinalWriter here on AmigaOS 4.0 Final (clean install as per documentation).

I can open/save/print without any issues.

_________________
ExecSG Team Lead

@mpiva

I can confirm that Final Writer 97 launches ok with mempatch in place, but I may have something else going on here with my installation. As long as I don't make a new document or modify and existing one, seems to work fine. I can quit and relaunch ok several times. As soon as I save a change to a document, the fun starts. I can quit ok but if I relaunch Final Writer 97 it crashes. IBrowse crashes as well after I modify a document with Final Writer 97. Not sure if that is just coincidence.

---
redfox

@dietmar

Where can I download the free GoldEd AIX version to verify?

_________________
Playstation Network ID: xeron6

@mpiva

I reinstalled Final Writer 97 from a different back-up and now it is working ok with mempatch in place. load/edit/save ok now. Relaunch ok.

---
redfox

Last edited by redfox on 29-Dec-2006 at 07:55 PM.

@xeron

Final Writer 97 working fine here now with mempatch in place.

very much indeed.
redfox

@redfox

I can also confirm that FW97 appears to work OK with Mempatch activated. Unfortunately it seems to make IBrowse crash. Will investigate further.

TrevorDick

_________________
No, I don't need no reason, I'm just breezin'

@TrevorDick

Quote:

TrevorDick wrote: @redfox I can also confirm that FW97 appears to work OK with Mempatch activated. Unfortunately it seems to make IBrowse crash. Will investigate further. TrevorDick

I can confirm that problem indeed.... Final writer works with patch but then ibrowse dusted an DSI, most likely repairable if you put mempatch higher up in the startup-sequence.

_________________
See my blog and collection website! . https://www.blog.amigaguru.com

@TrevorDick

IBrowse crashes here too. At first I thought it was random.

---
redfox

Last edited by redfox on 29-Dec-2006 at 11:04 PM.

@xeron: Nowhere (it was a 3-day free download opportunity earlier this year but it's not freely distributable per se).

@xeron

Stop coding on other things and do the hivelytracker release (os4/68k v1.2) damnit :D