Re: [Emc-users] gcmc - G-Code Meta Compiler

[Bertho S. <be...@va...>]

On 10/30/2013 02:07 PM, andy pugh wrote:
> A way to configure CSS and feed-per-rev would be useful.

Hm, yes, those will need to be implemented.

As a hack...
You can put anything into the generated code with the literal()
function. It simply outputs anything you like.


> Also the other typical switches, such as radius comp on/off,
> absolute/relative (G90, G90.1)

The relative/absolute question is actually a bit more involved. I have
been pondering that one from when I started.

The question is whether you want the meta-language or the generated code
to support relative/absolute.

Currently, the generated G-code is absolute by choice because it is
easier to debug. There is, as such, no added value to mix
relative/absolute in the generated target.

Supporting relative addressing at the meta-language level /is/ a good
thing. I just have not yet addressed that point. Some functions need to
be added for relative movement and at least:
- move()	move_r()
- goto()	goto_r()
- arc_cw()	arc_cw_r()
- arc_ccw()	arc_ccw_r()

I do track the current absolute position internally, and it is exposed
to the script under the variable name "global_position".

There is also a variable called "global_offset", which holds the
command-line added offset for all generated code.


> Is G43 implicit in toolchange() ?

No, currently the toolchange() only emits a T code. I think it may be
appropriate to send a "Tx" followed by "G43 Hx". It may also be
appropriate to enable/disable this behavior with an optional second
boolean parameter to toolchange().


-- 
Greetings Bertho

(disclaimers are disclaimed)