The Lounge - CodeProject

i mean just in general.

as i said i ported a Splay tree function from typescript to C# and it went beautifully.

i was impressed by TS for that.

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

In a complete hypothetical case that will almost never happen I would prefer to port TypeScript to C# rather than JavaScript.
But I'm not going to write all my front-end code in TypeScript in case I'll ever have to.

The only reason I'd use TypeScript is because some people really can't handle the dynamic nature of JavaScript.
So it's more a tool to "fix" the incompetence of others (I have a very specific someone in mind because I once had to use some JavaScript he wrote Dead | X|

)
Who the hell has three different return types for a single function!? Dead | X|

Best,
Sander

sanderrossel.com
Continuous Integration, Delivery, and Deployment
arrgh.js - Bringing LINQ to JavaScript
Object-Oriented Programming in C# Succinctly

Sander Rossel wrote:
Who the hell has three different return types for a single function!?

Do you really have to ask me that? Who do you think would do something so diabolical?

Sander Rossel wrote:
But I'm not going to write all my front-end code

A splay tree isn't front end specific and frankly I have no idea why someone wrote it in typescript.

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

honey the codewitch wrote:
Do you really have to ask me that? Who do you think would do something so diabolical?

Even you know better than that!
Despite your complete lack of code sanitation I actually hold you in pretty high regard Big Grin | :-D

~~Don't get any ideas though, I'll still prick you with my pitchfork.~~
That sounded better in my head Laugh | :laugh:

honey the codewitch wrote:
and frankly I have no idea why someone wrote it in typescript.

Exactly my point! Wink | ;)

Best,
Sander

sanderrossel.com
Continuous Integration, Delivery, and Deployment
arrgh.js - Bringing LINQ to JavaScript
Object-Oriented Programming in C# Succinctly

Sander Rossel wrote:
Despite your complete lack of code sanitation I actually hold you in pretty high regard

However, that will change when you see my C# ISAM database code.

In any case, you're being sweet. I won't hex you today.

As far as writing splay trees in typescript i guess one could ask why write a database engine in c#? Smile | :)

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

honey the codewitch wrote:
why write a database engine in c#

Why not? It does the job Wink | ;)

honey the codewitch wrote:
In any case, you're being sweet. I won't hex you today.

Probably because it's weekend and I don't have the energy to be mean Poke tongue | ;-P

honey the codewitch wrote:
my C# ISAM database code

You still wrote it though.

Best,
Sander

sanderrossel.com
Continuous Integration, Delivery, and Deployment
arrgh.js - Bringing LINQ to JavaScript
Object-Oriented Programming in C# Succinctly

Sander Rossel wrote:
You still wrote it though.

Am writing it. Trying to do B+ trees. And once i get them working for in-memory i have to make them diskable, and you can't just use serialization. You have to keep an open file pointer and seek periodically and read/write the blocks you need. It's that kind of file I/O, so i'm still a ways off. I hope to have the B+ tree at least solidly in progress by tonight but we'll see.

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

here's the main render loop for Prang Replay, my midi looper for FL Studio

C++

void _stdcall Replay::Gen_Render(PWAV32FS DestBuffer, int &Length) {
	// this takes almost no CPU, despite how it looks
	// there are almost no cases where all this code will run on the same call.
	unsigned long long p = looper.pos;
	int bts = looper.beats;
	float bps = looper.tempo / 60;
	float spb = looper.sampleRate / bps;
	unsigned long long i = 0;
	midi_message midi;
	vector<Loop>& loops = looper.loops;
	// rotate the loops left or right, if requested
	if (1 < loops.size() && looper.rol != looper.ror) {	
		if(looper.rol) {
			loops.push_back(*loops.begin());
			loops.erase(loops.begin());
			if (-1 < looper.lastPushIndex) {
				--looper.lastPushIndex;
				if (0 > looper.lastPushIndex)
					looper.lastPushIndex += looper.loops.size();
			}
			looper.rol = false;
		}
		if(looper.ror) {
			loops.insert(loops.begin(), (*(loops.end() - 1)));
			loops.pop_back();
			if (-1 < looper.lastPushIndex) {
				++looper.lastPushIndex;
				if ((int)looper.loops.size() <= looper.lastPushIndex)
					looper.lastPushIndex -= looper.loops.size();
			}
			looper.ror = false;
		}
	}
	else
		looper.rol = looper.ror = false;

