Hi George. The problem seems to be related to the differences between M0CFA and M1CFA. They expect different numbers of parameters on the stack and thus they...
... Of course it's off topic, John. But then... ... ...you have no competetion! Let's see the w32f code when it's done. There may be ways to simplify the...
... and ... and ... Actually it's the entry code that differs. M0CFA pushes the OP onto the return stack and then takes the new OP from the parameter stack ...
... In the event viewer, you will get this for a kernel crash; Event Type: Information Event Source: Application Popup Event Category: None Event ID: 26 Date:...
... Error : The ... In WIN98 an error dialog comes up. Clicking on Details brings up FKERNEL caused a divide error in module FKERNEL.EXE at 0167:00401c32. ...
Alex McDonald wrote: [SNIP] ... Thanks for the tip, Alex. I was not thinking clearly when I responded, my bad. Anyway, as I said before, if the kernel is...
... ===snipped The version you're using is a relative rather than absolute address W32F; you must add EDI to all addresses you calculate, so the code needs...
... I thought about this possibility, but I also thought that rel>abs conversion is only needed when we pass the params to external API calls ? The new...
... Yes, for Forth words that's true. For CODE words, you have to do it all yourself for all addresses. EDI contains the base address. Here's an example; CODE...
... -<snip>- ... an ... for ... Surely the 6.09.00 kernel can be built with fixed load by setting up in the meta-compiler dialog the same as 6.06 and 6.08 (...
... You're right in respect to the first part. What took all the time & effort was the second part; building something that could then get it loaded. -- ...
... <phiho.hoang@r...> ... up ... while ... it ... I thought PhiHo was still using the wrapper ( which is why he's using V6.09.00 rather than V6.09.12 ) which...
... entry point ... MAIN *** ... It initialises the DLL library and PROC structures; the very next word (init-malloc) makes windows calls that use them. -- ...
... setting ... time & ... get ... using ... Not for fixed load points. The wrapper just loads the code where it can allocate memory iirc; there's no way to...
... True. It was an address. Oops. That's why realtive loadpoints are a pain in the neck; REL>ABS, ABS>REL and having to add EDI to every address in code. -- ...
... <george@t...> ... <alex@r...> ... a ... than ... I thought if the image is for fixed load-point the wrapper asked the OS for the correct memory and if...
... You're right again; I've just take a look at the original code, and it's possible to specify the address. There _may_ be a range of addresses that work...
... My reaction is a bit late due to some problems with my phone. Thanks for the list. I will look into it. Jos == 4ePost: 1,313 bytes in mail. Elapsed time to...
George Hubert wrote: [SNIP] ... I got 6.08.00 on my machine. The boot screen says it's loaded at origin 0x00000000. and : 0 rel>abs . displays 9043968 WF32...
... Yes if it was fixed load-point the origin would be 0x8A0000 and 0 rel>abs . would display 0 ... 6.08 ... It's a dialog rather than a menu. IIRC you have to...
... Thanks, George. That did it . Now I have 6.08 with fixed loadpoint. Now, only if I know what part of meta.f should be scruntinized ;-) Cheers, PhiHo....
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
