Editing RubyEdit class

[Exo]

I noticed in the Flowstone install folder there is an fsm file called "system" in this schematic is a Ruby component that defines the RubyEdit class. I was thinking could we not alter this class to add line numbers ect?

I have played about with it a bit changing a few things in the syntax colouring ect, but couldn't get any line numbering going.
Can someone else that is fluent in Ruby see if that can be done?

Cheers,
Exo

[Nubeat7]

yes it should work with codeRay

this shows how to get the line numbers

Code: Select all

CodeRay.scan("5.times do\n  puts 'Hello, world!'\nend", :ruby).div(:line_numbers => :table)

but i don't know hoe to put them in?

http://coderay.rubychan.de/

[Exo]

Thanks.

Seems div is outputing HTML which will be why it doesn't work.

I have tried modifying a few things but no luck....

Code: Select all

require "coderay"

class RubyEdit

	Black = invertedColours ? [255,255,255] : [0,0,0]
	Brown = [128,64,0]
	Red = [163,20,20]
	Blue = [0,0,255]
	Green = [0,128,0]
	Pink = [200,0,200]
	Gold = [128,128,0]
	Purple = [100,0,200]
	Grey50 = [128,128,128]
	Silver = [100,100,64]
	
	def parse v,f

		sourceLines = Array.new
		v.each_line {|s| sourceLines.push s}
		startIndex=0
		lineStartIndeces=[0]
		startIndex=v.index("\n",0)
		while startIndex
			lineStartIndeces<<(startIndex+1)
			startIndex=v.index("\n",startIndex+2)
		end
				
		v.force_encoding "utf-8"
		tp=CodeRay.scan(v,:ruby).div(:line_numbers => :table) #Doesn't work it outputs HTML
		lines=[]
		nLines=sourceLines.count
		nLines.times {lines<<[]}
		ci=0
		strings=[]
		types=[]
		tp.each_with_index do |t,i|
			if i.even? then strings<<t end
			if i.odd? then types<<t end
			# Check for newlines inside the current string and split
			# where necessary
			if i.odd?
				if t != :string && t != :space && strings.last.class == String
					subs = strings.last.split("\n")
					if subs.size > 0 
						s = strings.pop
						t = types.pop
						subs.each do |str|
							if str.size > 0
								strings<<str
								types<<t
							end
							strings<<"\n"
							types<<:space
						end
						if s.index("\n") != s.length-1 
							s = strings.pop
							t = types.pop					
						end
					end
				end
			end
		end
		
		ln=0 
		for i in 0..strings.count-1
			item=[]
			t=types[i]
			s=strings[i]
			if s.eql? "\n"
				ln+=1
			elsif t!=:space && s.class==String
				i1=v.index(s,ci)
				if i1
					ci=i1+s.length
					item<<i1-lineStartIndeces[ln]<<s.length<<t
					if ln<nLines
						lines[ln]<<item
					end
				end
			end
		end

		# Set up colour scheme
		m={ :ident=>Black,:delimeter=>Black,:operator=>Black,
			:keyword=>Blue,:class=>Pink,:method=>Black,:comment=>Green,
			:class_variable=>Gold,:instance_variable=>Gold,:global_variable=>Silver,
			:symbol=>Grey50,:constant=>Pink,
			:integer=>Red,:float=>Red,:char=>Red,:content=>Red,:string=>Pink,
			:regexp=>Black,:shell=>Black,:schematic=>Purple }

		# Set up our own commands
		schemCmds = ["output","input","watch","releaseMouse","captureMouse",
					 "redraw","time","parent","caption","alert","showCursor"]

		li=0
		
		# Colour the lines
		lines.each do |l|
			f.newLine
			os = 0

			if li < sourceLines.size
				len = sourceLines[li].chomp.length
			end

			l.each do |i| 
				t=i[2] 
				