	// push the current loop, if requested.
	if(looper.push) {
		if (0xFFFFFFFFFFFFFFFF!=looper.tapPos) { // find our beats for bts
			unsigned long long len = p - looper.tapPos;
			i = (unsigned long long)ceil(len / spb);
			i = ((unsigned long long)ceil(i / bts))*bts; // quantize to Beats
			bts = (int)i;
			if (0 == bts)
				bts = looper.beats;

		}
		if (bts)
		{
			unsigned long long bp = p;
			vector<midi_event>::iterator it = looper.events.end() - 1;
			vector<midi_event>::iterator cur = it;
			unsigned long long bl = (unsigned long long)(spb*bts);
			unsigned long long bs = bp - bl;
			if (looper.events.size()) {
				while ((*it).pos > bs) {
					--it;
					if (looper.events.begin() == it)
						break;
				}
				while (looper.events.end() != it && it->pos < bs)
					++it;
			}
			Loop loop;
			loop.eventIndex = 0;
			loop.length = bl;
			loop.pos = 0;
			if (looper.events.size()) {
				while (it <= cur) {
					midi_event ev(*it);
					ev.pos -= bs;
					loop.events.push_back(ev);
					++it;
				}
			}
			loops.push_back(loop);
			looper.hanging = looper.keyContext; // store the last keys state
			looper.lastPushIndex = loops.size() - 1;
		}
		looper.push = 0;
		looper.tapPos = 0xFFFFFFFFFFFFFFFF;
		pEditor->ParamsToControls();
	}
	// iterate over the current loops, playing them.
	vector<Loop>::iterator it = loops.begin();
	while (it != loops.end()) {
		// play the next events from this loop
		Loop& loop = (*it);
		int lec = loop.events.size();
		if (lec) {
			while (loop.eventIndex >= lec)
				loop.eventIndex -= lec;
			unsigned long long n = Length + loop.pos;
			int nc = 0;
			while (true) {
				unsigned long long lp = loop.events[loop.eventIndex].pos;
				if (lp < loop.pos)
					lp += loop.pos;
				if(lp<n) {
					midi = loop.events[loop.eventIndex].midi;
					loop.noteCounter.process_message(midi);
					looper.noteCounter.process_message(midi);
					midi = looper.noteCounter.filter_message(midi);
					if (looper.enabled && midi.status)
						PlugHost->MIDIOut_Delayed(HostTag, (int)midi);
				}
				else break;
				++loop.eventIndex;
				if (loop.eventIndex >= lec)
					loop.eventIndex -= lec;
				 ++nc; if (nc == lec) break;
				
