@dan rathbun said:
How about a Ruby Cheatsheet?
Ruby QuickRef
http://www.zenspider.com/Languages/Ruby/QuickRef.html
Excellent!
I took the liberty of converting the cheatsheet to lowercase tags and attributes, then fiddled with the TOC a bit. I've written the author requesting permission to continue fiddling.
The camera is easy to position using Ruby and the API. Watch the first 30 seconds of my Airshow! movie for a good example.
What is the requirement for a separate C++ program?
@honoluludesktop said:
I often need a more baby one. Any suggestions?
@adamb said:
Strength Reduction is the transformation that would turn these loops into do_nothing loops.
Yes, but that's for really sharp compilers. I was just expecting func( 1.0 + 2.0 ) to be turned into func( 3.0 ).
@thomthom said:
Looks like it evaluates the string and treats the backslashes as special characters.
I had no end of troubles with slashes. Finally implemented this none-too-clever code and all the bugs went away:
=begin
qq() and qq!() do these replacements;
"'" -> "qq1"
'"' -> "qq2"
'/' -> "qq3"
'\' -> "qq4"
'\n' -> "qq5"
'\r' -> "qq6"
=end
def self.qq( str )
return str.gsub(
/\'/, 'qq1' ).gsub(
/\"/, 'qq2' ).gsub(
/\//, 'qq3' ).gsub(
/\\/, 'qq4' ).gsub(
/\n/, 'qq5' ).gsub(
/\r/, 'qq6' )
end # of qq()
def self.qq!( str )
str.gsub!( /\'/, 'qq1' )
str.gsub!( /\"/, 'qq2' )
str.gsub!( /\//, 'qq3' )
str.gsub!( /\\/, 'qq4' )
str.gsub!( /\n/, 'qq5' )
str.gsub!( /\r/, 'qq6' )
end # of qq!()
These reversers complete the Ruby:
def self.unqq( str ) # reverses qq()
return str.gsub(
/qq1/, '\'' ).gsub(
/qq2/, '"' ).gsub(
/qq3/, '/' ).gsub(
/qq4/, '\\' ).gsub(
/qq5/, '\n' ).gsub(
/qq6/, '\r' )
end # of unqq()
def self.unqq!( str ) # reverses qq!()
str.gsub!( /qq1/, '\'' )
str.gsub!( /qq2/, '"' )
str.gsub!( /qq3/, '/' )
str.gsub!( /qq4/, '\\' )
str.gsub!( /qq5/, '\n' )
str.gsub!( /qq6/, '\r' )
end # of unqq!()
These are the reversers in JavaScript:
single_quote = String.fromCharCode( 39 );
double_quote = String.fromCharCode( 34 );
ford_slash = String.fromCharCode( 47 );
back_slash = String.fromCharCode( 92 );
function unqq( msg ) {
s = msg.replace( /qq1/g, single_quote );
s = s.replace( /qq2/g, double_quote );
s = s.replace( /qq3/g, ford_slash );
s = s.replace( /qq4/g, back_slash );
s = s.replace( /qq5/g, '\\n' );
s = s.replace( /qq6/g, '\\r' );
return s;
}
@thomthom said:
...and I believed that. It is after all a scripting language we're talking about. But wow! - it's slow!
> > t=Time.now; 1000000.times { 1.0 }; puts Time.now - t
> 0.1
> nil
>
> > t=Time.now; 1000000.times { 1.0 + 2.0 }; puts Time.now - t
> 0.985
> nil
>
> > t=Time.now; 1000000.times { 1.0 + 2.0 + 3.0 }; puts Time.now - t
> 1.85
> nil
>
> > t=Time.now; 1000000.times { 1.0 + 2.0 + 3.0 + 4.0 }; puts Time.now - t
> 2.76
> nil
>
> > t=Time.now; 1000000.times { 1.0 + 2.0 + 3.0 + 4.0 + 5.0 }; puts Time.now - t
> 3.685
> nil
>
If I remember correctly, compilers started adding constant folding in the '80s. This is truly brain dead.
On the other hand, Ruby allows 1_000_000.times ... when counting zeros gets annoying. A peculiar mixture of great stuff and trash.
In my tutorial I suggest using a spare browser as a reference guide. One tab for the SketchUp Ruby API, another for Ruby, more for JavaScript, CSS and HTML.
I use this setup and it's very helpful, except for Ruby. My tab is http://ruby-doc.org/core/. In practice I find it's faster to do a Google search.
For the rest I've found references that I'm happy with, but not for Ruby. Suggestions?
Take a look at my 3-minute movie to see SketchUp in its full-4D glory.
Then use the code in my tutorial's Chapter 16 for classes that let you do all this without need for the APIs Transformation class.
In addition to TIG's info, you should move all the one-time code outside the loop.
model = Sketchup.active_model
entities = model.active_entities
path = Sketchup.find_support_file "mymodel.skp" ,"Components/"
. . .
