From: Pascal B. <pa...@dy...> - 2000-11-10 20:39:33
|
Well, I'm currently porting some of the changes of the DynAPI into my Dynacore dis. and I'm walking into a few things, so here's a list of questions/remarks: 1) The elm={}, css={} and doc={} lines have been removed from the DynLayer constructor.. Why? There's now more lines added to the other methods that check for the .created=true property this is slower and requires more lines of code, so a bigger file size.. ?! For webapps speed might not be an issue, but for multi-media (dhtml sites, games, etc) this IS a big issue. 2) setPageX() and setPageY() have been changed to call setX(), setY().. why? This is obviously slower (setX and setY call moveTo) AND more importantly not the same! the original setPageX/Y sets the pageX property, NOT the left property.. not sure how browsers handle the differences, but it doesn't sound right. 3) in the deleteChildElements there's no check to see if the child is actually created, I think it should be checked (again for reasons of speed, but also because it might use more memory then required, setting properties to false or empty while there not even created yet. 4) setBgImage() there was a bug-fix for NS in case the path was NULL, I think because netscape won't repaint the layer with an empty image, so you had to reset the bgcolor to make sure it was showing correctly.. I think that was the reason why it was there.. 5) Not sure about the current getClip() does it work correctly on an inline div that hasn't had a setSize() call? I think it should always read the real css.clip values that's it for now :) I'll be trying to create a new benchmark test this weekend doing better tests of all functionality (previous benchmarks didn't include all, this one should cover more of the DynAPI) cya, Pascal Bestebroer pa...@dy... http://www.dynamic-core.net |