				# Check if we ran onto another line (due to a line continuation '\')
				if i[0]-os > len 
					f.endLine len
					os += len+2
					f.newLine
					li+=1
					if li < sourceLines.size
						len = sourceLines[li].chomp.length
					end
				end
				if (i[0]+i[1]-os-1) < len
					if( t == :ident && schemCmds.include?(sourceLines[li][(i[0]-os)..(i[0]-os+i[1]-1)]) )
						t = :schematic
					end
					f.addSection i[0]-os,i[0]-os+i[1]-1,m[t]
				end
			end
		
			f.endLine len
			li+=1
		end
	end
end

[Nubeat7]

yes, it outputs html...

i think the main problem is that codeRay is able to spitout line nrs, but what we need would be a nr table in the editor itself which would need to be an own area beside the editor area itself..

so we cannot add something to any v.source because it would distroy the code or?

i was thinking about to add a nr after each \n which doesn't affect the code behind - like only whats behind that is valid code - it but i`ve no idea if this would be possible?

[Exo]

yes, it outputs html...

i think the main problem is that codeRay is able to spitout line nrs, but what we need would be a nr table in the editor itself which would need to be an own area beside the editor area itself..

so we cannot add something to any v.source because it would distroy the code or?

i was thinking about to add a nr after each \n which doesn't affect the code behind - like only whats behind that is valid code - it but i`ve no idea if this would be possible?

I don't know if it is possible everything I am currently trying breaks the code.

I have found a new class though LineFormat , that is the f variable.

I think it should be doable but currently I am struggling to understand the code and so I cannot say for definite.

Edit... I actually yes I think you are right, I don't think it is possible because adding extra characters would break the code that is passed to the interpreter.

Was a nice idea though I just didn't understand the code until now.

[Nubeat7]

but would it be possible to show the actual code in an extra window including linenr?

i mean if we send self (from the window where some code is written) into another ruby module which just draws the code as text (including linenr) using coderay?

this would also be interesting for documetation as coderay enables you to filter out comments too, so you could generate a textfile which includes all the comments of your code.

[Exo]

I think that should be possible.

The coderay lines doesn't seem to work. To work we would need to copy the original code as a new string object and append the number to the beginning of each line manually and then this could be displayed.

[Nubeat7]

hmm, got some troubles with encoding to utf-8!?

here is a schematic of how i think it could work convert code into txt (if this ascii 8bit to utf-8 encoding would work)
this

Code: Select all

v.CodeRay.encode(v,:ruby,:text)

also only works if you include the encoders folder form the original coderay zip file into the flowstone ruby folder:..ruby/libraries/syntax/coderay !! which is not included in FS
.encode does scan input (v) to language(:ruby) and output to given format (:text) all in one, which i think should be the easiest way if i'm right with all that..

[trogluddite]

To work we would need to copy the original code as a new string object and append the number to the beginning of each line manually and then this could be displayed.

Yes, and there is a sneaky way to do it.
Systems are down at home at the moment for some home improvements, but I have it working and can post later.

The trick is to hijack a RubyEdit method called "parse" - it gets called every time the code is edited. IIRC, I intercepted the input arguments by doing something like this...

Code: Select all

def parse(*args, &block)
  @code_string = args[0]  # May not be the right index!
  super(*args, &block)  # Hand control back to the original method
end

@Exo - take a look in the code tracer prototype I sent you. it is in there somewhere.

[Exo]

hmm, got some troubles with encoding to utf-8!?

here is a schematic of how i think it could work convert code into txt (if this ascii 8bit to utf-8 encoding would work)
this

Code: Select all

v.CodeRay.encode(v,:ruby,:text)

also only works if you include the encoders folder form the original coderay zip file into the flowstone ruby folder:..ruby/libraries/syntax/coderay !! which is not included in FS
.encode does scan input (v) to language(:ruby) and output to given format (:text) all in one, which i think should be the easiest way if i'm right with all that..

Nice Idea but the encoding cannot work because the RubyEdit is it's own object it isn't a String object and so cannot be encoded.

That file crashes flowstone here if I double click the file instead of opening it inside FS.