			}
			loop.pos += Length;
			if (loop.pos > loop.length)
				loop.pos -= loop.length;
		} else  
			it->pos+=Length;
		while (it->pos >= it->length)
			it->pos -= it->length;
		++it;
	}
	// play the unlooped main events
	while ((size_t)looper.eventIndex < looper.events.size()) {
		midi_event e(looper.events[looper.eventIndex]);
		if (Length <= e.pos - p)
			break;
		looper.noteCounter.process_message(e.midi);
		looper.incomingNoteCounter.process_message(e.midi);
		midi_message m = looper.noteCounter.filter_message(e.midi);
		if(looper.enabled && 0!=m.status)
			PlugHost->MIDIOut_Delayed(HostTag, (int)m);
		++looper.eventIndex;
	}
	// pop the current loop if requested, killing notes as needed.
	if(looper.pop) {
		if (loops.size()) {
			Loop* ploop = &(*loops.rbegin());
			for (int i = 0; i < 16; ++i) { // channels
				for (int j = 0; j < 128; ++j) { // notes
					looper.noteCounter.note[i][j]-=ploop->noteCounter.note[i][j];
					if (looper.noteCounter.note[i][j] <= 0 && ploop->noteCounter.note[i][j] > 0) {
						looper.noteCounter.note[i][j] = 0;
						midi_message m;
						m.extra = midiOut;
						m.status = 0x80 | i;
						m.data1 = j;
						m.data2 = 0;
						PlugHost->MIDIOut_Delayed(HostTag, (int)m); // send note off
					}
				}
			}
			loops.pop_back();
		}
		if (looper.loops.size() == looper.lastPushIndex)
			--looper.lastPushIndex;
		looper.pop = 0;
	}
	looper.pos += Length;
	Length = 0;
}

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

No - it's long gone.

Is it OK if I use one of yours, instead?

Ravings en masse^
"The difference between genius and stupidity is that genius has its limits." - Albert Einstein
"If you are searching for perfection in others, then you seek disappointment. If you seek perfection in yourself, then you will find failure." - Balboos HaGadol Mar 2010

use the above. It frightens animals and small children. Big Grin | :-D

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

No, but I remember what it was.

Duff's device laid on top of an if-else.

+1 I had to look up Duff's device. I've done it before, but i never knew it had a name. Laugh | :laugh:

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

I believe this was the function that made me completely reevaluate my approach to parsing. Ultimately, I came up with something FAR simpler, that wasn't going to be a nightmare to change when adding another condition to the code.

C++

bool parser::parseExecute()
{
   //Start at the bottom of the token and precedence stack.  Place these values onto the
   //workTokenStack and workPrecStack.  Check if the precedence value is lower than the
   //previous one and take action if it is the case

   tokStruct_ *workTokStack;
   precedence_ *workPrecStack;
   int numTok=0, numPrec=0;
   tokStackPos = -1;

   if (!setupTokStackMem(&workTokStack, ansStackAlloc) ||
             !setupPrecStackMem(&workPrecStack, ansStackAlloc)) {
      MsgBox("Unable to allocate memory in ExecuteToken");
      return false;
   }
   workPrecStack[0] = PrecGet;
   workTokStack[0].token = TokGet;

   do {
      ++tokStackPos;
      //pop on the next precedece and token
      if (tokStack[tokStackPos].token == TokValue) {
         //must place it on the ansStack and go for the next one
         ansStack[++ansStackPos] = valStack[tokStack[tokStackPos].varListPos];
         continue;
      }
      if (tokStack[tokStackPos].token == Tok0DimVar) {
         ansStack[++ansStackPos] = getValFromVarList();
         continue;
      }
      if (tokStack[tokStackPos].token == TokMDimVarEnd) {
         //need to place the correct array variable onto ansStack.
         //first, check to make sure that the last token on workTokStack is TokMDimVar
         if (workTokStack[tokStackPos-1].token == TokMDimVarAssign) break;
         if (workTokStack[tokStackPos-1].token != TokMDimVar) {
            MsgBox("']' encountered without arrayed variable");
            return moveCursorToError(origStr);
         }
         //now, need to place the appropriate value on the stack
         int Y, Z, pos, dim;
         Y = varList[workTokStack[numTok-1].varListPos].Y;
         Z = varList[workTokStack[numTok-1].varListPos].Z;
         dim = varList[workTokStack[numTok-1].varListPos].varDim;
         if (dim == 1) pos = ansStack[ansStackPos];
         else if (dim == 2) pos = ansStack[ansStackPos-1] +
                  ansStack[ansStackPos]*Y;
         else if (dim == 3) pos = ansStack[ansStackPos-2] +
                  ansStack[ansStackPos-1]*Y + ansStack[ansStackPos]*Y*Z;

