No, there is no caching going on, so it is probably slower than the original code. However, I suspect it is somewhat faster than the current implementation in the renderer branch, since it doesn't draw on an image, and then draw that image onto the paint canvas. The painter is passed directly to the code that draws the symbol.
<br>I'm curious how the speed of painting an image compares with the speed for painting a circle, square, etc. It seems to me that painting primitives might be faster than painting an image, but that needs to be tested.
<br>If someone can send me a large point set, I will try to test the speed of these different implementations.<br>Steven<br><br><br>I don't have a large point dataset to test it with at the moment.<br><br><div class="gmail_quote">
On Jan 9, 2008 9:04 AM, Magnus Homann <<a href="mailto:magnus@homann.se">magnus@homann.se</a>> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="Ih2E3d">Steven Bell wrote:<br>> I thought I had sent this out earlier, but realized I hadn't...<br>> I've committed my symbol rendering code changes into the composer_redesign<br>> branch. The major change is that point symbols are no longer drawn as
<br>> pixmaps on the canvas (which I think has been done in the renderer branch,<br>> although probably in a different way). This also happens to fix the<br>> clipping issues.<br></div>Steven,<br><br>How has this affected rendering speed for normal rendering. Are you
<br>still caching?<br><font color="#888888"><br>Magnus<br><br></font></blockquote></div><br>