DukRow

2017-03-21

I seem to be missing some of the fields? The “each” method isn’t working for behaviour_test.row.js

(started using GitLab to organise issues)

https://gitlab.com/g-pechorin/dukrow/issues/1

the variables being set by the script aren’t being saved … because … of … what?

the functions are being save just fine, but, the variables aren’t?

why aren’t variables being set?

2017-03-20

Had a “sick weekend” where I … think that I got “a lot” done … maybe …

Much progress implementing, very little written planning, now (sort of) stuck on pushing through encoding with the new system. If I didn’t have unit tests - this would be a nightmare.

Seems to be saving most-of the first unit test - yay!
Not sure why it’s skipping

Suboxone “closure” pushing

N pre-args are pushed on the stack;

	sub.push<const char*>("arg0-n")
	...
	sub.push<const char*>("arg0-2")
	sub.push<const char*>("arg0-1")

The “push closure” method is prototype

	template<nUpVals: uint16, argC: uint16, returns: bool>
	def suboxone::push_closure(code:(suboxone&) => void)

Before the code is executed, the pre-args will be pre-pended to the natural args After the code completes - the top arg will be the result, unless, “FALSE” was passed as the return type

The templates are … sort of … a bad idea for code-size

Defo take advantage of 0-arg stuff;

There will/should be a (shell based?) method pushing thing

	template<T: typename, nUpVals: uint16, argC: uint16, returns: bool>
	def suboxone::shell<T>::push_method(code:(T&, suboxone&) => void)

This one will wrap the above (I expect) but examine “this”

2017-03-17

It looks like it just needs more work to carry on.

I copied in scriptie/require() ; seems fine
I added extra warnings to the prototype code
I changed subspace to be … lighter … am calling it suboxone::subenv(std::function<void(void)>)

= I’m up to needing adaptations to the “inhume” module now = I think that I need an inheritance mechanism for vtables

so I added an extends function to do extensions and an “extends” function to lookup extensions
… maybe there’s something interesting here …

2017-03-16

My present troubles are associated with most recent rewriting of everything - I need a stable build soon. Ivan discussed “something” and I think that this framework might make an attractive way to run Haskell-as-a-scripting language or Haskell-as-embeddable; not sure. Both sound like relegating Haskell to a lesser role … which worries me.

I like the documentation that I wrote.

After creation of scriptie instances - it looks like the registered object … isn’t what I thought?

I moved the source code to the new home in a ??/??.cpp.??.hpp file
I’m going to try printing kv-pairs and see what’s in the table
…
looks like the tables are identical in both places … almost like the vtable functions aren’t being added in?

= I was using sub.setshell(-1); instead of sub.setshell(-2); - I was assigning the prototype as its own prototype