[WASTE-list] calculating linebreaks

[Alfred Van Hoek]

First of all, Waste 3 automatically recalculates line-breaks when  
necessary, while in Waste 2 we need to tell waste it by calling  
WECalText. No problem.

However, recalculating line-breaks when the destination rectangle  
changes leads to spurious mistakes. We claim that the mistake is in  
the line break algorithm.

Here is the code that will be called during a life resize:

		/* remember offset of first visible character */
		WEGetViewRect(&lr, text); 		// the old view rect
		topCharPosition = *(LongPt*)&lr; 	// old topChar offset
		
		topCharOffset = WEGetOffset(&topCharPosition, &edge, text);
		dprintf("rewrapping: topCharOffset = %i, edge = %i, view left = %i,
				view top = %i", topCharOffset, edge, lr.left, lr.top);
	
		/* reset destination rectangle and recalculate line breaks */
		_SetDestinationRect(instance);
		_SetViewRect(instance);
		
		if ( (err = WECalText(text)) != noErr ) {
			dprintf("error calculating line breaks...");
			goto cleanup ;
		}
		
		lineWidth = WEGetMaxLineWidth(text);
		
		/* scroll the destination rectangle to keep the previous */
		/* first visible character at the top of the view rectangle */
		
		WEGetPoint(topCharOffset, kHilite, &topCharPosition, &lineHeight,  
text);
		dprintf("rewrapping: topCharPosition = %i, %i, lineWidth = %i,  
lineHeight = %i",
				topCharPosition.h, topCharPosition.v, lineWidth, lineHeight);
		
		// note that sometimes the topCharPosition is way off with respect  
to topCharOffset
		// and potentially is related to WECalText doing something wrong??
		
		WEGetDestRect(&lr, text);
		
		offsetX = controltop.h - topCharPosition.h;
		offsetY = controltop.v - topCharPosition.v;
		
		dprintf("offsetX %i, offsetY %i", offsetX, offsetY);
		WEOffsetLongRect(&lr, offsetX, offsetY);
		WESetDestRect(&lr, text);

	cleanup: ;


A good pass through the above gives:

rewrapping: topCharOffset = 1728, edge = 0, view left = 8, view top = 37
rewrapping: topCharPosition = 8, 613, lineWidth = 312, lineHeight = 16
offsetX 0, offsetY -576


and the next pass gives:

rewrapping: topCharOffset = 1728, edge = 0, view left = 8, view top = 37
rewrapping: topCharPosition = 324, 485, lineWidth = 391, lineHeight = 16
offsetX -316, offsetY -448

and suddenly the field is scrolled to the left. This offsetX change  
relates to the length of the word used in the field.

The example used was:

WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset WEGetOffset WEGetOffset  
WEGetOffset WEGetOffset WEGetOffset etc

and the field width was less than 5 words.

When the field width is increased to 5 words, Waste is moving the  
dest rectangle 4 words to the left if that would be of value here  
(Waste 2) or 2 words to the left (Waste 3). Then moving to a width  
that can contain 6 of the above words on a line, the field snaps back  
to an offsetX of 0:

rewrapping: topCharOffset = 1728, edge = 0, view left = 8, view top = 37
rewrapping: topCharPosition = 8, 421, lineWidth = 470, lineHeight = 16
offsetX 0, offsetY -384

With Waste 2, but snaps further to the left as a function of the word  
length.

Any thoughts, comments, workarounds etc?

Alfred

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://six.pairlist.net/pipermail/waste-list/attachments/20071027/844bf076/attachment.htm>