I also heartily recommend that you use four-space indentation in spite of the fact that two spaces is a Ruby convention. Four is much more readable which is why it is mandatory in Python, convention in C++, Java, JavaScript, ... (every language I know, except Ruby).
New languages are always a PITA. Fortunately, quite a small subset of Ruby is all you really need to create SketchUp plugins.
@unknownuser said:
rescue Exception => eto catch everything.
I don't want to seem ungrateful, but that upgraded me to Bug Splat. Is there a
rescue BugSplat => e
???
I've set single click to open. An HTM opens in Chrome.
I'm writing a console. You input 2+2 and it outputs 2+2 # 4. Problem is, your input may not be perfect, so a little error trapping is needed.
In this part of the process() function (line #s added):
59 results = []
60 for line in lines do
61 rslt = line.chomp() + ' # '
62 begin
63 rslt += eval(line).to_s().chomp()
64 rescue => e
65 rslt += e.to_s().chomp()
66 end
67
68 results.push( qq(rslt) )
69 end # loop over lines
I'm thinking that an error at line 63 gets caught and reported at line 65. It usually does. Now, however, I've got this:
Any ideas why this isn't caught by the rescue statement? Any ideas on how to catch it?
Let's see - for performance I'm going to avoid iterations, arrays, hashes and objects.
What's left?
Ruby has its wonderful numbers (integers when possible, floats as needed) but they aren't C's much simpler (and faster) plain numbers. Cs numbers can be used directly by the CPU/FPU. Ruby's numbers are object instances.
@dan rathbun said:
You should have instead had the &Edit association pointing to Notepad++
Folder Options has disabled what I need:
I gave the user admin privileges, but that didn't help. I right-clicked in Explorer, Properties, and was able to change the default Open to a browser, getting openURL() working but that doesn't allow setting the Edit default.
Can I go back to KDE now?
Ruby handles text files correctly on both Windows and Mac. The line separator is the newline character (10, '\n'), regardless of platform.
This does not extend to text retrieved from a WebDialog textarea widget, which will be delimited with newlines on a Mac, with return/newline pairs on a PC. This little bit of extra code is needed:
text = web_dialog.get_element_value( 'textarea_id' ) text.gsub!( /\r\n/, "\n" ) # needed extra!
Thank Todd Burch for his Mac support.
@unknownuser said:
I personally see no productivity or functionality gains from bothering with each platforms newline patterns, unless you just want to.
This isn't the file I want given the input textarea:
You have to convert crlf to newline because Ruby converts newline to crlf as it writes the file on a PC. Ugh.
# ta_sep.rb
require 'sketchup'
html="
<html> <body>
<textarea id='ta'>one
two</textarea>
<script>
onload = function() { location = 'skp;go'; }
</script>
</body> </html>"
wd = UI;;WebDialog.new( "Test", true, "test", 400, 300, 100, 100, true )
wd.set_html( html )
wd.add_action_callback( "go" ) do | dlg, msg |
val = wd.get_element_value( 'ta' )
# val.gsub!( /\r\n/, "\n" )
pn = UI.savepanel(
"Save As ...", File.dirname(__FILE__), 't.txt' )
file = File.new( pn, 'w' )
file.puts( val )
file.close()
end
wd.show()
I think this is correct cross-platform code if you uncomment the gsub!(). For this I be thanking you and Notepad++ (View/Show Symbol/Show All Characters).
Ruby's lovely ability to handle everything with newlines, platform-independently, does not extend to text read from a browser's textarea widgets.
@unknownuser said:
You're welcome!
Thank you!
@unknownuser said:
Why are you coding at that level?
Stupidity or ignorance. Maybe both.
@unknownuser said:
Why not just bust the input data up into ruby "newlines"?
Hmmm. This is the code that gets a WebDialog <textarea>'s data into a disk file:
# write the file
str = wd_console.get_element_value( 'console' )
str.gsub!( /\r\n/, "\n" ) # why? doesn't look PC, but it works
file = File.new( pathname, 'w' )
file.puts( str )
file.close()
I'm thinking that will work fine on a Mac, because the gsub!() will find zero occurrences. Now the other direction: you read the file as a string and pass the string to the textarea and it will be Win/Mac compatible.
@cjthompson said:
even just
UI.openURL("notepad")will also work.
You never know when one little factoid will solve a problem.
I tried notepad (it works), then notepad++ (also works) and then an HTML file. It opened the HTML, but it opened it in Notepad++.
Now if only I can figure out how to specify both program and file ...
Don't think I want to dust off Decaf until I've got time to go after it. I forget where I put my mains in Java. There's tons of my Java on my site: http://www.MartinRinehart.com , including the original Decaf tokenizer.
There's a whole new vocabulary (words, not tokens; phrases, not expressions; sentences, not statements; etc.) that I'll need to document before anyone could realize that I wasn't just babbling.
See "The Project", my site, for more.