Programming in C, C++ for Mac and Windows?

thomthom

Would the code be simpler if it didn't integrate Ruby?
If it only passed data back and forth? Or is it required to integrate in order to share the various data types?

tbd

@thomthom said:

Would the code be simpler if it didn't integrate Ruby?

you dont integrate Ruby, just the needed headers for Ruby types. as you can see the output .so is small, 10kb, without any size optimizations.

@thomthom said:

If it only passed data back and forth? Or is it required to integrate in order to share the various data types?

you only pass data back and forth.

thomthom

And what about Classes defined by SketchUp? How does one handle them?

tbd

you can use rb_funcallX (see Ruby Interpretor Interface) to call methods from the classes or you can use array/hashes to pass data

thomthom

Say I got a collection of Point3d, Vector3d. Would it be best to convert those to arrays of basic numbers in Ruby and passing onto C? Or would it really not matter?


static VALUE test(VALUE vector)
{
  VALUE arr;
  arr = rb_funcall(vector, rb_intern("to_a"), 0)
}

This would leave me with arr being an Ruby Array representing the vector, right?
I see there are rb_each and rb_iterate - should one use ruby iterators to loop through Ruby Arrays in C? Or can one use regular C loops?

Are there further reading than this page: http://www.ruby-doc.org/docs/ProgrammingRuby/html/ext_ruby.html

tbd

@thomthom said:

This would leave me with arr being an Ruby Array representing the vector, right?

right

@thomthom said:

I see there are rb_each and rb_iterate - should one use ruby iterators to loop through Ruby Arrays in C? Or can one use regular C loops?

definitely rb_ functions as you are manipulating an array of pointers (VALUE objects)

thomthom

I am confused again now. I got Pellets. And I did what the article said about generating a makefile. But I'm confused to how I compile it.
From what I understand, make is a UNIX command?

thomthom

Ok, getting close I think, - from the Pelles C Project options I added C:\Ruby\lib\ruby\1.8\i386-mswin32 to the includes path.

But now I get:
Building helloworld.obj. C:\Ruby\lib\ruby\1.8\i386-mswin32\config.h(2): fatal error #1014: #error: MSC version unmatch. *** Error code: 1 *** Done.

thomthom

Pelles crashed on me - and I swear I heard Homer Simpson's "doh!" as it crashed...

tbd

just double click on suext.ppj and it will open in PellesC IDE and then you can Project | Build and you will get a suext.so ready to use

thomthom

@unknownuser said:

just double click on suext.ppj and it will open in PellesC IDE and then you can Project | Build and you will get a suext.so ready to use

Tried that - but it complained about the project being in the wrong location. Thereafter I got the error I posted in previous post.

thomthom

When I google the error - all I find is people suggesting that one remove
#if _MSC_VER != 1200 #error MSC version unmatch #endif
from config.h.

I take it that you found a way around that?

thomthom

When I open the .ppj files in a text editor I see some minor differences.
What kind of project template do you use?

tbd

added suext6.ppj in SUExt dir on github. it should work now. I use PellesC 5 and something was changed in 6 that made that error.

thomthom

Thanks to TBD I manage to make my first Hello World SketchUp Ruby Extension running!
The man deserves a big cookie!

Chris Fullmer

Hey Congrats Thom! So when are you moving out of Architecture and into Programmer/Developper full time?

Chris

thomthom

when it pays more..?

thomthom

Say that I have a module Foo defined in my ruby script containing a method bar.

How can I reference Foo in C so I can for instance call bar?
symbol = rb_intern("bar"); rb_funcall(??? , symbol, 0);

Would I use VALUE rb_define_module(const char *name) ?

module = rb_define_module("Foo"); symbol = rb_intern("bar"); rb_funcall(module , symbol, 0);

tbd

thomthom: yes

when in doubt always try it. faster than posting a question on forum

thomthom

Just wanted to make sure - some times something appear to work - but later turns out to be very wrong.