News 02.04.026

[Theo Gottwald]

Yesterday I learned that even if an Compiler has an Assembler and Optimizer - it does not mean that it works internally with real Assemblercode - like we know it.
Its working with binary Bytes ... so to get that into readable Assembler and make it visible need some sort of Dissassembly anyway.

While its not 100% ready (there seem to be some unwanted artefacts inside?)
I always wanted to see immediately what ASM-Code is been generated from my Basic Code and that will soon work.
See picture.



Also i am planning to integrate other Languages, between
#ASM ... #ENDASM (ASM include access to local variables)
#PCODE ... #ENDP (PILOT also with full access to variables)
#CCODE ... #ENDC (C-Code)
#BCODE ... #ENDB (BASIC)

Every day i have new Ideas what can be integrated.
Here you see an ... to my good old ATARI 400 Full-Screen Editor.
TO enter BASIC typical Commands like in that times ... (8-Bit/ 32 KB RAM)

It will also be possible to enter PILOT Commands like SAVE or LOAD etc. in an PILOT Window.



Ok, it seems to work now. The code is planned to be in a format that it can be pasted back into the editor and compiled.

[Theo Gottwald]

This is how I think it should be.
Left Side - the code.
Right side you can directly see what it becomes.

[Theo Gottwald]

What KI says about my project.
The really good news comes from Jürgen he seems to be near the first release soon.

[Theo Gottwald]

The AI Coding System we have still need Input from me all ?? Minutes.
I still think i am not using AI to its limits.
Seeing the situation, I think using the MCP-Server and the SPR i could do something better.

[Charles Pegge]

The really good news comes from Jürgen he seems to be near the first release soon.

Has Jürgen been tempted to improve or enhance the PowerBasic syntax ?

[Theo Gottwald]

He had some complex errors in memory management that would have left Memory leaks.
So after they are gone he thinks he is near an public release.
From what i know, he is vey near to powerbasic syntax, in a way that he can compile his own compiler that was originally compiled with PB now with his own compiler. That already works.
There are small differences like he wants variables after LOCAL etc. but i think he even makes a converter program for that.

PS: I knew C-Syntax is much easier to implement then Powerbasic. It looks like in 2 days it was done (this is the CX32 not the CX64 that is generally a bit more far).
Originally the KI Agent wanted to compile the C-Code using MingW32 Compiler to .obj and then link it into the final program. KI is lazy if the user does not watch.
But what is "Lazyness" other then optimization?

As already stated my Compiler will support Powerbasic Syntax (enhanced) and also (ATARI-)PILOT, and native C99, as well as #ASM .. #ENDASM; Syntax.
In Powerbasic i want addtional "MODULES" that will work like UNITS in Turbopascal.
And i want "Hooks" that can be attached to Variables and UDT's, and do something whenever any Value changes. UDT's will also be able to use dynamic strings. And i like the idea from charles to even add code to UDT's yet i have not yet a concept for that.

And all Languages can access all Variables. Means the ASM Code can directly access BASIC or C-declared variables. Full support for Variant included.

C, PILOT and ASM will be ready in short time. BASIC is much more complex due to the large number of >600
keywords that must be implemented in all stages.

[Theo Gottwald]

This looks morer like i think it should work.

[Theo Gottwald]

You need to understand german to understand this one ...