         //finally,
         ansStackPos = ansStackPos-dim+1;
         if (varList[workTokStack[numTok-1].varListPos].varType == Int) {
            ansStack[ansStackPos] =
            *((int*)(varList[workTokStack[numTok-1].varListPos].varAddress)+pos);
         }
         else if (varList[workTokStack[numTok-1].varListPos].varType == Double) {
            ansStack[ansStackPos] =
            *((double*)(varList[workTokStack[numTok-1].varListPos].varAddress)+pos);
         }
         else {
            MsgBox("Bad varType placed on ansStack");
            return moveCursorToError(origStr);
         }
         ansStackPos++;
         numTok--;   //take the variable off of the workTokStack
         continue;
      }

      ++numTok; ++numPrec;
      workTokStack[numTok] = tokStack[tokStackPos];
      workPrecStack[numPrec] = precStack[tokStackPos];
      if (workTokStack[numTok].token == TokLParen) continue;
      while (workPrecStack[numPrec-1] >= workPrecStack[numPrec]) {
         //pop the precedence stack down
         workPrecStack[numPrec-1] = workPrecStack[numPrec];
         numPrec--;
         switch (workTokStack[numTok-1].token) {
            case TokValue: //place the value on the answer stack
               ansStack[++ansStackPos] =
                     *(double*)varList[workTokStack[numTok-1].varListPos].varAddress;
               numTok++;
               numPrec++;
               tokStackPos++;
               break;
            case TokVarAssign: //place the value in the answer stack into the variable
               if (ansStackPos != 0) {
                  MsgBox("Malformed statement");
                  delete workTokStack;
                  delete workPrecStack;
                  return false;
               }
               *(double*)varList[workTokStack[numTok-1].varListPos].varAddress = ansStack[0];
               //pop the semiColon off the stacks
               numTok=numTok-2;
               numPrec--;
               ansStackPos--;
               goto pETBreakout;
            case TokAdd: //add the last two values in the answer stack
               ansStack[ansStackPos-1] = ansStack[ansStackPos-1] + ansStack[ansStackPos];
               ansStackPos--;
               numTok--;
               break;
            case TokSubtract:
               ansStack[ansStackPos-1] = ansStack[ansStackPos-1] - ansStack[ansStackPos];
               ansStackPos--;
               numTok--;
               break;
            case TokMultiply:
               ansStack[ansStackPos-1] = ansStack[ansStackPos-1] * ansStack[ansStackPos];
               ansStackPos--;
               numTok--;
               break;
            case TokDivide:
               ansStack[ansStackPos-1] = ansStack[ansStackPos-1] / ansStack[ansStackPos];
               ansStackPos--;
               numTok--;
               break;
            case TokPower:
               ansStack[ansStackPos-1] = pow(ansStack[ansStackPos-1], ansStack[ansStackPos]);
               ansStackPos--;
               numTok--;
               break;
            case TokSemiCol:  //it will reach here when the End token is reached
               if (ansStackPos != 0) {
                  MsgBox("Misformed Statement");
                  delete workTokStack;
                  delete workPrecStack;
                  return false;
               }
               ansStackPos--;
               break;
            case TokComma:
               //need to shrink workPrecStack and workTokStack to eliminate the comma.
               //workTokStack is popped down below.  workPrecStack has been popped down
               //up above
               numTok--;
               break;
            case TokLParen:
               //There are only two times this will be entered: 1) there are two or more
               //  left parenthesis on the stack back to back, or 2) a closing parenthesis
               //  and a opening parenthesis are back to back on the workTokStack
               if (tokStack[tokStackPos].token == TokLParen &&
                        tokStack[tokStackPos-1].token == TokLParen) {
                  numPrec++;
                  goto pETBreakout;
               }
               numTok=numTok-2;
               numPrec--;
               goto pETBreakout;
            case TokRParen:
               break;
            case TokMDimVar:
               //it only reaches here when assigning the last value on ansStack to the
               //appropriate location in memory for the variable.
                  int Y, Z, pos, dim;
                  Y = varList[workTokStack[numTok-1].varListPos].Y;
                  Z = varList[workTokStack[numTok-1].varListPos].Z;
                  dim = varList[workTokStack[numTok-1].varListPos].varDim;
                  if (dim == 1) pos = ansStack[ansStackPos-1];
                  else if (dim == 2) pos = ansStack[ansStackPos-2] +
                           ansStack[ansStackPos-1]*Y;
                  else if (dim == 3) pos = ansStack[ansStackPos-3] +
                           ansStack[ansStackPos-2]*Y + ansStack[ansStackPos-1]*Y*Z;

