I will reply for each. Detailing my thought pattern. I know all the "rules and conventions" of smalltalk but i "can't play" very well.
At first (Point>>#x) parseTree explore just looks weird. Then things start to settle. () Paranthesis means i don't actually have to care what is in there, conclusion is whatever happens the product of () will be an object. I'm actually scared of >>, since i don't know what it does… i've seen it in old code snippets of class definitions. parseTree seems to be a method, and so does explore. Going on… it seems to be a chain of unary messages. Based on first trial conclusion it might be a statement even though the "." at the end is missing. I paste in the Workspace and right click 'Do It' on the whole thing. A window opens with lots of information in it. I can't make sense of it since i have no idea what i just did. I need more information on what i just did. Restart from the beginning. "Point" might be a point like in a cardinal system… x is there so i might be right. I replace x with y and 'Do It' again… another window pop open… seems to be quite similar to the first one only it has y in it instead of x. I continue the assumption and replace y with z. I get a KeyNotFound:, i don't understand this but it's similar to MessageNotUnderstood:. Since success would have meant i would get a similar window as in the x y case, getting a different window means the test failed. I can only use x and y in the statement as it is now. Point might be a class, i search for it… it seems to have instance variable names by that name. I add z to its list of instance variable names, save, replace it in the command… error. Assumption was incorrect. I look at the Point class again notice it also has accessor methods by that name, just x and y … there is no z. Agrees with the failure of earlier experiment, i might be on the right track. I choose another method of Point. Execute the command again… get the "success" window with lots of information on it. I choose another class and one of its methods, execute the command again… get the "success" windows with lots of information on it. Conclusion… I can put any class name and choose any method it has and the command will pop up information about if. I don't understand the information from the pop up. Restart from the beginning. I search for parseTree in the browser and there seems to be a lot of implementors of that method. I search for explore in the browser and Object seems to implement it. I 'Inspect It' on just the (Point>>#x) part and it seems to be producing a CompiledMethod object. Based on assumption from the beginning whatever happens between () is an object this object then responds to parseTree. I 'Inspect It' on just (Point>>#x) parseTree and get an RBMethodNode object. Seems the parseTree gets CompiledMethod as input and outputs this different kind of object called RBMethodNode. Adding explore, based on the fact that explore is a method of object, seems to be just a generic printout of this entire object. The end. Time elapsed, a little over 3 min. Took longer to write the email. Pending questions: What exactly does >> do here? I know # is followed by a string literal normally in '' but since there are no spaces just the characters will do. Pe 26.01.2013, la 13:41, Stéphane Ducasse [via Smalltalk] <[email protected]> a scris: > > It's like being a two-eyed man in a dark room with no flashlight. Your eyes > > are there, but without the flashlight there might as well be no room. > > > > Without previous knowledge of how this AST is implemented my best bet is do > > the "architectural plan" example i gave Dale H. > > (Point>>#x) parseTree explore > will show you the tree then you can navigate in it. -- View this message in context: http://forum.world.st/How-do-i-list-the-execution-order-of-statements-inside-a-method-tp4665303p4665574.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.
