librelist archives

Re: [shoes] Moving code from C to Ruby

From:

Steve Klabnik

Date:

2012-01-05 @ 19:56

Okay, so I have a first version up here:
https://github.com/shoes/shoes/tree/c_to_ruby

The strangest thing is happening to me, though: it's not actually
running my Ruby. Even though I pulled out all of the C, when I run
this program:

Shoes.app do
  c1 = rgb 255, 120, 0
  c2 = rgb 0, 255, 120
  c3 = rgb 255, 120, 0
  c4 = rgb 20, 20, 20

  stack do
    if c1 == c2
      para "c1 == c2 (undesired behaviour)"
    else
      para "c1 != c2 (desired behaviour)"
    end

    if c1 == c3
      para "c1 == c3 (desired behaviour)"
    else
      para "c1 != c3 (undesired behaviour)"
    end

    if c1 > c4
      para "c1 > c4 (desired behaviour)"
    else
      para "c1 <= c4 (undesired behaviour)"
    end
  end

end

I get the old C behavior. Not my new Ruby behavior. I put alerts into
the Ruby, and they don't show up either.

Am I missing something? I did the whole clean thing, rebuilt, all of
that. The Ruby code does show up in the resulting .app.

On Jan 5, 2012, at 1:56 PM, Steve Klabnik wrote:

> Okay, so I have a first version up here:
> https://github.com/shoes/shoes/tree/c_to_ruby
> 
> The strangest thing is happening to me, though: it's not actually
> running my Ruby. Even though I pulled out all of the C, when I run
> this program:
> 
> Am I missing something? I did the whole clean thing, rebuilt, all of
> that. The Ruby code does show up in the resulting .app.

It's possible that the makefile is fubar. You should remove the *.o files by hand.

  % find ./path/to/topoftree -name "*.o" -print0 | xargs -0 rm

And then try building again.

cr

There aren't any .o files. I hand-rm'd dist/.

I'll try to give it a try tomorrow

On 01/05/2012 10:25 PM, Steve Klabnik wrote:
> There aren't any .o files. I hand-rm'd dist/.

> I'll try to give it a try tomorrow

Awesome. Note that it's still failing on Travis:
http://travis-ci.org/#!/shoes/shoes/builds/482673

So you might need to do a little bit of screwing around with it, if
you're doing stuff on Linux or Windows.

On 01/05/2012 10:29 PM, Steve Klabnik wrote:
>> I'll try to give it a try tomorrow
> Awesome. Note that it's still failing on Travis:
> http://travis-ci.org/#!/shoes/shoes/builds/482673
>
> So you might need to do a little bit of screwing around with it, if
> you're doing stuff on Linux or Windows.
With my lack of knowledge of the concrete shoes implementation and C 
screwing around unfortunately isn't so easy, I get the following error: 
https://gist.github.com/1573067

I believe it's the following line:
shoes/native/gtk.c:1196:46: error: ‘cColor’ undeclared (first use in 
this function)

I guess I need to tell shoes somewhere to use the ruby version instead 
of the C version - but where?

Thanks and good night everyone!
Tobi

Ah! add a 'extern VALUE cColor' to the top of gtk.h

Sorry for being so late, lots of university stuff to do...

There is no gtk.h :-( I put it in native.h instead (I hope that's ok?) 
but I still get an error

(where I as a non Cer thin the error comes from)
shoes/native/gtk.c:1094:15: error: ‘_color’ undeclared (first use in 
this function)

(full error log for more knoledgable people)
https://gist.github.com/1595531

Tobi



On 01/07/2012 02:58 AM, Steve Klabnik wrote:
> Ah! add a 'extern VALUE cColor' to the top of gtk.h

On 01/05/2012 10:29 PM, Steve Klabnik wrote:
>> I'll try to give it a try tomorrow
> Awesome. Note that it's still failing on Travis:
> http://travis-ci.org/#!/shoes/shoes/builds/482673
>
> So you might need to do a little bit of screwing around with it, if
> you're doing stuff on Linux or Windows.
Linux. First finishing a paper until 16:00 and explaining why I think 
that the task is bad and hasn't been thought through, then meetup with 
friends and then some open source, sounds like a good day ;-)