By some unknown reason, it compiles the engine quite well but at the second stage, when it calls the newly created GForth, it just dies with Segmentation Fault.
This, truly, just makes more and more evident that as soon as possible, I have to go and build the Mex4th. It's specifications are almost complete and instead of being playing with systems that are not of my own and that I do not understand completely (or I take an endless time until I do), I'll be playing with systems that I designed from scratch and that I do understand entirely.
The 4th-HTML has yet some limitations due to 2 factors: in first place, is not yet totally developed. There are a large number of features that are yet to be developed. And, for second (and the most "critical" reason) due to limitations of the GNU Forth itself.
Don't take me wrong; GNU Forth is quite complete but should have more commands to interact with the Linux OS (or any other). It's completely possible to implement them but I do not master GCC and I don't have the time to spend with it. So, it remains a to much complex system for what is to be accomplished.
Instead of relying on GCC to build the forth system for any target, the cross compilation (or meta compilation) is quite superior.
So, any new system of Forth should be developed from within Forth it self, by meta compiling it's sources and not depending from an external C compiler or any other external compiler.
That's how WinForth32 works and it is quite more simple (in it's engine and not only).
But I can't also stop the development of final products just to produce the necessary tools.
So, as soon as my present situation clears a little more and stabilizes, part of the dispensable time will be allotted to tools development and the other part will be allotted to the development of systems with those and other more primitive tools.
And I'm specially interested to see when final products will be possible to be seen and used.