                  //finally,
                  if (varList[workTokStack[numTok-1].varListPos].varType == Int) {
                     *((int*)(varList[workTokStack[numTok-1].varListPos].varAddress)+pos) =
                              ansStack[ansStackPos];
                  }
                  else if (varList[workTokStack[numTok-1].varListPos].varType == Double) {
                     *((double*)(varList[workTokStack[numTok-1].varListPos].varAddress)+pos) =
                              ansStack[ansStackPos];
                  }
                  else {
                     MsgBox("Bad varType placed on ansStack");
                     return moveCursorToError(origStr);
                  }
                  ansStackPos = ansStackPos-dim;
                  numTok--;   //take the assignment off of the workTokStack
                  continue;

            default:
               MsgBox("Still Programming!");
               delete workTokStack;
               delete workPrecStack;
               return false;
         }
         //pop the workTokStack down
         workTokStack[numTok] = workTokStack[numTok+1];
      } //end of while loop for processing previous tokens.
pETBreakout:
   } //end of do loop
   while (workPrecStack[numPrec] != PrecEnd);
   delete workPrecStack;   workPrecStack = NULL;
   delete workTokStack;      workTokStack = NULL;
   return true;
}

The forgotten roots of science | C++ Programming | DWinLib

yeah, two stacks is complicated.

you should just need the one at most (not even that if you're borrowing *the* stack (call stack) using recursive descent.

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

Pfft. That's art to the stuff I've written. My stuff is so bad if I posted it here your computer would keel over in fright and refuse to reboot.

cheers
Chris Maunder

I'm glad i'm not the only target of Sander and his pitchfork. Laugh | :laugh:

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

Yes you are.
Speaking of which...

Best,
Sander

sanderrossel.com
Continuous Integration, Delivery, and Deployment
arrgh.js - Bringing LINQ to JavaScript
Object-Oriented Programming in C# Succinctly

i'll turn you into a VB6 developer

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

Been there done that Big Grin | :-D

Best,
Sander

sanderrossel.com
Continuous Integration, Delivery, and Deployment
arrgh.js - Bringing LINQ to JavaScript
Object-Oriented Programming in C# Succinctly

In that case I'll hex your boss so you get stuck on front end work.

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

I'm my own boss Sigh | :sigh:

Best,
Sander

sanderrossel.com
Continuous Integration, Delivery, and Deployment
arrgh.js - Bringing LINQ to JavaScript
Object-Oriented Programming in C# Succinctly

Even better. Laugh | :laugh:

When I was growin' up, I was the smartest kid I knew. Maybe that was just because I didn't know that many kids. All I know is now I feel the opposite.

Is it part of the code behind CP...? Unsure | :~

dandy72 wrote:
the code behind CP

No.

That code's even worse Wink | ;)

cheers
Chris Maunder

Yes, and it's a work of art. Picasso is also art mind you.

I spent two weeks on it and it solved quite a few problems in one go.
<edit>Two different ways to parse a tree and a statemachine for the nodes.</edit>
I was very proud of it at the moment.

Today I look at it and think WTF!

Wrong is evil and must be defeated. - Jeff Ello

modified 13-Sep-19 18:37pm.

Welcome to the Lounge