Short: A memory `doubler' that WORKS (v1.1) Author: Cucumber Pythagoras Uploader: hoss satech net au (Ian Hoskins) Type: util/boot Replaces: MemPlus.lha Requires: Kickstart 3 Architecture: m68k-amigaos Mem+ v1.1! Unlike other so-called memory doublers, this one WORKS! Well it doesn't actually DOUBLE the memory, but can give you up to aroundabouts an extra 65-90% of memory! What you need is Kickstart 3 and at least 128k free. (A memory doubler that requires memory, it's a sad fact, but true.) This program runs as a commodity, but you can't remove or disable the program. If you do all your memory would be corrupt, and would probably cause a major GURU! So it doesn't let you. Nyah. The response time of the program is about 1 second behind, so if you click on something and nothing happens for a while, don't panic. *New for v1.1:* The response time - DRAMATICALLY improved! Almost instantaneous... OK and before you go and say this program doesn't work because Workbench still says you have 1 meg free rather than 1.5, well, this program doesn't patch that, but it should go down more slowly. (So, a program that usually takes up 400k will only take about 300 instead. On average. Or something like that.) To run it, either double click the icon or in CLI type: run <>Nil: MemPlus That's all you need to know. Updates!!! Yes! My program has had 3 posts in the first week, and I didn't even give out an EMail address! OK, yup, you're right... Post to me at the uploader address up the top, but address it to Cucumber Pythagoras cuz it ain't my account! So, what's new in v1.1? Well, thanks to Andrew Grant, you can now sorta redirect the cruncher workpile to a location on a disk-like device, "T:" is what he suggested, but I wouldn't recommend it. It makes yer harddrive REAL busy and noisy, and slows things down quite a bit. It however does save you 32k of ram. Whoopdee. Oh, before I go any further : I WILL NOT BE HELD RESPONSIBLE FOR ANYTHING GOING WRONG WITH YOUR SYSTEM FROM THE USAGE OF MY PROGRAM. IF YOU DO DECIDE TO USE THE VMEM OPTION AND YOUR HARDDRIVE PUTS ON A FIREWORKS DISPLAY COULD YOU PLEASE PHOTOGRAPH AND DIGITISE IT AND U/L IT ON THE NET SOMEWHERE SO WE CAN ALL HAVE A JOLLY GOOD LAUGH. AT YOUR EXPENSE. To use the VMem option, however, just click on the really friggin' COOL minature drawer icon (or the really shocking one if you're running a less-than- eight-colour-Workbench-setup), which will bring up a requestor asking to pick a path, any path. Any pathname with a ":" in it somewhere is a valid one, although if Mem+ can't actually locate that path VMem is turned back off. And if you think you could be a smartarse and scrub the tempfile it creates to try and crash the program :) then you're wrong because if the file ain't classified as `in use' by DOS, then it didn't need it anyway :) And if it is classified as in use, you can't delete it. Nyah. Of course you could reset the computer while it's writing the file, which means then you could delete the file (after your harddrive has finished revalidating) but hey - the joke's on you buddy, after you reset all your memory is cleared and the program isn't running anymore anyway :) A Hide Interface button has been added, which really should have been done in the last release, so you don't have to bring up the Commodities Exchange. And the Save Prefs button will save whether or not you have the VMem option on and the path, so you can run the program and not worry about changing that every bootup. Although I'd be worried about the noise your harddrive is making if I were you :) The buttons assume if you have 8 or more colours that the MagicWB standard palette is installed and if you have less than 8 colours, that the standard Workbench 2-3 palette is installed. Otherwise the buttons will look a little weird. But they still do the same thing. Now, Georgia Pristo (sounds like a good new name for Mr Sheen, Mr Pristo! Spray it on or whatever and Pristo! a brand new oventop!) suggested I patch Workbench to say the actual amount of memory you've got after crunching. Well, no. I won't. But there's a reason :) The crunching process is a rather complicated one and a badly written one, which makes it hard to incorperate code that keeps Workbench up to date. Also calculating exactly how much you have after crunching all the time makes things quite a lot slower. (And it would have to be calculated ALL the time, perhaps even create a seperate process for it so I can forget about it) Otherwise if, for instance, it's only calculated once every second or so, a program could (and probably would) ask for AllocMem() within the have-crunched/will-update period, and well, I don't know exactly what would happen, but it would probably not be too friendly, I can tell you. So if I made it recalculate after every crunchcycle, well, that's just another set of commands to slow the whole process down and honey, it's slow enough I promise :) So I have left that bit the same. In all actual fact, that is still the amount of memory you have, but programs are just taking up less of it. HOWEVER if you are really getting anxious to see how much newly-aquired-ram this program has given you, you will be given a rough estimate in the Interface window. If you activate it, the Workbench titlebar will be modified to show you a quick guess of how big the largest chunks of memory are after crunching and will update it every 10 seconds or so. Notice I said largest chunks, so there's a chance of it saying you've got less memory than what Workbench actually says, but Workbench counts all the fragmented bits of memory as well. They usually can't be used (most of the time memory fragments are only a few K, and not much fits in a few K these days) so I didn't bother with them. Problems: Yes, to get those buttons looking the way they do, I have used Shape Gadgets rather than just text ones! Why bother? Because Blitz doesn't apparently handle the hitting of textgadgets too well (it works for my A600 buddies, but not for me) ... I use GadgetStatus(0,x) for the True/False feedback it gives me, but when operating textgadgets, if it hasn't been hit the result is True, if it has been hit the result is On, and then it stays On, and if I toggle the gadget back Off, it goes Off, rather than True. There doesn't seem to be a way to Toggle it back to True :) I didn't setup the gadgets as cycle gadgets, so it should be being handled just the same as when a shapegadget gets hit, right? Of course, I must be wrong. Cuz it isn't. Oh well, at least these buttons look nice. :) Text gadgets are a little more boring really. And they only take up another 7k which I guess everyone can afford :)